reproductions supplied by edrs are the best that can be made - … · 2014-06-09 · museums and...

DOCUMENT RESUME

ED 482 044 IR 058 756

AUTHOR Bearman, David, Ed.; Trant, Jennifer, Ed.

TITLE Museums and the Web 2001: Selected Papers from anInternational Conference (5th, Seattle, Washington, March 15-17, 2001).

INSTITUTION Archives and Museum Informatics, Pittsburgh, PA.ISBN ISBN-1-885626-23-1PUB DATE 2001-00-00NOTE 218p.; Accompanying CD-ROM is not available from ERIC. For

selected individual papers, see IR 058 757-776.AVAILABLE FROM Archives & Museum Informatics, 2008 Murray Ave., Suite D,

Pittsburgh, PA 15217. E-mail: [email protected]; Web site:http://www.archimuse.com/. For full text:http://www.archimuse.com/ mw2001/.

PUB TYPE Collected Works Proceedings (021) Non-Print Media (100)EDRS PRICE EDRS Price MF01/PC09 Plus Postage.DESCRIPTORS Access to Information; Archives; Computer Oriented Programs;

Conferences; Exhibits; Information Management; InformationTechnology; *Museums; Online Systems; TechnologicalAdvancement; *World Wide Web

IDENTIFIERS Museum Collections; Museum Libraries; *Virtual Museums; *WebSite Design

ABSTRACT

In this selection of papers from the conference, authors from10 of the more than 35 countries and every continent (except Antarctica)provide discussions covering all levels of museum Web design. They brought awide variety of experiences and backgrounds to the conference, all of whichensured new perspectives and new ideas. The meetings opened on Thursday witha pienary address, followed by a full day of sessions and two Crit Rooms.Friday featured sessions on accessibility, portals, and data interchange.This day also featured a variety of Mini-Workshops. The final day began withtwo groups of demonstrations that were followed by sessions on design,licensing, schools, and new technology. MW2001 ended with a closing plenary.This proceedings includes the register; schedule of workshops by leaders inthe museum computing field; schedule of sessions, with links to abstracts andfull text papers; a table listing speakers (identifying country, presentationtitle and indicating whether the paper is online); schedule of interactionsessions (Mini-Workshops and Crit Rooms), with links to abstracts of Mini-Workshops; schedule and description of demonstrations; list of exhibits withdescriptions of vendors; schedule of events; "Best of the Web" awards; keydates for future meetings; links to Seattle and Puget Sound area museums andtourist information; and list of sponsors. An accompanying CD-ROM includes: alist of all the speakers at the conference and links to their abstracts,biographies, and papers (where available); an overview of the Museums and theWeb 2000 conference program and links to abstracts and paper biographies; andthe results of the Best of the Web 2000 conference (requires Internetconnection) . (AEF)

Reproductions supplied by EDRS are the best that can be madefrom the original document.

Museumsand the

Web 2001Selected Papers from

an International Conference

PERMISSION TO REPRODUCE ANDDISSEMINATE THIS MATERIAL HAS

BEEN GRANTED BY

D. Bearman

TO THE EDUCATIONAL RESOURCESINFORMATION CENTER (ERIC)

U.S. DEPARTMENT OF EDUCATIONOffice of Educational Research and Improvement

EDUCATIONAL RESOURCES INFORMATIONCENTER (ERIC)

(Z121 This document has been reproduced asreceived from the person or organizationoriginating it.

0 Minor changes have been made toimprove reproduction quality.

Points of view or opinions stated in thisdocument do not necessarily representofficial OERI position or policy.

edited byDavid Bearman and

Jennifer Trant

2

BEST COPY AVAILABLE

www.archimuse.com/mw2001/

Museumsand the

Web 2001

Archives Ce Museum Informatics

Consulting, Publishing and Trainingfor Cultural Heritage Professionals

Informatics: The interdisciplinary study of information content,representation, technology and applications and the methods andstrategies by which information is used in organizations, networks,cultures and societies.

Archives & Museum Informatics organizes conferences,workshops and seminars, publishes journals and mono-graphs, and consults for cultural heritage organizationsworldwide. For the past 10 years, our educational goalhas been to provide specialists in archives and museuminformatics with timely and challenging opportunitiesfor professional exchange and training. Our consultingservices emphasize inter-institutional collaboration,strategic planning and standards-based solutions.

To be kept informed of coming events, conferencesand new publications contact us at:

Archives & Museum Informatics2008 Murray Avenue, Suite DPittsburgh, Pennsylvania15217 USAphone: +1 412 422 8530fax: +1 412 422 8594email: [email protected]://www.archimuse.com

Ask to be put on our mailing list.

www.archimuse.com/mw2001/

Museumsand the

Web 2001

Selected Papers froman International Conference

edited byDavid Bearman and

JenniferTrant

Museums and the Web 2001Selected papers from an international conference

edited by David Bearman and Jennifer TrantPittsburgh, PA:Archives & Museum InformaticsISBN: I -885626-23- I

@Archives & Museum Informatics, 2001

Printed in the United States of America

6

Contents

IntroductionThe Web as a Fact of Museum Life

David Bearman and Jennifer Trant,Archives & Museum Informatics, USA 3

User Influenced Design

Bringing the Curatorial Process to the WebCarrie Adams,Traci Cole, Christina De Paolo, Susan Edwards,Seattle Art Museum, USA I I

"Less Clicking, More Watching": Results from the User-Centered Design of a Multi-Institutional Web Site forArt and CultureJohn Vergo, Clare-Marie Karat, John Karat, Claudio Pinhanez,Renee Arora,Thomas Cofino, Doug Riecken, Mark Podlaseck,BM T.J.Watson Research Center, USA 23

Accessibility Techniques for Museum Web SitesSusan Anable,The Dayton Art Institute and Adam Alonzo,Wright State University, USA 33

Evaluating the Usability of a Museum Web SiteIlse Harms and Werner Schweibenz, University of Saarland,Saarbrücken, Germany 43

Design for(re-) Use

A Re-usable Software Framework for Authoring and ManagingWeb ExhibitionsChristian Breiteneder, Hubert Platzer,Vienna University of Technology,Martin Hitz, University of Klagenfurt,Austria 55

Design and Analysis ofVirtual MuseumsGilles Falquet, Jacques Guyot, Luka Nerima, University of Geneva,Switzerland and Seongbin Park, Information Sciences Institute, USA 65

Re-Purposing of Content and Digital Delivery Convergence:Implications for Interface DesignSlavko Milekic, M.D., PhD,The University of the Arts, USA 77

7

©Archives & Museum Informatics, 2001

Table of Contents

New Content / New Uses

A New Way of Making Cultural Information ResourcesVisible on the Web: Museums and the Open ArchivesInitiativeJohn Perkins, CIM1 Consortium, Canada 87

Using Interactive Broadband Multicasting in a MuseumLifelong Learning ProgramLeonard Steinbach,The Cleveland Museum of Art, USA 93

Streaming Audio and Video: New Challenges and Opportuni-ties for MuseumsJim Spadaccini, Ideum, USA 105

Personalizing Web Experiences

Co-operation Metaphors for Virtual MuseumsThimoty Barbieri, Paolo Paolini, Dept. of Electronics andComputer Science, Politecnico di Milano, Italy I 15

The HyperMuseum Theme Generator System: Ontology-based Internet Support for the Active Use of DigitalMuseum Data for Teaching and PresentationPeter Stuer, Robert Meersman, Steven De Bruyne,STARLab VUB, Belgium 127

Providing Personal Assistance in theSAGRES Virtual MuseumAna Carolina Bertoletti, Museu de Ciências e Tecnologia PUCRS,Marcia Cristina Moraes, PPGCC-UFRGS,Antônio Carlos da RochaCosta, Escola de Informatica UCPEL, Brazil 139

OPALES:An Environment for Sharing Knowledge amongExperts Working on Multimedia ArchivesHenri Betaille, Marc Nanard, Jocelyne Nanard, LIRMM, France 145

Being a Museum, Digitally

"Beyond museum walls" -A critical analysis of emergingapproaches to museum web-based educationKevin Sumption, Powerhouse Museum, Australian MuseumsOn-Line, Australia 155

8

vi © Archives & Museum Informatics, 2001

Museums and the Web 2001

1000 years of the Olympic Games:Treasures of AncientGreece: Digital Reconstruction at the Home of the GodsSarah Kenderdine, Powerhouse Museum,Australia 163

Engaging Museums, Content Specialists, Educators, andInformation Specialists:A Model and ExamplesC. Olivia Frost, University of Michigan, USA 177

Unseen Discussions: [email protected] Huberman P.S.1 Contemporary Art Center, USA 189

A Curatorial Resource for Upstart Media BlissBeryl Graham and Sarah Cook, University of Sunderland, UK 197

Maintaining the Magic

The Virtual Aura - Is There Space for Enchantment in aTechnological World?Susan Hazan,The Israel Museum, Jerusalem, Israel 209

About the AuthorsLearn about the People Represented Here 221

About the CD-ROMFull documentation of MW200 I and digital versions of submitted papers 230

©Archives & Museum Informatics, 2001 vii

Introduction

10

The Web as a Fact of Museum Life

David Bearman and Jennifer Trant,Archives & Museum Informatics, USA

The fifth year of Museums and theWeb comes at a time when the use of theWebby museums and the need of museums to take advantage of the Web aregivens. No-one questions the appropriateness of a museum Web site, only howwell it achieves the goals the museum has set,and how fully those goals encompassthe mission that the museum could be taking on in an age of nearly ubiquitous,increasingly broad-band, and narrow-casting. Each of us now expects to be ableto use the Web to get instant information about museum offerings both in-townand around the world. In North America we imagine receiving Web content onever larger monitors at home and the office; in Europe, with the proliferation ofcell phones and wireless PDAs, the dynamics of access may soon (if only for ashort time) feel very different. But the assumption of instant information, ondemand, anywhere, anytime, remains.

For museums, the opportunities have never looked greater. In the wake of thebursting of the e-commerce bubble of the year 2000, there is a new focus onmission both in the corporate and non-for-profit sectors and a reaffirmeddesire to use the medium to connect with our audience and to create somethingreally useful.

.museumThe past year also saw the first, brave moves of ICANN,The Internet Corporationfor Assigned Names and Numbers (http://www.icann.org), towards opening upthe name-space of thematic Internet Top Level Domains beyond the familiar .com,.edu,.org,.gov and the two-letter country designations.The place of museums onthe Web was reaffirmed when MuseDoma (The Museum Domain ManagementOrganization http://www.musedoma.org) was granted the right to administera .museum top level domain. (See http://www.icann.org/tIds/mus 1 / for theapplication, and http://www.icann.orglannouncements/icann-pr 1 6nov00.htm forthe announcement.) MuseDoma, was founded by the International Council ofMuseums (ICOM) and the J. Paul GettyTrust purposefully to submit the .museumapplication to ICANN, "for the benefit of the museum community world-wideand for the Internet itself. This new TLD was intended to give museums thepossibility of registering Internet addresses with a distinct museum identity andwould allow the users of the Internet to recognize this as a sign of authenticity"(http://www.musedoma.org/).

MuseDoma now faces all the challenges of administering a collaboration in theopen world of the Internet. As "registration in .museum will be restricted tomuseums and museum organizations" (http://www.musedoma.org) the refinementof the ICOM definition of museum and the development of MuseDoma operatingpolicies and procedures that both maintain the distinct sense of museum spaceand encourage the inclusive and participatory philosophy of much net art andmuseum Web activity will be a difficult task.

Museums and the Web 2000:ProceedingsArchives & Museum Informatics, 2001, p. 3

Bearman &Trant,The Web as a Fact of Life

We offer our collective thanks to Cary Karp, of the Swedish Museum of NaturalHistory, and Ken Hamma of the J. Paul Getty Trust for their leadership in developingMuseDoma and the .museum domain.The Museum Web community owes it to itselfand its users to stay abreast of this debate, both through the online discussions hostedby MuseDoma (http://listserv.musedoma.org/archives/musedoma-news.html) andthrough membership in that organization itself,when that becomes possible.Our domainwill only be as good as we make it!

User Influenced Design

Awareness of the needs and requirements of users was one of the main themes ofMW2001.The first four papers in this volume explore means and methods of involvingusers of museum Web sites in their conception, design and evaluation, and offer ideasof how we can integrate the needs of all kinds of users in to the Web presence of ourmuseums.

Carrie Adams, Traci Cole, Christina DePaolo and Susan Edwards of the Seattle ArtMuseum lead the volume with an essay on "Bringing the Curatorial Process to theWeb".Working with sixth-grade students from the Seattle area, their My Art GalleryProject used the curatorial metaphor to help students understand and appreciate theintricacies of museum operation. An engaging guide, Mona, personalizes the site.Designed in an iterative process with direct input from users, Mona, and the applicationshe unveils, set the tone desired by the target audience.

Other ways of finding out what users want are equally valid,and can lead to developmentof appropriate applications that surprise us with their ground rules. Research conductedat the IBM T.J.Watson Research Center by John Vergo and his team over ten intensivemonths revealed what user-centered design might dictate for a model museum Website. Using interviews, questionnaires, and surveys, they identified a preference in ageneralist audience for "less interactive narratively structured streaming experienceswith a human voice" as a way to for many users to augment their cultural experiencesand attendance at cultural institutions. Is it distressing to find that users want enhancedTV like experiences? Or is it a challenge to convert them with more interactive offerings?Or is the even greater challenge for us to change the character of what we often wantto deliver?

We are also c6rged to reach out to disabled users whom we may initially think areunable or unwilling to visit the virtual museum. Enabling access for a broad range ofusers is the theme of Susan Anable and Adam Alonzo's paper.They point out how easilymuseums can accommodate the needs of people with disabilities, with a small amountof planning and forethought. Not surprisingly, the design discipline required makes theresulting Web site better for all users.

Developing the skills to conduct rigorous user studies of museum Web activity is achallenge for many museum professionals. Ilse Harms and Werner Schweibenz, of theUniversity of Saarland in Germany report on a research project that applied methodsof usability engineering to a museum Web site. Their methods, built on foundationsconstructed in usability engineering and systematic qualitative human-computerinteraction research, offer a model to which many museums will aspire.The partnershipbetween computer scientists and museums is also a model to keep in mind.

12

4 ©Archives & Museum Informatics, 2001

ESTCOPYAVAILABLE-


Designing for (re-) Use

Effective and efficient information design remains a challenge as we struggle with thecomplex nature of cultural information. The second group of papers in this volumereports on projects, methods and tools to help us create information architecturesthat support the design of virtual exhibitions and the re-use and representation oftheir informational components.

The ViEx system developed by Christian Breiteneder, Hubert Platzer and Martin Hitzin Austria is one such tool to assist in the authoring and management of a virtualexhibition. Their framework of content repository, editorial tools, and Web pagegenerator supported the creation of the complex exhibition Bhutan Fortress of theGods (http://www.bhutan.at) and is being generalized and tested in other applications.The approach beginning at the end of a traditional multimedia design and reverseengineering a toolset that can make the same end product is definitely worth furtherstudy and application.Detail sof the content and presentation choices made in presentingthe Bhutan experience can also be found in a complementary paper on the accompanying

CD-ROM.

Gilles Falquet, Jacques Guyot, Luka Nerima and Seongbin Park also address issues ofreusability of museum information in their paper on the "Design and Analysis ofVirtualMuseums". Using a database publishing tool called Lazy, this group demonstrates theefficiencies in design that result from a rigorous analysis of structure, and the refinementof that structure using grammatical formalisms. If the building blocks are sufficientlysolid and modular, constructing new multimedia content packages could indeed besomething even the 'Lazy' could do.

Slavko Milekic explores the implications of content repurposing on interface design inhis paper, which builds on his experience in creating informative and engaging userinterfaces. Milekic proposes a typology for content repurposing, and points towardssome of the benefits that intelligent repurposing could offer. Adaptive delivery ofinformation content tied to user interest or ability, and effective redesign of informationpresentation, are two such areas where further research and exploration is required,particularly in the realm of converging media.

New Content, New Uses

Museums house and deliver a wide range of information content in many media. Muchof the deep and exciting content of museums is, of course, served from databasesrather than presented as static HTML pages.As a consequence, even the informationthat museums would like to make widely available to the public is hidden in the "DeepWeb" (Bright Planet, 2001 , see: http://www.completeplanet.com/Tutorials/DeepWeb/index.asp) where search engines cannot tread and most users never go.

John Perkins reports on the CIMI Consortium's work with the Open Archives Initiativeas one way that museum can surface their deep content, and make it available tometadata servers that support information discovery on the Web. CIMI members areworking with other partners in the international metadata community to developmethods for museum and other deep content servers to place metadata in harvestablepackages for easy collection and updating of metadata servers (or search engines, asthey are more popularly known).

13@Archives & Museum Informatics, 2001 5

BEST COPY AVAILABLE

Bearman &Trant,The Web as a Fact of Life

The Cleveland Museum of Art has taken on another of the challenges of movingmuseums out to interested publics in trying to create the partnerships and methodsneeded to launch a broadband network for the delivery of museum content to seniorsin assisted living residence facilities, community centers and disabled persons in theirown homes. Len Steinbach reports on this unique collaboration, funded by theTechnology Opportunity Program of the US Department of Commerce, that bringstogether museum and other cultural organizations' content expertise,the technologicalexpertise of local ISPs and public broadcasters, and a network of community partners.These diverse players are working together to bring life-enhancing programming tounder-served, and ready audiences and it seems that they are having fun while theyare doing it. (Maybe it's just too early in the project...)

Jim Spadaccini reflects on his experience in delivering streaming audio and video in anumber of museum contexts, and offers some thoughts on the limitations of currenttechnologies and implementations. After reviewing the relative experiences andevaluations of participants 'on the museum floor' and on-line, he designed modificationsin theWebcast viewing environment that would greatly enhance the viewers experience

moving them from a passive recipient of pushed content to an active participant in amulti-media dialogue. Spadaccini's thought experiment cum requirement statementshould inform the next generation of streaming content viewers the 'killer app' ofthe users John Vergo and his colleagues at IBM interviewed.

Personal Web Experiences

Methods that enable visitors to museum Web sites to have active and relevantexperiences have been the goal of museum Web designers since we began to takeadvantage of theWeb's dynamic capabilities.The next group of papers explores methodsand means for enabling the personalization of the museum Web experience.

Timoty Barbieri and Paolo Paolini continue their exploration of the nature of on-linecollaboration in their paper on "Cooperation Metaphors forVirtual Museums". Buildingon their experience in developing the Web-Talk environments presented at previousMW conferences, they systematize and enumerate the requirements for virtual groupbehaviorThis analysis offers a strong guidance to the designers of collaborative systems,and articulates the benefits of a group experience over a solitary visit. By making itclear how much social experience depends on shared time and space, Barbieri createsan exciting conundrum for virtual experience designers; they need to create virtualworlds in which groups of people have the same grounding they feel in the real world!

In their discussion of the "HyperMuseumTheme Generator System" Peter Stuer, RobertMeersman, and Steven De Bruyne explore another way for virtual museum visitors togain a personalized take on museum multimedia content in this case, it's a Take Away.Their prototype system enables the creation and extraction of 'personalized themes'from the HyperMuseum Server, and to create a their own view of that information, ina chosen information display and/or manipulation environment.

Providing Personalized Assistance is the goal of the designers of the SAGRES VirtualMuseum. Brazilians Ana Carolina Bertoletti, Marcia Cristina Moraes and Antonio Carlosda Rocha Costa have developed multiple paths through their virtual collection forVisitors,Teachers and Students. Each type of visitor also has access to a software agentor Personal Assistant, represented as an animated character.Their evaluation, shows

146 © Archives & Museum Informatics, 2001


that users are satisfied with this kind of on-line help, and would choose to have theassistant join them on their next visit.

Customizing the environment for experts working with multimedia archives is thegoal of the OPALES project, reported here by Henri Betaille, Marc Nanard,and JocelyneNanard. Working with the French Institut National de l'Audiovisuel this group hasdefined authoring and reading points of view, which enable categories of annotationand viewing of specified information content within personalized workspaces. Theyare also defining a collaborative work environment which enables geographically andtemporally separated experts to work together annotating documents and videos.

Being a Museum, Digitally

Since its inception in 1997, Museums and theWeb has explored the impact of networktechnology on museums and museum programs. Creative use of theWeb has expandedand enhanced museums' missions.

In a very few years, theWeb has evolved a number of distinctive genres. Kevin Sumptionof the Powerhouse Museum and Australian Museums Online takes a broad look at themodels and metaphors informing presentation of content by museum Web sites intheir delivery of Web-based Education. He presents a typology of learning styles andprovides examples of museum Web sites that have exploited each of these approaches.While Sumption concludes that further evaluation of museum learning is required todevelop consensus around when and how particular on-line educational methods areeffective, his initial typologies provide much for designers to use and students of thisevolving communications medium to ponder.

Sarah Kenderdine, also of the Powerhouse Museum in Australia, profiles an innatelyvirtual exhibition created to celebrate the 2000 Olympic Games in Australia. Shehighlights the active interpretive choices involved in making digital archaeologicalreconstructions and reviews the components of ;:his digital museum that integratesmany kinds of media in a compelling manner.

C. Olivia Frost explores the changing nature of authorship through the model of thecollaborative on-line exhibitions developed by the CHICO project at the University ofMichigan. Uniting content specialists, educators, and information specialists in cross-disciplinary teams she explores changing professional roles and the nature of creationin a collaborative environment.

The"Unseen Discussions" that took place in conjunction with the PS 1 ContemporaryArt Center Exhibition Greater New York: New Art in New York Now, changed thenature of this museum exhibition from a one-way delivery to a two-way dialog.AnthonyHuberman outlines the E-Mail Project, created as a vehicle to enable directcommunication between contemporary artists and the people who viewed their worksin this show. By giving each artist an email address and distributing these widely (includingon the exhibition labels), PS I enabled museum-goers and other critical publics toconnect directly with the creators of challenging works of contemporary art.

Conceiving of theWeb as the museum has been the logical next step for contemporaryartists and curators alike. Beryl Graham and Sarah Cook of the UK have explored theissues involved in the creation and curation of this new media arts space through the


Bearman &Trant,TheWeb as a Fact of Life

CRUMB Web site, a "Curatorial Resource for Upstart Media Bliss".Artists,Archives,Audience and Aesthetics are analyzed and the authors' experience shared in thisoverview of approaches to the confluence of digital media and museum reality.

Maintaining the Magic

Susan Hazan sends us off with a warning not to disregard the importance of theethereal in aesthetic experiences.Following the requisite tour of Malraux and Benjamin,she challenges us conceive of the virtual museum as one with all of the magic andmystery of other museum spaces.Using examples from virtual (and not so real) museumsaround the world, her exploration reminds us of the emotive and evocative power ofthe arts, and the seduction of the truly beautiful.We mustn't lose sight of that whichwe cannot completely grasp, and we must strive to maintain the mystery of the museumexperience.

In this selection of papers from the conference, authors from ten of the more thanthirty-five countries represented at the meeting present a taste of the research onwhat is rapidly becoming a world-wide phenomenon.The creativity reflected in thesepapers, and in the more than 120 other presentations made at the conference (anddocumented on the enclosed CD-rom) demonstrate again that Museums and theWeboffer a fruitful conjunction of content and a need to interpret and communicate, witha technology that thrives on rich resources and making connections possible.


plIIIPIII

0111011101

UserInfluenced

Design

1 7

Bringing the Curatorial Process to the Web

Carrie Adams,Traci Cole, Christina De Paolo, Susan Edwards,Seattle Art Museum, USA

Abstract

The Seattle Art Museum's My Art Gallery Web site was born out of an actual experiment in the galleries of the museumitself.The project was called "Growing Up With Art" and was funded by a four-year grant from the Pew CharitableTrusts. SAM invited sixth-grade classes from local schools into the museum to curate two exhibitions using themuseum's permanent collection. It was a project requiring collaboration from many corners of the institutioncurators,educators, and registrarsas well as teachers and students in the public schools.Achieving an actual exhibition in themuseum's galleries after only a ten-week lesson was an ambitious goal. Educators and a curatorial associate developedlessons for the students that would take them step by step through a "curatorial process." This process was designedto incorporate key elements of a curator's exhibition-development process and to tie into a sixth-grade curriculum aswell as Washington State learning objectives.

The very nature of this project was clearly experimental and brought up many issues within the museum about thedisplay of art and the role of the museum in relation to its community. Because of the intense time commitment andexpense required for the project, the museum was able to give this experience to only a limited group of sixth-gradestudents. Not wanting to retire our efforts to the archives, we turned to the Internet as a way to extend a version ofthe unique experience to a wider audience.

As the project was winding down, SAM had completed the transfer of its object collection data to a new system,allowing the data to be accessed through the Web. Using the Web to engage our audience with the permanentcollection became a new possibility. The idea was to build a Web site where students could be introduced to theelements of the curatorial process online:a process of visual analysis that includes observation, questioning, research,comparison, and label writing. Students, specifically in grades 6-10, could create a virtual exhibition by choosing fromthe group of works used in the actual student-curated exhibition and writing the results of their research and observationsinto a database-driven notebook.The notes and images would then dynamically generate html pages displaying theirwork.As a reward, students could pick their own gallery backdrops for their exhibition and send "gallery opening"emails to friends and family.

This paper will explore the transformation of classroom curriculum into an engaging,Web-friendly, interactive experience,pointing out the Web-management and structural-design challenges that were faced to achieve this goal. Similar to theprogram that resulted in the physical exhibitions, theWeb site project was ambitious; it required museum staff to workcollaboratively and face new issues that were pushed to the forefront by this medium.

Introduction

Using theWeb to teach the complexities of curatingan art museum exhibition, with the goal of havingthe user complete the process in twenty minutes,is no small featYet perhaps giving the user a senseof the complexities involved, from research and vi-sual analysis to design and writing, is possible.TheSeattle Art Museum developed an interactive sitethat allows users to interact with the museum'spermanent collection in a new way and create theirown online exhibition, thereby beginning to graspthe knowledge and experience required of today'smuseum curators.

My Art Gallery is aWeb site that was developed outof a four-year grant entitled Growing up with Art (seeAppendixA for granting information), which brought

local sixth-graders into the Seattle Art Museum toco-curate two actual exhibitions using objects fromour permanent collection. In ten one-totwo-hourlessons in their classrooms, museum staff taught thesixth-graders a simplified version of the curatorialprocess. It was an ambitious pedagogical programfor the sixth grade, introducing the students to vi-sual analysis, research, comparison, developing in-terpretations, and writing explanatory labels, allwithin a period of six months.A central goal of theproject was to find new ways to present themuseum's permanent collection and improve howthe public interacts with it.Another important goalwas to foster collaboration both within the museumand without, working with museum departments,teachers, and students.

BEST COPY AVAiLABLE

18Museums and the Web 2001: Proceedings

Archives & Museum Informatics, 2001, p. 1 I

Adams etal., Bringing the Curatorial Process to the Web

After the exhibitions were installed, the questionremained of how to maintain the relationshipsforged between the public schools, the teachers,the students,and the museum,and how to use whatwas learned from the project to help future usersengage with the permanent collection.These exhi-bitions were intense administrative feats made pos-sible by a large grant and will not likely becomepart of the museum's regular program due to lim-ited funding, time, and resources. In addition, thoughthe project was a great experience for about onehundred students, we wanted to find a way to ex-pand our reach.Thus, three years into the four-yeargrant,the curatorial associate involved in the projectapproached the Web team with an idea to transferthe curatorial process that the students underwentin the classroom on to the Web, allowing the les-sons developed for the project to live beyond thegrant period. It was an ambitious ideato create ahighly interactive learning environment on theWebmodeled on ten classroom lessons.The site wouldgive any users a glimpse into the world of an artmuseum curator and give them their own virtual"gallery" where they would choose a work of art,research it, and develop their own interpretations.

Many of the goals for the exhibitions were carriedthrough to the Web site, though they sometimesmanifested themselves in new incarnations. Fromthe beginning, the project to create student exhibi-tions was envisioned as a cross-cultural one becausethe Seattle Art Museum's collection includes worksof art from North and South America, Europe,Af-rica, and Asia, with particular strengths in Chinese,Japanese, African, and Northwest Coast NativeAmerican art. Development of the Web site alsocorresponded with the launch of our online collec-tion database, which is accessible through the mu-seum Web site.MyArt Gallery would provide a struc-tured environment in which a user of the SeattleArt Museum site could explore a portion of thepermanent collection online.

In addition, the cross-cultural comparisons andthemes chosen for the exhibitions came from thestudents' own impressions of the world aroundthem. We hoped this approach to the exhibitionsand the Web site would help to make a connectionbetween our collection and our users' personalexperiences. Finally, theWeb site allowed us to con-tinue collaboration and outreach efforts into the

future by marketing the site to teachers and allow-ing users to experiment with our permanent col-lection.Theoretically, the expense and risk involvedin pursuing such experimental projects in the physi-cal space of the museum is reduced on the Web.The assumption was that the Web is more flexibleand forgiving, and therefore a perfect place for theeclectic and experimental nature of these exhibi-tions.

The process of developing this site was complicated,especially for a Web team of three developers(graphic designer, programmer, and manager) andone content provider (curatorial associate).Whatmade this project such a challenge was transform-ing a model for teaching students in a structuredclassroom setting into a free-for-all Web environ-ment.The existing process for the museum's Website development presented a great stumbling blockfor this type of projectwe had simply never at-tempted such an interactive, process-oriented Website before. We knew we needed to build a sitewhere graphics, images, and rollovers were used tocreate an interactive learning experience that wouldengage our users and take them through the pro-cess. Users needed to learn on the Web from do-ing, and the site needed to give them all the toolsnecessary to do this without supervision.

The idea of developing a"notebook," a place wherestudents would do their own work and start creat-ing their own exhibition, was essential to early sitedevelopment.A vehicle was needed for students toselect art images from the collection and write notesabout their selections.A database had to be devel-oped where "notebook" images and text could besaved and retrieved to create the content for theindividual online exhibitions. It needed to be both aprogramming tool and a learning tool, giving usersownership of their process.

One early realization was that if we didn't engageour users immediately in the learning process, wewould lose them.We also couldn't assume studentswould know who a curator is or what one does. Sothe first task was to introduce the idea of a curatorin a fun and interesting way. Once we knew how toaccomplish that, we could roll up our sleeves andbuild the remainder of the site.

B19 ESTCOPYAVAILABLE

I 2 Ardlives & Museum Informatics, 2001

User navigates throughlessons In notebook torn

tobe. User has to havecompleted the learning

sateen of Me site toreceive the notebook

From notebook, user canchoose new artwork, mite

about It & retum to theirexharit where the art workwill be up dated in virtual

Callen/ almae


paler I -st ,tht,

sx, 11 metaliereserai

<weber OM,

1

artwork here

P. Pleatielheats earbeenta

Initial notebook concept

Some of the questions we needed to answer in-cluded, How do you get students to think aboutlooking at art online? How can we get them to leavethe site to research the work of art they chose andthen come back to the site? What happens if theygo to a search engine and can't find any informa-tion? Finally, how do we make them feel rewardedfor completing the process of building their ownexhibition?

It took several months of meeting, developing flowcharts, drafting text, and building models to comeup with a design solution to the "questions" thesite posed.The key challenge was coming up withan interface design that combined all the compo-nents into one dynamic learning experience. Every-thing on the site had to work together.We neededto simplify a very complex set of ideas.We knewthis thorough development process was necessaryto avoid the pitfalls before uspossibly building asite where the components competed with eachother, lacked focus, and were confusing to the user.

Throughout the months of the production process,we constantly had to redefine and lock in our es-tablished goals.This was necessary for two reasons.First, the site kept evolving and changing as weworked; we needed to produce one section of thesite in order to understand how to build the next.We weren't always sure of the end result, so revis-iting our goals kept us focused and on the rightpath. Second, each person working on the projectcontributed distinct expertise and brought a par-ticular perspective to the project.This dynamic in-tensified when staff members from several differ-ent departments were pulled into the project at

nen a lot ot the =elaternotoe on the internat. haatwere bleak and chi., and wereabout children tn theale/rata= It Stead the

Final notebook

various stages of development. It took a lot of ne-gotiating and working together to pool the neces-sary skills into a shared vision for the site. Unlike atraditional museum project with a clear beginning,middle, and end and clearly defined staff roles, thesite development was much more dynamic and cy-clical.

Over a year and a half, we worked through a pro-cess of constant redesign and rewrites of text andcode.We developed new models for management,as well as site structure, design, and content. Theevolution of our process grew out of the lesson wewere learning about the Web: that it is a uniqueenvironment where complex concepts can becomelucid in a way that is not possible in the classroomor in the gallery.

Simple Structures for a ComplexProject

As various museum staff members contributed anddeveloped content, the Web team found it neces-sary to establish content-organization tools whichwould be used both to guide the project along in afocused fashion and to guide the user along withintheWeb site itself.A main challenge was to restruc-ture the original goals and materials for the actualexhibition process into content suitable for theWebenvironment. In order to simplify both the manage-ment of the project and the structure of the site,the Web team developed a clear mission statementand defined the target audience.Then, the graphicdesigner developed a storyboard, using it as a toolto create one interface to integrate all the compo-nents of the site: the introduction, the curatorial

@Archives & Museum Informatic2

BEST COPY AVAILABLE

1 3


lessons, the notebook idea, and the ability for usersto create their own exhibition site. Crucial to thisstoryboard process was the development of an il-lustrated character (and eventually the developmentof two characters) to guide and motivate the userthrough the site.

A Focused Mission StatementCreates a Focused Web Experience

Defining a target audience is key in the early stagesof developing a Web site, as it can often lead to thesuccess of a site. It may be necessary to elaborateon the definition of the target audience and outlinethe kind of experience they should have.This canbe called a "mission statement," where the targetaudience demographics and goals for their experi-ence are melded into a clear statement.

Choosing a target audience for the site became anessential tool for focusing the content during earlydevelopment of the site. Not only did we need agood solid understanding of the audience, but wealso needed to understand how this group uses theWeb.The target audience for theWeb site differs inmany ways from the students who participated inthe physical exhibition project. One difference isthat Web users can abandon the process any timethey want by clicking out of the site.Another differ-ence is that they can't ask questions or have directdialogue with a teacher if they have difficulty under-standing a step in the curatorial process.Within theWeb site, the users are on their own and in controlof their experience.

These limitations challenged us to further stream-line the content and take several approaches thatdeparted from the classroom experience. Initially,we assumed our target audience would be the samedemographic as the students, with the exceptionthat the target age would be expanded, focusing ongrades six through ten. As we proceeded furtherinto the Web project, it became clear that, alongwith retrofitting the content, we would need tocreate a focused "mission statement," outlining thekind of experience we wanted the user to have.

For example, we decided that the Web experienceneeded to be completed within a shorter period oftime than originally planned.We initially thought ofthe notebook as a tool the users could save their

work in and then log back into if they wanted totake a break. They could log in upon return andcontinue where they had left off in the process. Onfurther thought, we realized that this kind of open-ended learning experience could result in confu-sion. Allowing the user the freedom to break theflow of learning could come at the cost of losingthem all together.As a result, we decided to createa two-part Web experience.

The first portion of the site would be dedicated tointroducing the users to the curatorial process, hav-ing them participate in five "lessons" as they wentalong. On the Web the five lessons are called learnfrom looking, asking questions, research, compare,and interpret. Only after the learning occurred andtheir online exhibit was posted did we give themfree rein to work within their notebook and chooseadditional works of art.The mission statement wasour saviour in many regards, as it helped us clarifythe user goals and focus on creating a more orga-nized Web experience.We would often refer to themission statement as a focal point as we continuedto develop the site.

Storyboards: Streamlining Contentinto aWeb Experience Conducive toLearning

Storyboarding can be helpful if there is a lot of con-tent to organize but design is not yet developed.The storyboards we created helped to approximatethe number of pages the content would requireand how much text and graphics might go on eachpage. Simply presenting a masked-out visual repre-sentation of each page helped the team easily graspwhether our ideas and content were working.

Originally, content for My Art Gallery came from vari-ous worksheets, diagrams, and notes from educa-tors who participated in the physical experience.This content was useful in the physical experience,as it promoted two-way interaction with the stu-dents and educators.The lesson tools could be usedto explain complex concepts and generate discus-sion.Working with this content, however, the Webteam found the Web site was developing into anunfocused experience for our users. Based on theoriginal content, many of the five lesson compo-nents repeated concepts, provided too many op-tions,and offered intricate cross-references to other


WT.

aorgazD-

rO. wane

V11IIM/AINe.........

CZLIZotlff7=NW.

to-KWEGIMM, V.EireelairumW

.`-`

alcaors

&T.11Z=

w.wo

111


Wog88001w

111:1-1

2

OFIrr1E2 .k rt7-

=ww...MOLOOMMOr

WOOwwfoo.10141

a:10,rAws

WINSIMP.

4. WO....

uI 000

art ..... ono ww. m*,

:Cr

aoarw

UM:011r

INWWWwww

Storyboard

lesson components. We realized that the originalcontent and ideas generated for the site were im-possible to fit into a restricted Web environment.

As a result we prioritized content, focusing on thematerial crucial to understanding the curatorial pro-cess. Content was streamlined into a straightfor-ward learning approach that would work on theWeb. Other ideas and components that were in-teresting but did not fit into this approach werediscarded.The storyboard we developed illustratedto all non-Web team staff how the original contentneeded to be streamlined for the web. It gave ev-eryone working on the project a visual guide thatbecame the framework from which all content andideas were generated.

Mona,A Hip 'Learning Tool' for Kids

During the first project meetings we brainstormedthe idea of having a guide to take users through thelessons. The graphic designer developed an illus-

Olk., '\

Mona

trated character who would act as an anchor pointwithin the site, add personal interest, and work withthe content in a way that our target audience couldrelate to.

We felt it was important that the character be acompelling peer who goes through the curatorialprocess on a level similar to our users, but alsoexplains the concepts and lessons as they proceedthrough the site.The character developed was firstcalled "Annie Art" but she was eventually christened"Mona:' Used as a main tool for organizing the con-tent, she became the backbone around which weworked the other elements.

Through her positioning, and the use of dialogue(called "bubble text") and rollovers, Mona operatedinteractively with the users, providing a context forthe content and accommodating different learningstyles. Some users have found they read Mona'sbubble text while others find they learn better re-lying solely on the visuals.

22

Now ifs YOUR turn to pick yourwork of' art. What color and shape

appeal to you?

.Mona with bubble text

BEST COPY AVAILABLE

© Archives & Museum Informatics, 2001 15


(fir Iteirlorpris

Pablo

In addition to Mona, we developed a second char-acter, her dog Pablo, to keep students motivated,guide them through help screens, and chart theirprogress throughout the site. Pablo became the icongiving a consistent presence to the help function.

Evolution of the Character and Script

The original text for the Web followed the courseof the lesson plans despite the fact that in the class-room the lessons were presented for one to twohours over a period of ten weeks.The resulting textwas cumbersome, repetitive, and dry. Part of theproblem was lack of experience.The curatorial as-sociate who helped write the original lessons andtaught them in the classroom was also writing thecopy for the site. Her task was formidable; she hadto write for a format she had little experience work-ing with and for an interactive framework that did

Help box icon

not yet exist. Likewise, the Web team understoodthe technical and design possibilities of the Web,but did not have the experience of creating an ex-hibition with the students, nor were they expertson the curatorial process. Clearly, the linear modelthat worked on simpler site projects"first copy, thendesign, then post" was not going to work here.

Initially, we felt that Mona would appear intermit-tently throughout theWeb site, appearing only whennecessary to add visual focus and help explain acomplex lesson. As the lesson sections were laidout graphically with the first draft of text, it becameapparent that the text and our character seemeddisjointed. In fact, in early testing, our user was notreading the text at all. This was partially a designissue, but the illustrative examples that came out ofthe classroom were detracting from the visual andinteractive nature of the Web experience.

ar-J1

learn from looking

,SAIG111.

41111e...raimE, How do you deotribe what you see when you look at a work of art?Artists use elements such as Ilea, ruler, torture, space, and shape to make art. Let me

show you. Click on the green paintbox* to go forward, or the red one to go beck!

1 2 5 5 FIN.M.1

Round I Design:Mona with no "bubble"text - lesson text at bottom

2316 © Archives & Museum Informatics, 2001 BEST COPY AVAILABLE


My Alt G Yle 3.t.04

learn frOm loOkiflg

Final Design: Lesson text integrated with Mona

During the next redesign Mona became the unify-ing structure for the site, and text was streamlinedto work with her. We added a text bubble aboveher head and drastically cut the script to fit com-fortably within it.We were able to move much ofthe longer explanatory text into Pablo's help screens,assuming that those who needed this extra infor-mation would seek it out. Finally, text that did nothave a direct visual relationship to what was goingon in that page was eliminated.

Placing all of the instructional text in Mona's bubbleintensified the interactive quality that she now em-bodied.As users rolled over the various interactives,Mona's position and bubble text changed accord-ingly.A once static page was now full of action.

With Mona and Pablo established as our tour guides,the text for the site shifted from a descriptive lec-ture into a narrative dialogue.The curatorial asso-ciate and Web manager had many two-hour writingsessions that were more like drama workshops.Thegraphic designer also rewrote much of Mona's text;perhaps because the design is so central to a site ofthis nature, she clearly understood what Monashould say.

Much of the narrative couldn't be written until thebasic design of each section of the site had beensketched out. We needed an idea of what Monawould be doing and how the user to the site would

be interacting with her before we could write care-fully focused text. This back-and-forth exchangebetween design and copy writing allowed us to makeuse of the unique abilities of the Web to illustratecomplex ideas (especially visual ones), and to stream-line the text as we went.As the programmer noted,there was one point where she realized that onerollover allowed us to get rid of a whole page oftext.

Project staff had decided to write Mona in the youthvernacular in order to appeal to the target audi-ence.We did not want her to "talk down" to theuser but be an equal, someone they could feel com-fortable with when learning about the complexi-ties of art. Internally there was some concern aboutthis language.We wondered if using words such as"dude" and "rock on" could potentially alienate thestudents and trivialize the process.We were sensi-tive to the fact that, as writers, we are much olderthan our target audience.We didn't want to assumewe knew what would work for them.We also didn'twant the site to become out of date by using trendylanguage.

Testing Our Audience

During the writing process we struggled to keepour users in mind.We realized that we were reallyworking in the dark and were plagued by the ques-tion,"Will kids really do thisrThe next step was to

24©Archives & Museum Informatics, 2001 17

research

Question?Ask!

r.unkod: liicn s 4(10:60,1,I hm Nand mown to the tallowirte tatddlent:

WIns movie...War TV shows hat . beta its?Isms cIIP ...et <IP Masa cm Med Om paps fee 1"...." ?

47:e71.,14 dst Mere etsi I tee epicure arca. Cabeitle death certificate?

at Ken Mule =tiled Sionsalee sod tar tom colelmity now 4.149 we.. C. 10 piowa otilw auk okbrityllkw

fal, Kw, Cow., OD Go year uncover ham to Expert an Etanrodianent FXP,raa,

had.artaddegeateaePre wow. 4*he.*

(011.0 C01.1a lkst.ww: Now now low sow 5Wwull7owl ...

CO 19* 0.10. inAwyo161.(belf...1eon.*

at ...it T.. CAN

SID awl ceethi memo

/dicts,trm Now look what luwiwool when I typed in "Who is Kurt alei Askloowo. Thwzlrd toiler narrow down the search!

.1 2

Round I Design: Research section with search engine screen shots

have both students and teachers review our mate-rials to let us know if we were on the right track. Itwas important to get feedback on what we werebuilding and also to get a sense of whether or notour target audience would be interested in usingthe site.

Through informal testing we evaluated the site withthree different groups of users that we knew wouldbe invaluable to our development process: teacherconsultants, students in the target age group, andmuseum staff. The graphic designer created twoquestionnaires one for the students and one forthe teachers.She also created design boards featur-ing Mona and Pablo, the site layout, and text. Wescheduled several one-hour testing sessions andrecorded them.These sessions kept us focused onthe needs of our user throughout the rest of thedevelopment process.

What we found is that each student and teacherhad contrasting opinions of theWeb, based on bothlevel of interest and access to technology resources.Regardless, each person responded with genuineinterest in learning what a curator does and in build-ing their own exhibition site on the Web. In fact,everyone seemed excited and interested in the con-cepts of the site and the direction it was going. Re-action to the character illustrations was positive, aswas reaction to the text.

18

Teachers were among the first testers of the site.They encouraged a youth-oriented dialogue and al-leviated our initial concerns about the text.This coregroup of teachers became vital consultants whosefeedback proved invaluable throughout the devel-opment of the site. Because much of the site's con-tent had to be reshaped from the original class-room lessons for the web, we also wanted to verifythat the content was still solid from a pedagogicalpoint of view.We also got their advice on the logis-tics and feasibility of using an interactive Web site ina classroom.The teachers supported our decisionto make the site a personalized process for the stu-dents, pointing out that the tools to let them picktheir own works of art and write their own noteswere important motivators.

As part of the testing process, we also intervieweda group of middle school students. We were curi-ous to see how they would respond to Mona andthe language in the script, as well as what interestthey would have in creating an art exhibition online.The graphics and text were generally well received,with requests from various students to"tone down"the use of the vernacular. We learned that theymainly use sites targeted to adults, feeling stronglythat education sites developed for their age groupwere "cookie cutter" and that the language usedwas often "too cute:' They encouraged us to cre-ate a Web site that gave them the ability to makechoices and trusted their intelligence.

25© Archives & Museum Informatics, 2001 BEST COPY AVAILABLE


MA's the, lantser Why to Klevtlesa:1trattr4 such Sa tntle thstovaptsAltar Wtteclantter*Cuttholganffltra

da err:mares.

aitaViSta:Kum Co-Na:n

InalEto°Z

Google

t:Itvatta

ateewfIALSos

lisho was Kutc Cdsalnl alt446=1..

75; rt.:0,K

Final Design: Research section with search engine logo portal

It was important to schedule constant review intothe many phases of the production process. Projectstaff , often consisting of more that the central groupof four, had long review meetings going over eachpage of the site both during the initial design andthroughout the various redesign phases.We askedother museum staff to walk through the site at sev-eral key points during production.These feedbacksessions were instrumental in fine-tuning the flowand troubleshooting technical problems.

Redesigning the Curatorial ProcessOnline

A good illustration of the development process forMy Art Gallery is the evolution of the research les-son design, which was the most challenging to build.This section needed redesigning three times, man-aged to dodge several deadlines, and caused wide-spread panic among the Web team. The problemwith this section was threefold: I .) We invited ourusers to leave our site to conduct research; 2.) Wecouldn't assume they knew how to do this; and3.)We had to acknowledge that researching worksof art could be an ambiguous and open-ended pro-cess.

When we decided to encourage users to leave thesite and use search engines, we feared we wouldlose them. We worried that there would not be

enough incentive for them to return to our siteeven if we used all the limited programming toolsavailable to make coming back effortless. As a re-sult of our concern, Pablo was engineered to ea-gerly await their return to the site.

Another concern was how to explain the complexprocess of conducting research online. Similar tohow we first approached the original text,we startedwith classroom technique and tried to model theprocess by example.We took screen shots of somepopular search engine sites that Mona "used" toresearch the work of art she chose, hoping visualswould effectively explain how to use search enginesto some one who may have never done it. How-ever, we had problems securing permission fromseveral of the search engines to capture screenshots, so we had to abandon this first design con-cept.

This turned out to be a mixed blessing. We weretrying to balance the need to guide novice research-ers while providing a more streamlined presenta-tion that would not bore those with more onlineexperience. In the end, we broke with the narrativescript and replaced it with a central research por-tal.

Using various search engine logos, users can link toseveral examples of search results and explore them

ESTCOPYAVAILABLE26

©Archives & Museum Informatics, 2001 19


in detail if they wish, before moving on to do theirown research.This simple design provides a neededstructure without being too didactic.

Research is a vital step to the curatorial process,including understanding that there isn't necessarilyone right answer to a question raised. Sometimesfinding an answer at all can be challenging, especiallyfor some of our collection's works that the stu-dents can chose from.To prepare for this eventual-ity, we used Mona to explain that there can be manydifferent opinions about one work of art, and theyare not necessarily right or wrong. She lets the userknow that it is ok if they can't find specific informa-tion about a chosen work of art. Mona selects aphotograph as an example, but fails to find explicitdetails about it. However, after more research, shefinds information about the photographer and thesubject of the photo, allowing her to hypothesizewhat the answer to her question might be.

Realistically, our concerns about this section couldnot be dispelled easily since the actual process ofresearch is, in and of itself, inconclusive at times.There was also little we could do other than hopethe site was compelling enough to draw users backin to finish their exhibit.

Use

During our testing phase we asked the teacherswhat they thought the best way to market the Website to Seattle-area educators would be.Each teachersaid an attractive postcard would stand out amongthe multitude of solicitations they receive. Our Webdesigner produced a postcard that was mailed inconjunction with the launch of the site in Septem-ber 2000.We used our own mailing list of PugetSound-area schools and education-related institu-tions that represented our target audience. Ourpress department also sent out a release that gen-erated national coverage in education and art-re-lated publications.

The postcard mailing was very successful in reach-ing a local audience, and several area teachers haveintegrated the site into their curriculum. Someteachers have contacted us to discuss the site andprovided informative feedback. Most reviews havebeen favorable, but we ran into an unforeseen road-block: notebook identification. Following protocol

20

adopted by other Web sites, we had chosen theusers' email addresses as the unique identifier forthe notebooks.This identifier has the advantage ofallowing us to email forgotten passwords to usersimportant in ensuring continued use of the site.However, we learned that many students do nothave their own email addresses. In one instance, aresourceful teacher allowed her students to use heremail address; unfortunately, this meant that all thestudents were actually writing to the same note-book, resulting in quite a mess. In other instances,the students mistyped their email address in theinitial setup, rendering their notebook inaccessiblewhen returning to the site. The programmer andWeb manager worked with affected teachers tocome up with short-term solutions so their stu-dents could continue to work on the site.

Many of the teachers who contacted us did so spe-cifically to discuss technical difficulties, but this alsogave us the opportunity to interview them to un-derstand how the site was being used. It is interest-ing to highlight one teacher's experience with thesite. She had an ongoing discussion with her stu-dents about attitudes toward works of art. Her stu-dents argued that all artworks could be judgedequally, that merely by being art they were "all thesame:' She wanted to find a way to prove to themthat works of art are held up to standards and thateach work of art had to be understood within itsown context. After receiving our postcard, shethought the MyArt GalleryWeb site would be a goodtool to use to prove her pointAlong with using thesite, she also required the students to hand in apaper. She did feel that the students' assumptionsabout art were being tested as a result of using thesite.Through their exploration of the five interac-tive lessons, they were discovering the complexi-ties involved in researching art.As a result, they hadmany questions about the works they chose, andsome students came to her for guidance when theycould not find adequate research to explain them.This feedback supported our initial fears about de-veloping the research section. However, with thisteacher's guidance, the students did not give up andlearned a valuable lesson about the process of re-search itself and also the intricacies of researchingtwo very different works of art

Another indication of use is the online exhibitionsthat are created after the users go through the les-

27©Archives & Museum Informatics, 2001


sons.As of this publication, 254 users have loggedon and 64 exhibitions have been created. Many ofthe exhibitions submitted are similar; the majorityinclude two works of art and one paragraph de-scribing and/or comparing the works.This may indi-cate that our user expectations were too high, sincewe were hoping the submissions would be moreextensive and dissimilar.We knew from the begin-ning that our goals were ambitious, expecting usersto go through the five lessons to launch their onlineexhibition.After going through all of that, it may beunrealistic to expect the students to go back totheir notebook to continue to choose, research,and write about additional works of art.

In the spring of 2001, we will evaluate the use ofthe site to determine what further enhancementsshould be made. We will review a wide range ofissues: long-term solutions for the technical prob-lems, analysis of the quality of the exhibition sub-missions, limitations of the research section, andwhether or not we should provide more informa-tion about the works of art from our collectionthat are on the site for users to select. Importantto this discussion is what amount of museum re-sources is available to further market, maintain, andenhance the site.

Conclusion

The Seattle Art Museum staff working on thisproject learned several invaluable lessons. First, welearned that the organization of a complexWeb sitehas to come from the Web site producers in theform of a storyboard, mission statement, and pro-duction timeline. Staff with Web expertise under-stand how a project idea can be best translated intoa successful online experience. Second, a level oftrust in the Web team is needed from all staff par-ticipants in order to do this. Staff members need totrust that their goals and expertise will not be lostor misinterpreted through the Web developmentprocess.This trust is built when all project partici-pants gain an understanding of how theWeb works.Also, positive working relationships among staff canbe further established if all participants on a projecthave a clear understanding of their roles and re-sponsibilities and are kept informed.Third, effectiveproject coordination and communication is difficultto implement with new initiatives such as the MyArt GalleryWeb site, but it is one of the most impor-

tant factors in its success. We encountered manystumbling blocks in this area, but came out of itwith clear communication models that we are nowusing, making currentWeb site projects much easier.

Interestingly, the Web site ignited many of the con-tentious issues raised as a result of the Growing upwith Art project. This project sparked discussionwithin the institution about the role of the curatorin a museum and whether a sixth-grader could be asubstitute for a real curator. Our efforts to high-light this experience and put it on the web, in a verypublic format, brought this issue back into the spot-light.

The management challenges of a Web project withthis level of complexity seemed daunting at timesbecause protocol was lacking within the museumto guide us through the unique issues that arose.Many of these issues demanded attention,and meet-ings were held to explore and create new guide-lines. Some of the questions raised included, Do weneed to review the online exhibitions users cre-ated? Or is it acceptable to post them automati-cally to the site? Since this was a project of com-plex collaboration across museum departments,who was ultimately responsible for it? The meet-ings convened to answer these questions resultedin innovative problem solving, but also led to morequestions that we are confident will be answeredover time. Eventually, we started to see My Art Gal-lery as one pioneering answer to meeting themuseum's important initiatives and as an excitingpilot project to lay groundwork for future Webprojects.

Acknowledgements

The Documents International exhibitions at the SeattleArt Museum are part of a four-year project calledGrowing Up with Art which has been generously fundedby The Pew Charitable Trusts; M. J. Murdock CharitableTrust; Microsoft Corporation; the National Endowmentfor the Arts with the H.J. Heinz Company Foundation;The Coca-Cola Foundation; the Discuren CharitableFoundation; the King County Arts Commission Hotel/Motel Tax Revenues; PONCHO; and the WashingtonState Arts Commission.The Corporate Council for theArts provides support for this exhibition through itsgenerous annual funding of the Seattle Art Museum.

28BEST COPY AVAILABLE


"Less Clicking, More Watching": Results from the User-Centered Design of a Multi-Institutional Web Site for

Art and Culture

John Vergo, Clare-Marie Karat, John Karat, Claudio Pinhanez, ReneeArora,Thomas Cofino, Doug Riecken, Mark Podlaseck,

IBM T.J.Watson Research Center, USA

Abstract

This paper summarizes a I 0-month long research project conducted at the IBM T.J.Watson Research Center aimed atdeveloping the design concept of a multi-institutional art and culture web site. The work followed a user-centereddesign (UCD) approach, where interaction with prototypes and feedback from potential users of the web site weresought throughout all the phases of the design process. In the first phase of the research, we conducted interviewswith museum curators, issued questionnaires to brick and mortar museum visitors, performed on-line surveys withusers of museum web sites, ran focus groups, and performed usability walk-throughs of best-of-breed museum websites and of our new web site design ideas.

The results showed a surprisingly strong interest in streaming multimedia,TV-1ike experiences where users watchedexperts and artists talk about art and culture, augmented by links to additional and in-depth information. Such a designstrategy was clearly favored by our subjects over more interactive experiences such as chat rooms, information searchsystems, or user-curated tours. In other words,when users go to a web site of arts and culture looking for entertainingand educational content, they seem to be less interested in clicking to find information than in watching people,especially experts and celebrities, present and discuss artistic works from their perspective.

Based on the results of the first phase, we developed a prototype of a web site based on the concept of "tours" 5-10 minute streaming multimedia experiences led by experts and artists, accessible by users with normal telephoneconnections (56Kbps modems).The tours resemble a slide show enhanced by narration, music, occasional video clips,and hot spots for extra information.Two pilot tours were fully produced and tested with users who reported greatsatisfaction with the tours in terms of their entertainment, engagement.and educational values.We also observed thatthe most satisfied users were those who interacted and explored the least during a tour.

The research described in this paper suggests that people interested in art and culture have an affinity for web siteswhere they can have entertaining, curated experiences in many ways similar to shortTV-1ike documentaries. Yet thereare many other occasions when they are looking for engaging educational and entertaining experiences. In suchsituations, less interactive but narratively structured streaming experiences with a human voice seem to be a veryappealing format to engage and satisfy such users. Importantly, they do not see such experiences as a substitute foractual visits to museums or attendance at live performances, but as an enriching and highly accessible way to augmentthe cultural experiences and performances they enjoy in brick and mortar cultural institutions around the world.People are often looking for specific information about artists and art works.

Introduction

In the summer of 1999,IBM decided to investigatethe possibility of sponsoring and creating a multi-institutional web site for art and culture.The goalof this web site is to give users around the worldways to increase their appreciation of artistic works,cultural treasures and performances by providinginformation and learning opportunities about artand culture.

User centered design

User-centered design (UCD) is a design approachin which multidisciplinary development teams cre-ate useful, usable and satisfying user experiences.

Central to UCD is the notion that the design pro-cess is iterative, and is accomplished by nearly con-tinual input from users. As user input is collectedand analyzed, it is fed back into the design process,resulting in progressive refinement and improvementof the system under design.

We followed six general UCD (IBM,2000) principleswhich guided our activities:

I. Set goals for the project.This early phase of theproject is also the right point to spell out any com-petitive goals, along with an identification of theprimary competitors.

'4)9Museums and theWeb 2001: Proceedings

Archives & Museum Informatics, 2001, p. 23BEST COPY AVAILABLE

Vergo etal., Results from User-Centered Design

2. Understand users. Define the intended user popu-lation and the context of use.The context of useincludes the physical context and the social con-text, if both are relevant.

3.Assess the competitiveness of the system underdesign. If the goal is to design a system that breaksnew ground, it is risky to attempt such a designwithout a thorough understanding of the com-petitive landscape. Having clearly stated goals forthe intended user allows testing of the systemagainst competitive alternatives and comparing theresults.

4. Design every aspect of the user experience. Ev-erything a user sees and touches is designed to-gether by a multidisciplinary team. In the case ofthe design of a web site for art and culture, theuser experience goes well beyond the content toinclude navigating the site, interacting with virtualexhibits, and even downloading and installingbrowser plug-ins.

5. Create and evaluate designs iteratively with rep-resentative samples of the targeted user popula-tion. User feedback is gathered early and often,and this feedback drives design and development.In the early phases of design, lower fidelity proto-types are presented to and tested with the users.As the design team's understanding of the effec-tiveness of various design alternatives and designelements is clarified, higher fidelity prototypes aredeveloped and tested.This allows inexpensive test-ing of many ideas early in the design process.Oncea design idea (or set of design ideas) tests wellwith a low fidelity prototype, it is less risky tocommit development dollars and resources inpursuit of the idea(s).

6. Manage the project through continual user ob-servation and testing.Throughout the life of theproject, continue to monitor and listen to yourtarget user group, and let the feedback inform anymodifications to the system design.

These guidelines are a set of heuristics that we typi-cally follow on any project They provide tremen-dous flexibility with respect to the precise activi-ties a project team might employ and include a widerange of established user research methods andparticipatory design methods in use today.The next

section explains the specific techniques we em-ployed on our project, along with the rationale forour choices.After each UCD activity is presented,we report some of the most significant findings re-sulting from the activity. Each activity yielded a wealthof data not presented in this paper but publicly avail-able (Karat 2001, Pinhanez 200I).Those findings thatsupport our ultimate design choices are presented,along with findings of general interest to the Muse-ums and the Web community.

Our UCD Approach

Project goals

The initial requirements for the project were ex-tremely broad and open.We set out to create onthe Web a major cultural destination (not a portal)that would attract a large number of visitors. Thesite had to provide access to content supplied by5-10 of the largest and best-known cultural institu-tions around the world.The institutions were cho-sen to represent different areas of art and culture,including various performing arts, fine arts, naturalhistory and science. Finally, each exhibit on the siteneeded to combine content from multiple institu-tions, to leverage the powerful institutional part-nerships and to ensure the uniqueness of the site.

An additional goal was that the web site must notbe a database of cultural artefacts or knowledge,but instead it must attract users by enabling enter-taining and educational experiences similar to thoseprovided by visiting a museum, attending a perfor-mance, or watching a cultural TV program.The suc-cess of the web site would be measured by its popu-larity and, specifically, by its rate of return visits. Giventhe openness of the initial specifications of theproject and the strong requirement of users' satis-faction, we adopted a user-centered-design (UCD)approach to the development of this web site.

We ran five major user research activities in thediscovery phase of the projectWe conducted in-terviews with museum curators, issued question-naires to brick and mortar museum visitors, per-formed on-line surveys with users of museum websites, ran focus groups, and performed usability walk-throughs of best-of-breed museum web sites andof our new web site design ideas. Before engagingin any of the user research activities, we took thetime to define our target user population.



Defining the target user population

Our definition of the target users for our projectwas initially shaped by the project business goals.The sponsoring organization within IBM came tous with some initial parameters.We eventually de-fined our target users to include members of sev-eral different population segments:

1.Adults 21-34 years old2.Adults 35-50 years old, with and without children3.Adults 50 - 70 years old, and4. Children 9-15 years old, typically accompanied

by adults from segment "2", above

While we did not target people 16-20 years old,we did not explicitly exclude them from our userpopulation.

We assumed that our typical users spend an aver-age of 10 or more hours a week on a computer,and of that time, five or more hours are spent onthe Internet.They find out about sites from searchengines, advertisements, friends and family, and websurfing.They think that having cultural events as apart of their lives is somewhat to very important.We also anticipated that visitors see the site as away of enriching their experience at the brick andmortar center, not as a substitute for it.

Focus groups

Our first project activity was to run a series of sixfocus group sessions,which were conducted on theeast coast (New Jersey), mid-west (Michigan) andwest coast (California). Each group consisted of eightindividuals, gender balanced. Two groups of agegroups 21-35,35-50 and 50-70 were conducted.Anindependent, objective third party moderated thesessions.

Some of the initial ideas that were explored duringthe focus group session were:

Giving users the ability to create their own ex-hibitLive virtual tours where artists or curators areavailable on-line, in real time, to conduct a tourA cultural community center allowing users tomeet, talk and exchange ideas on a wide rage ofcultural topics

3ti

A site that could be personalized to an individual'scultural interests, so that an individual visiting thesite any time sees cultural content that is morelikely to have personal interest

During the focus group sessions, participants werepresented with verbal descriptions and simple vi-sual representations of the new design ideas, andwere asked a wide range of questions with the fol-lowing objectives:

Gaining a better understanding of the target audi-ences' use of cultural resources.Uncovering consumers' experiences with exist-ing cultural web sites.Understanding current web site usage andbehaviour.Assessing reactions to the proposed web site fea-tures/activities."Brainstorm" potential features for inclusion onthe web site.Determining relevance of and interest in the website concept.Understanding the receptivity to various paymentmethods (i.e. advertising, sponsorship, subscrip-tions/memberships, etc).

Focus group results

Overall, the newly proposed web site featuressparked favorable reactions from consumers."Live"tours were overwhelmingly considered one of themost compelling and unique ideas proposed.Theiparticularly liked the opportunity to interact withartists and tour guides.

Respondents were adamant that they should notbe charged a fee to use the features offered on theweb site.Advertising banners and third-party spon-sorships were deemed the most acceptable meansof funding the site. However, it must be noted thatthe use of advertising banners was a forced choiceresponse. In the usability walkthroughs, people wereclear in their disdain for advertising banners.

Most respondents viewed the newly proposed website as a supplemental source of information and/or enhancement to the actual live cultural activityor event.While east-coast participants usually useticketing and venue web sites to gather logisticalinformation regarding a cultural activity, mid-westand west-coast respondents also venture to cul-

O Archives & Museum Informatics, 2001 25

Vergo eta, Results from User-Centered Design

tural "home sites" for background information re-lating to the activity or event. Most respondentsconduct some form of preparatory research be-fore attending a cultural activity or event.

Time,geographic limitations and financial constraintsare the primary factors that hinder respondentsfrom attending cultural events as often as they wouldlike.

Curator interviews

We conducted a series of interviews with curatorsfrom two partner institutions. Five curators wereinterviewed, with each session lasting between oneand two hours.

The interviews were conducted with a specific setof goals but did not follow a rigid script.We que-ried curators to understand the design criteria forexhibits. We were interested in understanding ifthere are differences between institutions on howdesign criteria are created and implemented. Weasked them how the design criteria affect the finalexhibit.We wanted to know if curators go throughthe explicit step of defining the target audience foreach exhibit Finally, we asked the curators how they(or their management) measure the success of theirwork.

Curator interview results

Curators at both partner institutions stated thatmost individual exhibits are attractive to narrowaudiences, and that many different types of targetaudiences exist Curators from both institutionsexpressed an interest in drawing new people (i.e.new visitor segments) to the institution. Curatorsat an institution indicated they specifically target20-30 year old visitors on a regular basis with thegoal of establishing a satisfied clientele that will en-gage in repeat visits to the institution for many de-cades to come. Both institutions consider "familieswith children" a target audience.

All institutions have "dead zones", times when visi-tation is slow, and they try to draw people in to themuseum by putting on exhibits that have appeal toa wide audience.The curators reported that theirinstitutions perceive a strong need to keep exhibits"fresh", in order to attract repeat visitors.

26

No institution employs a quantitative approach inmeasuring success other than simply counting visi-tors.The institutions use ad hoc observational tech-niques in the museums.They also gauge reactionsthrough published reviews in magazines and peri-odicals.

Usability walkthroughs

The usability walk-throughs were run in 12 sessionswith a total of 70 participants who were screenedfor cultural interest and basic familiarity with theweb. Subjects were first shown excerpts of existingweb sites related to culture. Groups of approxi-mately six subjects each were assembled in the TJWatson Usability Lab in Hawthorne, NY.As muchas possible, the groups were composed of 50%women, 50% men, and represented the diversity inincome, education, background, race, and ethnicityin the tri-state area.

Each group was shown sections of two existingcultural web sites (children were shown only onesite), followed by two new design ideas for the pro-posed site.After each demo,a facilitator led a groupdiscussion of their opinions of the content in thedemo. Then participants recorded their individualthoughts about the demo content (children had nowritten part).Then the next demo was shown.Atthe end of the adult session, participants were askedto answer several summary questions (children didnot do this part). Over the course of all thewalkthroughs, five existing web sites were shownand five new design ideas were presented.

In the first part of the session, we showed existingweb sites where users could navigate through 3Dreplicas of exhibitions, interact with panoramic im-ages, visually explore paintings and 3D sculptures,write and publish stories about the content, andobtain information about different layers of a paint-ing.

The mockups of our new design ideas shown in thesecond part of the usability walk-throughs includedfive different possibilities for exploring cultural con-tent

I.A filtering system based on direct manipulationof large databases with visual feedback (such as inAlberg & Shneiderman, 1994)

32@Archives & Museum Informatics, 2001


2.A set of lenses (tools) to manipulate the way con-tent could be viewed (such as in Stone, Fishkin,and Bier, 1994)

3.A chat system where people could talk about aparticular artwork (Viegas and Donath, 1999)

4.A notebook system where the user collects andcomments on artistic content, and later publishesthe notebook for publidprivate viewing

5. A multimedia system where the user watchesguided multimedia tours, interacting wheneverinterested in related information

Usability walkthrough results

A major finding of the usability walk-throughs wasthat most of the participants did not express inter-est in web sites that involved active interaction withthe content or other people, such as when using afiltering system, creating a notebook, or chatting.The multimedia prototype was clearly the best re-ceived among the design ideas.Among the existingweb sites, there was a preference for sites wherethe user was guided through an experience or dis-covery process,and in this case, participants stronglysuggested the replacement of text by audio.

We summarize those findings by hypothesizing thatin this kind of entertaining web experience, userswant less clicking more watching. Users seem to bevery comfortable with the idea of a streaming webexperience that leads them through artistic andcultural artefacts where, unlike television, the streamcan be paused, replayed, or interrupted for furtherexploration. In fact, we found a strong desire foravailability of additional information throughhypermedia links and in-depth analysis of the worksof art. Interestingly, some of the participants viewedthe more interactive design concepts and existingweb sites as work, not entertainment.

The idea of more passive experiences was furtherreinforced by many complaints about the difficultyof 3D interfaces, the excess of textual information,and the amount of work required to find informa-tion.The streaming multimedia concept seems toaddress this major concern by neither requiring the

user to learn and exercise a large set of new skillsnor requiring a large amount of attention and deci-sion-making.

The usability walk-through also detected a desirefor a"human voice" behind the multimedia experi-ence; that is, a personal viewpoint in the expositionof the content. However, this voice needs to havethe authority of an expert in the field or the humaninterest of a celebrity. In fact, a major complaintidentified with chat systems was precisely the lackof trust and perceived value in the information pro-vided by generic chatters.

We explicitly asked the usability walkthrough par-ticipants what factors would be important in theirdecision to re-visit the site on multiple occasions.The most frequent responses to this question were:

The quality of the contentThe ease of use of the siteHaving a cultural calendar that would send themremindersThe quality of the images, video and audioHaving new exhibitsAccess to archived old exhibits.

The modal response to the question "How oftenshould the site be updated with new material" wasone month, but there was considerable variance inthe data.

Web site surveys

The web surveys were conducted by using a ques-tionnaire that automatically popped up, in a sepa-rate window, whenever users accessed the homepage of two partner art museums. Answering the17 questions was optional, but throughout themonth the survey was conducted, 830 and 1417users answered the survey on each of the museumweb sites.The survey was centered on questionsabout how the respondents use the web to obtainart and culture-related information and entertain-ment.

Web site survey results

About 81% of the surveyed visitors were visitingthe web sites for the first time.About half of them

© Archives & Museurrarffirmatics, 2001 27

Vergo etal., Results from User-Centered Design

(45%) got to the site through search engines and17% through links from related sites.An estimated10% came directly from other web pages.

About 25% were researching information on spe-cific content, while 32% were interested in specificinformation about the museum (schedules, access,etc.).About 12% were looking for fun-like activitiesand only 2.4% declared interest specifically in shop-ping.

The absolute majority of the visitors (90%) werealone while visiting the sites. The most importantaspects in the visit were the "content" (46%) andthe ability to pre-visit the museum (31%). Only 14%ranked "usability" as their major concern.

About half of the visitors considered the sites "in-teresting",21% saw them as"educational", and only4% thought of the sites as "entertaining". This, insome ways, correlates with the objectives of thevisitors as described above.

Visitors considered actually getting the informationthey wanted as the most engaging part of the expe-rience of the site (39%). Similarly, when asked whatpart of the experience could not be replicated,50%of the visitors mentioned the facility of getting in-formation, and 35% mentioned the convenience ofvisiting the museum from home. The biggest frus-trations with the web sites were

The inaccessibility of the full content of the mu-seum (35%)Not finding the sought-for information (16%)The inability to access the schedule (9%)The inability to ask questions of the curators (5%)

There was a slight predominance of males (58%)among visitors to the site.The age distribution wasquite homogeneous, with a strong participation ofpeople in their forties (22%) and fifties (11%).

Most of the visitors (58%) belonged to families hav-ing more than 5 hours/week of web usage and only22% used the web less than 2 hours/week.As forconnection time, 76% had a connection equal to, orfaster than, 56Kbps, and 99% have a 28.8Kbps orbetter connection available.

28

To the users who answered the survey, the mostpopular education web sites were Discovery (6%),Yahoo, PBS,and the NewYorkTimes.The most popu-lar entertainment web sites were Yahoo (6%), theInternet movie database, ESPN, and the onion.

Brick and mortar visitor surveys

A total of 99 visitor surveys were conducted attwo of our partner institutions.The surveys at eachinstitution were conducted on multiple days in or-der to get information from weekend and weekdayvisitors, as well as to survey individuals who mightbe visiting to attend a special event at the institu-tion.

Brick and mortar visitor survey results

People enjoy integrating cultural experiences intotheir activities in order to create a rewarding so-cial, entertainment, and educational experience -also,visitors to the institution are interested in learn-ing and enriching themselves and their friends andfamilies.

Many people selected the institution they were vis-iting because it provided an event that would sat-isfy both social and educational goals.A visit couldalso be integrated with other social events (visitorsappeared to schedule a trip to the institution alongwith other activities such as a walk in the park, shop-ping, and dining). A majority of visitors visited toenrich themselves by viewing exhibits and specialevents. Many visitors indicated that the use of dif-ferent presentation techniques (e.g., kiosks, films,human guides, lighting effects, etc.) greatly enhancedthe learning experience.

The nature of the content dictates whether peopleare more focused on entertainment than educa-tion, and is primarily responsible for shaping theexperience.As for the design of the web site, thesesurveys point out that there should be differenttreatments and different goals for different kinds ofcontent. For instance, scientific content asks for anexperience designed to be more educational thanentertaining, while content related to comedy anddrama should be entertaining and emotionally en-gaging.

3 4



There is a significant difference between the nor-mal visitors and the audience for special events.Special events are less "social" than regular visits. Interms of the design of the site, this data should in-form the number of community-related features (forinstance, chatting) available for regular browsing visi-tors versus the audience for special events.

Conclusions of the Discovery PhaseUCD Activities

From the responses of the web surveys, we foundthat only a minority of people (4 to 5%) label theexperience they have at museum web sites as "en-tertaining". In many ways, the participants in ourresearch seem to lean towards defining an enter-taining web experience as something closer to tra-ditional TV, but enriched by the opportunity to ex-plore and find extra information. However, our re-search also indicated that to access the majority ofour targeted audience today, such web experiencesmust be available to users sporting modems of56Kbits/second or more.

Based on the results of the discovery phase, wedeveloped a design concept for the cultural website based on the idea of multimedia tours guidedby experts and artists. In our design, a tour pre-sents information to the user continuously, frombeginning to end, unless the user signals a desire toexplore extra material or to exercise control. Dur-ing the tour, opportunities to obtain additional in-formation are presented as hot spots on the screen.All of the extra information is also available for ex-ploration at the end of the tour.

The proposed web site consists of a collection oftours covering different areas of art and culture,such as painting, music, theater, and sculpture. Inorder to satisfy a large number of users, it seems tobe necessary to provide a large number of toursand to add new ones regularly. Some of the designdecisions presented below are, therefore, the re-sult of the need to reduce production costs.

To cope with the requirement of maximum 56Kbandwidth, we decided to explore multimedia ex-periences primarily based on still pictures and sound,

A

instead of low quality video.At 56Kbits/second,videotends to become small and blurry, while at thatspeed it is possible to download audio with rea-sonable quality and reasonably good-looking pic-tures. Dealing mainly with stills also serves to re-duce production costs, since shooting video is nor-mally quite a bit more expensive than just audiorecording and picture taking.

In our design, the main multimedia experience, orsimply main tour, is enriched by the addition of usercontrols such as pause/resume/rewind/forward andby the inclusion of hot spots for two kinds of extracontent side tours and branches. A side tour is anextra, self-contained segment of multimedia, nor-mally focusing on more specific aspects of the toursubject, which can be"inserted" into the main tour.A branch is a static web page with text, pictures,and links to extra information on a specific subject.The decision to create a side tour instead of abranch is based on the expected reach of the con-tent. Since side tours are much more costly to pro-duce than branches, we tend to produce side toursonly for highly desirable extra information.

Figure I shows a snapshot of a tour. Most of thearea is used by the content being presented by thetour (pictures,text, occasionally very short segmentsof video). On the bottom left side, a pictorial mapgives the user a basic idea of the duration of thedifferent scenes of the tour and the elapsed time.Rolling the mouse over the map brings extra tex-tual information about each scene, while clickingon the picture of a scene interrupts the currentscene and immediately starts the scene correspond-ing to the clicked image.

As the tour progresses, hot spots indicating sidetours and branches appear on the screen.When aside tour is selected, the main tour is interruptedand the side tour is played.When the side tour fin-ishes, the main tour automatically resumes fromthe point where it was leftA click on a branch pausesthe tour and opens a new window on the browser,displaying the web page associated with the branch.To resume the playing of the main tour, the user isrequired to click on the pause/resume marker abovethe map.

35© Archives & Museum Informatics, 2001 29

Vergo et.al., Results from User-Centered Design

mon for a aide tom

mon for a Mancn

rovsganon map

Figure I. Typical scene of a tour with itsnavigation map, including links to a

side tour and two branches

When the tour finishes, all the content - main tour,side tours, branches - of the tour becomes easilyavailable on the explore page. Figure 2 depicts theexplore page associated with the tour offig.l. Click-ing on the tour map restarts the tour from the be-ginning of the scene that was clicked.Similarly, click-ing on side tours and branches immediately startsthem.The explore page is automatically presentedto the user at the end of the main tour. However,during any moment of the main tour, the user canaccess the explore page by clicking on the corre-sponding hot spot on the right of the map (fig. 2).

The Two Pilot Tours

The design concept described above instructed thedevelopment of two pilot tours that are the maininstruments used in the evaluation methods laterdescribed by Pinhanez and Karat (Pinhanez 2001,Karat C. 2001, Pinhanez 2001).

The first tour features the work of a toy pianist,Margaret LengTan. In the tour, the pianist talks abouther involvement with toy pianos, how music is ar-ranged for a toy piano, and her connections toSchroeder, the famous cartoon character createdby Charles Schulz.Two side tours describe the his-tory and functioning of toy pianos and the work ofMargaret Leng Tan before becoming a toy pianist.The main tour lasts 4m15s, and the side tours takeIml8s and 50s, respectively.The tour also includesfive branches.

1

Figure 2. Exploration page that allowsaccess to the main tour and all side tours

and branches

The second tour talks about Ludwig van Beethovenand his Ninth Symphony.ln this tour, a narrator leadsthe user through aspects of Beethoven's life andwork, with comments also by a maestro and a con-cert pianist.Three side tours are provided, one aboutBeethoven's deafness and two others that let theuser explore Beethoven's scores and his famousHeiligenstaden testament. Beethoven's main tourlasts 10m1Os and the first side tour 2m00s.The otherside tours, since they incorporate interactive ele-ments, have no fixed duration, although their ex-ploration usually takes around 60s each.There arealso 5 branches available for user exploration.

After completing the development of the proto-types, we again engaged a sample of our target userpopulation in a usability study to assess the successand value of our designs.The prototypes were testedwith men and women from 21-55 years old. Testresults indicate high user satisfaction with the tours.Users interacted relatively infrequently with thetours, and the less they interacted, the more theyreported feeling engaged and entertained by theexperience.This supports our initial hypothesis thatthe tours are most entertaining when they are ex-perienced as they were designed to be; that is,watched as a TV-like experience.The results of theusability study give support to the contrarian "lessclicking, more watching" design approach identifiedin the discovery phase.

36

BESTCOPYAVAILARLF



References

Alberg, C. & Shneiderman, B. (1994).Visual informationseeking:Tight coupling of dynamic query filterswith starfield displays. Proceedings of CHI '94(Boston MA,April 1994), ACM Press, 313-317.

IBM, 2000, (http://www-3.ibm.cornlibm/easyleou_ext.nsf/Publish/ I 3)

Karat, C. et al. (2001), Less Clicking, More Watching:Results of the Iterative Design and Evaluation ofEntertaining Web Experiences. Submitted toInteract '01.

Pinhanez, C. et al. (200 I ),"Can Web Entertainment BePassive", Submitted to International World WideWeb Conference, Hong Kong,

Pinhanez, C. et al. (200I),"Less Clicking, More Watch-ing":An Option for Entertainment on the Web?"Submitted to CHI '0 I , Seattle.

Stone, M., Fishkin, K., & Bier, E. ( I 994).The movablefilter as a user interface tool. Proceedings of CHI'94 (Boston MA, April I 994),ACM Press, 306-312.

Viegas and Donath (1999), Chat Circles Proceedings ofCHI '99 ACM Press.

37

© Archives & Museum Informatics, 2001 3 I

Accessibility Techniques for Museum Web Sites

Susan Anable,The Dayton Art Institute and Adam Alonzo,WrightState University, USA

AbstractLike other public institutions, museums strive to make their facilities accessible to people with disabilities, yet thesesame patrons may be hindered in their use of museum Web sites by electronic accessibility barriers.This presentationwill demonstrate that access was a primary design factor in the Virtual Museum Tour, part of the Web site of TheDayton Art Institute. Developed in collaboration with Wright State University, planners of the Virtual Museum Tourfaced the challenge of making a variety of Internet technologies accessible to people with visual or hearing impairmentor mobility limitations. By adopting a set of Web accessibility guidelines and utilizing them in creative ways, the VirtualMuseum Tour provides an enlightening art experience for all visitors.

Introduction

The Dayton Art Institute has served residents ofsouthwestern Ohio since 19 I 9.The museum oper-ates on the premise that art is for everyone, andprovides the community with free general admis-sion 365 days a year. In addition, the Art Institute'sphysical facility is free of architectural barriers thatmight prevent people with disabilities from access-ing exhibitions or participating in programs. Rated"superb in quality" by the American Association ofMuseums,The Dayton Art Institute has one of thefinest mid-size museum collections in the country.As the Art Institute began to make this Collectionavailable on the World Wide Web, the same com-mitment to barrier-free access was demonstrated.

The Virtual Museum Tour is an interactive learningenvironment for Web visitors to explore The Day-ton Art Institute (http://www.daytonartinstitute.orgO.The Tour came about through a collaboration be-tween the Art Institute and Wright State Univer-sity (http://www.wrightedu/), located in Dayton andknown for its outstanding support program for stu-dents with disabilities. One of the primary goals ofthis project was to provide Web visitors with aninformative and absorbing experience equal to orbetter than what they might have in the Museum'sactual galleries.This naturally implied the inclusionof visual and auditory technologies to enhance thevisitor's examination of works of art.Yet the Touralso needed to be as accessible as possible for ev-eryone, including people with disabilities, as pre-scribed by the U.S. Department of Education grantthat funded the development of the Tour.Althoughthe idea of accessibility may seem incongruous to a

Web site enriched with images and sounds, it is notimpossible.Through cooperation and creativity, theproject partners were able to achieve both objec-tives.

One important factor that had to be determinedearly in the project was the definition of accessibil-ity as it pertains to aWeb site.There is a great dealof information about the topic available from books,Web sites and conferences, but the recommenda-tions given by these sources do not always concur.Fortunately, a single authoritative source hasemerged from the World Wide Web Consortium,the international body that leads the developmentof the Web.The Web Accessibility Initiative (WAI)within that organization has published a set of four-teen Web Content Accessibility Guidelines (http://www.w3.0rg/TR/WAI-WEBCONTENT/) to helpdesigners create accessible sites.Developed througha process of extensive research and review by ateam of experts, theWAI Guidelines were adoptedas the accessibility standard for theVirtual MuseumTour project.

Because the WA1 Guidelines are designed to beuniversal in scope, they do not specifically addressthe needs of unique audiences such as museums.Asa result, they often require interpretation or simpli-fication in order to be applied.This paper documentshow the Guidelines were implemented in practicalways on the Virtual Museum Tour, thereby creatinga type of accessibility style guide suitable for othermuseum Web site designers. Due to the limitationsof this paper, only a few of the fourteen Guidelines

438Museums and the Web 2001:Proceedings

Archives & Museum Informatics, 2001, p. 33

Anable,AccessibilityTechniques for Museum Web Sites

can be addressed, but more extensive informationabout accessibility techniques will be available onthe Virtual Museum Tour site itself.

Accessibility TechniquesAmong the fourteen WAI Guidelines, there are afew which have particular relevance to museumWebsites. These are listed in the sections below, alongwith the techniques used to implement them onthe Virtual Museum Tour.

Guideline I. Provide equivalent alternativesto auditory and visual content

Provide content that, when presented to the user,has conveys essentially the same function or pur-pose as auditory or visual content.

This is perhaps the most basic of the fourteenGuidelines, but it is also the most important. TheWeb in general, and museum sites in particular, relyheavily upon images to convey information; audiocontent is also found on manyWeb sites.But peoplewith visual or hearing impairments may be unableto perceive multimedia features such as images,sounds or video. If a Web page includes equivalentalternatives, however, all Web visitors will be ableto experience to some degree the content con-veyed by these visual and auditory elements. Equiva-lent alternatives frequently take the form of text,which is the most broadly accessible type of con-tent available on the Web. It is important to notethat Guideline I does not prohibit the use of multi-media such as images and sounds, but rather sug-gests how to convey that content in different waysso all users can understand its meaning.

Audio

The Virtual Museum Tour contains a large numberof audio clips, spoken by various commentators,containing information about works. In order toimplement Guideline I, each narration is accompa-nied by a text transcriptAll of the recordings werecarefully transcribed to ensure that the speaker'sexact words were accurately rendered into text(including sentence fragments and grammatical er-rors), and to ensure that names, places and termswere spelled correctly.

34

A typical audio segment on the Tour contains thecomments of Alexander Lee Nyerges, Director ofThe Dayton Art Institute, recounting how the Mu-seum acquired the painting Waterlilies by ClaudeMonet (Fig. 1). The excerpt below shows how thetranscript captures the conversational style of thenarration, and also demonstrates the need for ac-curacy in transcribing names and places.

It really isn't 'til you get to Monet's highestpoints, his series which include his series ofWaterlilies, that we find the fully developedand fully mature essence of what we nowregard as classical Impressionism.This wasdone in his garden in 1903, and probablyrepresents the highest point of his entireWaterlilies series.This work was given to theArt Institute by a gentleman by the name ofMr.Joseph Rubin,who ran a company in NewYork City called the Loma Dress Corpora-tion.And sometime before 1953, Dr. EstherSeaver, who was the Director of The Day-tonArt Institute in the first part of the 1950s,had encountered this picture and was ac-quainted with a dealer in New York by thename of Silberman ...

Fig. I: Claude Monet ( I 840-1926) French,Waterlilies, 1903, Oil on canvas, 32 x 40

inches, Gift of Mr. Joseph Rubin, 1953.11

When incorporated within theVirtual Museum Tour,text transcripts are provided alongside the audioclip itself, so Web visitors can choose the formatthat is appropriate for them.They can select eithera link that plays the recording or a link that leads to


Museums and theWeb 2001

the transcript on another page. In this manner, thetwo formats appear truly equivalent, since one isnot presented as subordinate to the other.

Although not applicable to theVirtual Museum Tour,it may be necessary to expand text transcripts ifthere is meaning conveyed by more than just thewords themselves.For instance, if an audio clip con-tains music or sound effects, these should be brieflydescribed in the transcript. In addition, if the toneof voice is significant, or the speaker is portraying aparticular character, this may also be alluded to inthe transcript

Images

In addition to audio, Guideline I also applies to vi-sual content such as images, certainly the most com-mon medium found on theWeb. In the context of amuseum Web site, images can be roughly catego-rized into two types:simple graphics such as logos,and complex pictures of artwork. Depending upontheir type, two different approaches may be usedto provide equivalent alternatives to images.

Simple graphics can be dealt with relatively easilyby taking advantage of a feature of HypertextMarkup Language (HTML), the language that con-trols the format and function ofWeb pages.To placean image on a Web page, a designer inserts a par-ticular HTML element into the document this ele-ment is referred to as an "image tag" and appearsas a three-letter abbreviation between brackets.

<img>

In addition to this term, another statement (calledan "attribute") must be placed inside the brackets.This attribute indicates the filename or"source" ofthe image; it too is abbreviated.

<img src = "filename"

While no further information is required to pro-duce a visible image on the page, there are severaloptional attributes that may be used to controlproperties of the image such as its size and its rela-tion to surrounding screen elements. One suchoptional attribute can be added to provide a textalternative that briefly describes what the image is.The alternative attribute is also abbreviated, and is

used alongside the source attribute within the sameimage tag.

<img src = "filename" alt ="description"

Text alternatives are not captions that appear along-side the image, and in most cases they will not evenbe visible on the page. However, there are severalways in which they benefit visitors, particularly thosewith visual impairments.Web visitors who are blindtypically use specialized computer software andhardware that reads pages aloud to them. If a Webpage includes text alternatives, they too will be vo-calized, informing users about the images found onthe page.Without text alternatives, users will notknow what the images are or why they are there.Further information on the use of text alternativesis available from the World Wide Web Consortium(http://www.w3.org/TR/htm14/struct/objects.html#adef-alt).

Every image on the Virtual Museum Tour is accom-panied by a text alternative that is generally no morethan a few words in length. Because of their brevity,text alternatives are not used to describe the ap-pearance of images, but simply to identify what theyare. For example, the following line of HTML codecould be used to identify the Art Institute's logo(Fig. 2):

<img src = "logo.gif" alt = "TheDayton Art Institute logo"

Fig. 2: Th e Dayton Art Institute logo

BEST COPY AVAILABLE


For simple graphics such as a logo, a brief text alter-native may be adequate, but pictures that showworks of art require a more detailed visual descrip-tion.Visual descriptions are short paragraphs thatverbally describe the appearance of images for thebenefit ofWeb visitors who are unable to see them.They serve the same role in the visual arts as audiodescriptions do for the performing arts, since bothprovide participants with a verbal impression ofthings they cannot see. The Virtual Museum Tourprovides a visual description for every, piece of art-work presented.

The WAI Guidelines suggest different techniquesfor including visual descriptions with complex im-ages (http://www.w3.org/TR/WCAG 10-HTM L-TECHS/#1ong-descriptions), but there is currentlyno standard practice in use. The Virtual MuseumTour uses visual descriptions to supplement ratherthan replace text alternatives:a text alternative givesthe title of a work while a visual description tellswhat it looks like. Links to visual descriptions areprominently placed on a navigation menu alongsidelinks leading to other sources of information aboutthe artwork, thereby presenting them as simplyanother resource for everyone, not just people withdisabilities.

Developing quality visual descriptions is a time-con-suming task that should be done thoughtfully byindividuals with excellent writing and editing skills.No instructions for doing so were available duringthe creation of the Virtual Museum Tour, but in thecourse of preparing approximately 100 visual de-scriptions, a standard process was developed.Thisprocess is summarized in the following six Recom-mendations which can be applied to other Webprojects that involve images of artwork.

Recommendation One: Be ObjectiveThe sole function of a visual description is to de-scribe the appearance of a work of art. In brief, itshould simply answer the question,"What does theobject look like?" Descriptions should avoid ana-lytical interpretations or emotional responses. Inother words, they should not explore questions like,"What does it mean?" or "How does it make youfeel?" Descriptions help visitors to visualize an ob-ject, thereby providing a context for other infor-mation about the work found elsewhere, such ashistorical background, the artist's style, or critical

36

e"


commentary. By combining this factual informationwith an objective description,Web visitors are suf-ficiently informed to make a personal analysis orachieve their own emotional response.

Objectivity should also be exercised when refer-ring to characters portrayed in a painting or otherwork of art. Though it is appropriate to describetheir appearance, clothing and actions,visual descrip-tions should not attempt to explain their motiva-tions or feelings, even if these are implied by ges-tures or context. If the emotions of the charactersare obvious in the artwork, then they will likely beapparent in the visual description as well.

/'

4

Fig. 3:William Adolphe Bouguereau (1825-1905) French,The Song of the Nightingale,

1895, Oil on canvas, SS x 35 inches, Gift of Mr.Robert Badenhop, 1954.12



In this example (Fig. 3), the description was writtenwithout references to emotion, although the moodimplied by the painting may still be perceived.

...The girl's body is turned partially to herleft while her head is turned to look to herright. Her brown hair is parted in the middleand pulled to the back of her head. She hasheavy eyebrows, dark eyes and reddishcheeks. Her mouth is small and her red lipsare closed. Both arms are extended and herhands rest on her knees with their fingersintertwined ...

Finally, because artistic taste varies according to theindividual, visual descriptions should not containvalue judgements on the quality of the art object,nor on the skill of the artist who created it.

Recommendation Two: Be BriefAlthough length will vary according to the piece ofartwork being described, visual descriptions shouldbe as brief as possible. In some settings, visual de-scriptions are presented as audio recordings readby a narrator. Other times they are provided as textthat visitors have to read by themselves. In eithercase, since extremely long descriptions are tiresome,they should be limited to 250 to 300 words. De-pending on the context, visual descriptions may beaccompanied by catalogue data about the artwork(such as the artist's name and life span, title, me-dium and dimensions). If so, this information neednot be included in the overall word count.

For better word economy, exclude redundantphrases like"rectangular in shape" or"blue in color"Simply use "rectangular" or"blue" instead. Likewise,avoid obvious statements such as"she wears a neck-lace around the neck" or "gloves on her hands:'

Recommendation Three: Be Descrip-tiveVisual descriptions should utilize a broad vocabu-lary of vivid terminology to describe various fea-tures of art objects. Some common terms are cat-egorized below.

Shapes can be described as square, cubed, rectan-gular, flat, straight, circular, spherical, cylindrical,curved, rounded, triangular, conical, pyramidal, an-gular, irregular, jagged, sloped, diagonal, horizontal

and vertical.These words can be used not only toidentify the overall shape of the object, but also todescribe geometric patterns within it.Avoid wordsthat imply action unless the object actually doesmove; for instance,"curved" is preferable to "curv-ing:'Also, colloquial terms such as "squiggle" or"zig-zag" should not be used.

Size can be described as small, tiny, short, miniature,large, tall, monumental, th ick, th in, narrow, wide, life-size, true to size, large scale and small scale. Theobject's dimensions, provided with the cataloguedata, will inform visitors of its actual size.

Texture can be described as smooth,glossy, coarse,grainy, rough, worn, weathered, scratched, cracked,broken, rippled, grooved, patterned, striped, dottedand perforated.

Color can be described as intense,vivid, bright, light,dark, dull, pale, faint, solid or blended.There is noneed to avoid references to color on the assump-tion that they will be meaningless to visitors whoare blind. First of all, descriptions will be used bypeople without visual disabilities. Second, manypeople who are now blind were able to see in thepast and are able to recall colors.Third,colors some-times have symbolic meaning in works of art (al-though interpretive phrases like "warm gold" orangry red" should not be used).

Composition (or the arrangement of elements in awork) can be described as low, high, above, below,parallel, perpendicular, in the foreground (or back-ground), and to the left (or right).When referringto relative locations, describe objects from theviewer's perspective, unless referring to the left orright of a character portrayed in the work.

Artistic technique can be described as realistic, ab-stract, unnatural, simplified, detailed, precise, impre-cise, sharply defined, blurred, splashed, brushed orstroked.

Recommendation Four: Be LogicalIn order to be easily understood, visual descrip-tions must describe objects according to a logicalsequence. Descriptions should begin with a generaloverview of what the object is and what it por-trays. Depending upon what type of object it is, itmay be appropriate at the outset to mention its

4 2© Archives & Museum Informatics, 2001 37

Anal) le,AccessibilityTechniques for Museum Web Sites

color and surface texture, and perhaps its construc-tion. Following the overview, the various portionsof the object should be described in detail, in someorderly fashion such as left to right or top to bot-tom. After one portion of the work has been de-scribed, an explicit transition should be used to iden-tify the next area and its spatial relationship to thelast. If part of the object is extremely complex, it isbest to describe each segment separately, perhapsin a numbered sequence.

In the following example (Fig. 4), the location ofobjects in a painting are described in relation toother elements, providing the reader with an un-derstanding of the overall composition of the work.

...To the left of the barn is an area of criss-crossed black strokes like a shadow. Belowthe barn is a red and gray machine placedawkwardly on its side. Its large wheels andgears give it the appearance of a piece offarm machinery.To the right of the machineare two brown and angular tree branchesthat rise upward from the bottom edge ofthe picture. Between the branches, severalsmall scenes are crudely rendered one aboveanother ...

Fig. 4: Stuart Davis (1894-1964) American,Landscape with Broken Machine, 1935,

Gouache on paper, 15 1/4 x 22 1/8 inches,Bequest ofVirginia Rike Haswell, 1977.39

Depending on their design, sculptures or otherthree-dimensional works will likely need to be de-scribed from more than one angle. A logical se-quence should be used when doing so, as if theviewer was moving in a circle around the object.

When using descriptive words such as adjectives, itis often best to place them after the word theymodify, so visitors know what the object is beforethey are told what it looks like. For example, use"his fingers are long and thin" instead of "he haslong, thin fingers?'

Recommendation Five: Be AccurateBecause visual descriptions are part of the overalllearning experience of art patrons, they should befactual and consistent with other sources of infor-mation about the artwork. Background research maybe necessary to correctly identify historical figures,religious personalities, geographical locations, typesof clothing, breeds of animals, architectural elementsand so on. However, descriptions should avoid jar-gon or specialized terminology that may be unfa-miliar to most visitors. For example, the stylisticterms "abstract" and "realistic" are likely to be un-derstood, but "Geometric Abstractionist" and"French Academic" are not.

It is strongly recommended that a describer lookat the actual object when making initial notes, be-cause color and small details are most accuratelyobserved in person.When these notes are refinedinto the actual description, slides or photographsmay be used as a reminder of the object's appear-ance and composition.

Recommendation Six: MiscellaneousBecause visitors on a virtual tour may access worksof art in any order, writers of visual descriptionsshould not assume that readers have followed aparticular sequence. References to other artworkshould not be made within a visual description, noteven to works by the same artist or from the samegallery.

For the benefit of Web visitors using screen read-ing software, numbers should be spelled out aswords, as in "nineteenth century" rather than" 19thcentury," and "two and a half" instead of "2 I /2:'

Descriptions ought not to assign gender to animals,such as "the dog stands on his hind legs:'

When describing the clothing of characters in awork, monotony can be minimized by using variedphrases like "dressed in,"is wearing" and "has on:'

4 3


Once visual descriptions have been written andedited, several reviewers (including people with vi-sual disabilities) should provide suggestions for im-provementAs a final check, descriptions should becompared, by a team member, in person, to the ac-tual works of art in the galleries.

Video

Although the Virtual Museum Tour currently doesnot contain video clips, the principle of equivalentalternatives also applies to video. In general, it isnecessary to provide synchronized captions for allspoken words, as well as descriptions for sound ef-fects, music and visuals. Further details and examplesare available at the National Center for AccessibleMedia (http://main.wgbh.org/wgbh/pages/ncam/richmedia/index.html).

Guideline 4. Clarify natural language usage

Use markup that facilitates pronunciation or inter-pretation of abbreviated or foreign text.

The purpose of this guideline is to make it easierfor people using adaptive computer equipment tounderstand the content ofWeb documents, particu-larly those that contain periodic language changes.For example, a Web visitor who is blind will prob-ably utilize a certain type of adaptive software calleda screen reader to vocalize the words on a page. Byidentifying the language of the text, the screen readermay be able to respond with the proper pronuncia-tion and accent. This technique can also benefitpeople who readWeb pages with the aid of a Brailledevice, since words or characters may be displayeddifferently depending upon the language in use.

For a museum Web site this Guideline is crucialdue to the frequent use of terminology from a vari-ety of languages. As the Virtual Museum Tour wasbeing developed, a large number of new documentswere written,and existing documents were adaptedfor the site. In both cases, these catalogue entriesand commentaries were closely checked, and varia-tions in language were noted.This process was alsoapplied to all visual descriptions and text transcriptsof audio segments, which are accommodationsthemselves.

In order to implement Guideline 4, the primary lan-guage of a document must first be identified,: thenany subsequent changes in language must be indi-cated.The initial step is accomplished by adding thelanguage attribute to the opening HTML tag foundat the top of the document. Instead of using the fullname of a language (such as English), a two-letterabbreviation is used.

<html lang = "en">

A lengthy list of these two-letter language codes isavailable from the Organization for the Advance-ment of Structured Information Standards (http://www.oasis-open.org/cover/iso639a.html).

Once the primary language of a document is iden-tified, changes in language that occur within the textalso need to be declared.This can be done by usinga pair of span container tags with the language at-tribute.The opening span tag marks the beginningof a new language segment while the closing tagindicates the end. After the closing span tag, thedocument reverts to the primary language previ-ously declared.The following example is an excerptfrom commentary written by Clarence W. Kelleyabout a pair of Japanese screen paintings (Fig. 5).

4 4

... By the end of the 17th century, a distinc-tive association had developed between clas-sical literature and a type of screen paintingknown as byobu-uta (poems for screens), as seen inthis pair of screen paintings ...

Fig. 5: Japan, Edo period ( I 615 - I 868) MuashiPlain, late I 7th - early 18th century, Pair of

six-fold screens: ink, colors and gold on paper,Each panel 66 1l2 x 24 I /4 inches, Museum

purchase, I 960.24a-b

@Archives & Museum Informatics, 2001 39


In addition to language changes, Guideline 4 alsoapplies to the use of abbreviations and acronymswithin Web documents. Further details are avail-able from the Web Accessibility Initiative (http://www.w3.org/TR/WCAG I 0-HTML-TECHS/#text-abbr).

Guideline 8. Ensure direct accessibility ofembedded user interfaces

Ensure that the user interface follows principles ofaccessible design:device-independent access to func-tionality, keyboard operability, self-voicing, etc.

This Guideline refers to various Web site featuresthat present users with some sort of onscreen con-trols besides the typical buttons and menus foundin a browser. It is important that these controls beoperable through different methods, to accommo-date as many people as possible. Computer userswith disabilities frequently utilize input devices otherthan a mouse, since using a mouse requires bothsight and manual dexterity.A person who is blind,for instance, might rely upon the keyboard, whilesomeone with a mobility impairment might use ahead-mounted pointing device.While Web brows-

eyeapyRetores That ny

Fig. 6: Frans Geubels (1535 - 1590) Flemish,Kink Abimelech Restores Sarah to her

Husband, Abraham, around 1560 - 1570,Tapestry: dyed wool and silk, 166 x 185 inches,

Gift of Mr. Robert Badenhop, 1952.10

40

ers themselves can be controlled through thesemethods, additional interfaces may not have thesame functionality. A common example found onmuseum Web sites is an image enlargement systemthat allows visitors to view works of art in detail.Depending upon the technology, users may be re-quired to click the mouse on the image itself tozoom in. If so,this feature may be difficult for peoplewith certain disabilities to use.

In the process of selecting an image enlargementsystem for the Virtual Museum Tour, keyboard op-erability was a primary concern. Many popular tech-nologies failed in this respect, but a suitable prod-uct was found in the EyeSpyTM Image Server fromAXS Technologies (http://www.axs-tech.com/). Itallows visitors to use a mouse to click on the im-age, but it also provides a keyboard-accessible in-terface (Figs. 6 and 7).

ConclusionsWhen considering the topic of Web accessibility, itis beneficial to note how the addition of ramps andautomatic doors have made access to museum build-ings easier not only for people with disabilities, butalso for all visitors.A similar phenomenon is likely

45

7LY,Fly ? 00009Fig. 7: Frans Geubels (1535 - 1590) Flemish,

King Abimelech Restores Sarah to herHusband,Abraham. (detail)

© Archives & Museum Informatics, 2001


to take place as museums improve the accessibilityof their Web sites. By observing the WAI Guide-lines, designers will make their pages more usableby everyone. For instance, text transcripts and de-scribed videos can be used by those who do nothave the computer hardware or software to playmultimedia, and by people with a slow Internet con-nection. Likewise, visual descriptions may be appre-ciated even by people with normal eyesight, sincethe descriptions point out details that they mightnot have noticed. By incorporating the WAI Guide-lines into their design procedures, museums will takea major step toward accomplishing the goal of mak-ing art accessible to everyone.

AcknowledgementsThe Accessible Arts! program at Wright State Univer-sity operates with funding from the RehabilitationServices Administration of the U.S. Department ofEducation, and in collaboration with a consortium ofarea arts organizations and businesses.

Thanks to Alan Woods and Nancy Van Voorhis for theirinformative audio description workshop; to Nancy VanVoorhis and Nancy Mitchell for writing numerous visualdescriptions; and to Mike Paciello for tips on theimplementation of visual descriptions in Web pages.

References

AXS Technologies EyeSpyTM Image Server (2000).Consulted January 14, 2001. hap://www.axs-tech.com/

National Center for Accessible Media Access to RichMedia Project (2000). Rich Media Accessibility:Resource Center for Developers. Last updatedDecember 6, 2000, consulted January 3, 2001.http://main.wgbh.org/wgbh/pages/ncam/webaccess/arm/index.html

Organization for the Advancement of StructuredInformation Standards (1989). Code for theRepresentation of the Names of Languages. Lastupdated July 15, 1998, consulted January 8, 2001.http://www.oasis-open.org/coverliso639a.html

World Wide Web Consortium HTML 4.01 Specifica-tion (1999). Objects, Images, and Applets in HTMLdocuments. Last updated December 24, 1999,consulted January 3, 2001. http://www.w3.org/TR/htm14/struct/objects.html#adef-alt

World Wide Web Consortium Web AccessibilityInitiative (1999).Web Content AccessibilityGuidelines 1.0. Last updated May 5, 1999, con-sulted January 2, 2001. http://www.w3.org/TR/WAI-WEBCONTENT/

4 6@Archives & Museum Informatics, 2001 41

Evaluating the Usability of a Museum Web Site

Ilse Harms and Werner Schweibenz, University of Saarland,Saarbriicken, Germany

Abstract

The paper presents a research project conducted by the Department of Information Science in cooperation with theSaarland Museum, the art museum of the Federal State of Saarland, Germany.The study had two aims.The first was toevaluate some methods of usability engineering for the Web, and the second was to evaluate the usability of theSaarland Museum's Web site and improve it. The applied usability engineering methods were an expert-judgment-focused evaluation using heuristic evaluation with the Heuristics forWeb Communication and a user-focused evaluationconducting a laboratory test with actual users and the thinking-aloud method.The combination of heuristic evaluationand laboratory testing provided interesting results.The heuristic evaluation detected a vast number of usability problems.The laboratory test confirmed most of these findings as usability problems and added some usability problems thatexperts did not discover because actual users often have a different perspective.The evaluation led to a re-design ofthe Web site.

I. Usability Engineering for MuseumWeb Sites

The World Wide Web offers museums the possi-bility to disseminate information about their col-lections to a world wide public.Therefore, one canstate that"museums are in the communication busi-ness" (Silverstone 1988, 231) and that there areinteresting parallels between museums and the massmedia (Schweibenz 1998, 187). When attributingcommunicative and interactive functions to a mu-seum Web site, it is especially important to keep inmind the users and their needs for easy interac-tion.Therefore it is essential that"the Web is a do-main which must be instantly usable" (Rajani &Rosenberg I999).This statement evokes the ques-tion of what usability means for museum Web sites.According to Garzotto, Matera & Paolini (1998),usability is "the visitor's ability to use these sitesand to access their content in the most effectiveway. As a consequence, it has become compellingto provide both quality criteria that WNW sitesmust satisfy in order to be usable, and systematicmethods for evaluating such criteria:' ThereforeWeb usability and Web usability engineering meth-ods have become an important issue (Garzotto,Matera & Paolini I 998,Teather 1999, Cleary 2000).

The usability of Web sites can be tested and im-proved in a process which is called usability engi-neering.According to Kromker (1999, 25), usabilityengineering is a set of methods to design user-friendly products and enhance the quality of theproduct.The methods of usability engineering can

be categorized in expert-focused and user-focusedmethods. Expert-focused methods like heuristicevaluation and user-focused evaluation methods likelaboratory testing with actual users can be used incombination (Nielsen, I 997a, 1543). Undisputedly,the combination of heuristic evaluation and labora-tory testing achieves the greatest value from eachmethod (Kantner & Rosenbaum 1997, 160; Nielsen,1993, 225).Therefore, the Department of Informa-tion Science at the University of Saarland,Germany,developed a usability engineering process (Harms& Schweibenz,2000,19-20) and tested it in a usabil-ity study evaluating a museum Web site.This paperpresents the experiences from the study which wasconducted with an expert-judgment-focused evalu-ation using heuristic evaluation based on the Heu-ristics for Web Communication and a user-focusedevaluation using laboratory tests with actual usersand the thinking-aloud method.

2.A Survey of Methods for UsabilityEngineering

The methods of usability engineering can be cat-egorized into expert-focused and user-focusedmethods. Among the expert-focused methods areseveral variations of heuristic evaluation.Accordingto Nielsen ( I 997a, 1543) "heuristic evaluation is away of finding usability problems in a design by con-trasting it with a list of established usability prin-

I '7Museums and tne Web 2001: Proceedings


BESTCOPYAVAILABLE

Harms and Schweibenz, Evaluating the Usability of a Museum Web Site

ciples".The established usability principles are listedin guidelines or checklists like Keevil's Usability In-dex (Keevil, 1998) or Molich & Nielsen's nine prin-ciples for human-computer dialogue (Molich &Nielsen, 1990) or the Heuristics for Web Com-munication. In the evaluation process, experts com-pare the product with these guidelines and judgethe compliance of the interface with recognizedusability problems.The advantage of expert-focusedevaluation is that it is a relatively simple and fastprocess.A comparatively small number of five evalu-ators can find some 75 per cent of the usabilityproblems of a product in a relatively short time(for details see Levi & Conrad, I 996).The disadvan-tages are that experts have to do the evaluationand that experts cannot ignore their own knowl-edge of the subject, i.e. they cannot "step back be-hind what they already know:' So they will alwaysbe surrogate users,i.e. expert evaluators who emu-late users.

Web site

suggestönsfor the

redesign

In contrast to expert-focused methods, user-focusedmethods rely on actual users to test the usability ofa product This process is called user testing, andaccording to Nielsen (1997a, 1543) it "is the mostfundamental usability method and is in some senseirreplaceable, since it provides direct informationabout how people use computers and what theirexact problems are with the concrete interfacebeing tested:' There are various methods for usertesting. One of the most popular and most effectivemethods is the laboratory test with the thinkingaloud method (Nielsen 1993, 195) which was usedin our case study. The advantage of user-focusedevaluation is that the tests supply a huge amount ofqualitative data that show how actual users handlethe product The disadvantages are that the teststake place in a laboratory situation and that a lot ofequipment and coordination is necessary to con-duct the test, making it labor-intensive.

heuristicevaluation

detected usa-bility problemsseveriV rath g

develcpingtest materid

questionnaires

test tasks

analysing tte data data from user testswithof user tests aid

ieuristicevaluationthe user

teststhinkingaloud

Figure I:The evaluation process of the usability study

4 8 BEST COPY AVA !LAD_ F



3. Description of the Usability Study

3. I The Process

In cooperation with the Saarland Museum - StiftungSaarlandischer Kulturbesitz, the art museum of thefederal state of Saarland, the Department of Infor-mation Science at the University of Saarland evalu-ated the museum's Web site (http://www.saarlandmuseum.de).The site is a graphicallydesigned Web site of the third generation (Siegel,1997) and went online in summer I 999.The evalu-ation project was carried out by sixteen graduatestudents who had received training in usability en-gineering in a research class, and two lecturers ascoordinators.The study had two aims.The first aimwas to evaluate some evaluation methods, espe-cially the Heuristics forWeb Communication; the sec-ond was to improve the usability of theWeb site ofthe Saarland Museum.Therefore we decided to usea combination of heuristic evaluation and user test-ing as suggested in the research literature. Figure 1illustrates the process.

3.2 Heuristic Evaluation With theHeuristics for Web Communication

The first step was an heuristic evaluation.As men-tioned above, there is a multitude of heuristics.Heuristics can be specific for a certain domain orgenerally applicable.They can be design-oriented orevaluation-oriented or both.They can be based onresearch or on experience of practitioners.There-fore, heuristics vary in extent and quality. In ourcase study we used the Heuristics forWeb Communi-cation, developed by the faculty of the Departmentsof Technical Communication of the University ofWashington, Seattle, and the University of Twente,the Netherlands. The heuristics are based on re-search findings in technical writing and cognitivepsychology and were evaluated in a workshop with40 participants, both students and lecturers of tech-nical communication, and professional Web devel-opers from various Web design companies in theSeattle area.The heuristics were revised accordingto the feedback of the participants of the work-shop and the Web developers and were publishedin a special issue of the Journal of TechnicalCommunication in August 2000.

The Heuristics forWeb Communication consist of fivedifferent heuristics.The five heuristics deal with all

important aspects of Web sites: displaying infor-mation, navigation, text comprehension, role play-ing (i.e. author-reader relationship), and data col-lection for analyzing interaction.The content of theheuristics can be summed up as follows:

The heuristic Displaying information on the Web con-sists of guidelines for visuals, e.g.

how to design and arrange display elementshow to ensure that text is readablehow to use pictures, illustrations, icons and mo-tion

The Heuristic forWeb Navigation deals with hypertexttheory and offers guidelines for navigation and ori-entation, e.g.

how to design orientation information on eachpage

how to coordinate navigation deviceshow to design site-level orientation information

The heuristic Text Comprehension and theWeb: Heu-ristics for Writing Understandable Web Pages focuseson text comprehension and issues of text quality,e.g.

how to select, design, and organize contentwhat style to usewhat makes Web pages credible and trustworthy

The heuristic Role Playing on the Web discusses thetypical rhetorical roles of the implied author andreader of the Web pages and their rhetorical roles,e.g.

how rhetoric is used to describe author roles andreader roles, andwhat kind of relationship exists between authorroles and reader roles

The heuristics Web Data Collection for Analyzing andInteracting with Your Users focuses on analyzing theaudience of a Web site and building a relationshipeither between you and your users or among theusers themselves, using for example

server log data for analyzing the use of Web pagesand their audience, andmeans to build a relationship and create a senseof community with the audience.

4 9@Archives & Museum Informatics, 2001 45


The four content-oriented heuristics (the heuris-tics on Web Data Collection was not applied due toaccess restrictions to log files) were applied accord-ing to Kantner & Rosenbaum (1997, 155).The gra-duate students, who had received an introductionto the heuristics, worked in teams of four. Each teamconducted an evaluation of the same selected num-ber of pages from the chosen Web site. In a two-hour session of individual evaluation, the team mem-bers applied one of the four heuristics to the Website. During the evaluation process they took notesof usability problems according to the various pointslisted in the heuristics. Then the team membersgathered and discussed their findings.The usabilityproblems detected in the heuristic evaluation weregraded in a severity rating according to Nielsen(1997b) by each team.The rating ranged from 0 (nousability problem) to 4 (usability catastrophe) (cf.Table I ) and was conducted with respect to thefrequency and persistence of the problems and theimpact they have on users.At the end of the evalu-ation, the four teams presented their findings in aplenary meeting.The evaluation process took aboutfive hours.

0 I don't agree that this is a usability problem at allI Cosmetic problem only: need not be fixed unless

extra time is available on project2 Minor usability problem: fixing this should be

given low priority3 Major usability problem: important to fix, so

should be given high priority4 Usability catastrophe: imperative to fix this

before product can be released

Table I:Severity rating according to Nielsen

The most frequent usability problems were naviga-tional and orientation problems as described in theHeuristic forWeb Navigation, followed by general de-sign problems as named in the heuristic Displayinginformation on the Web.After the rating the two lec-turers collected the written findings of the teammembers and compiled a list of all usability prob-lems for a re-design of the Web site. The findingswere also used to design tasks for a user test in thelaboratory.

3.3 UserTesting in the Usability LaboratoryWith The Thinking Aloud Method

The next step of the evaluation was a user test inthe usability laboratory. Figure 2 shows a sketch ofthe laboratory in which the tests took place.

In the lab, real users have to work on tasks whilethinking aloud, i.e. they verbalize their thoughts andcomment on their actions while they handle thecomputer.This "allows a very direct understandingof what parts of the dialogue cause the most prob-lems" (Nielsen 1993, 195). During the test, userswork on standardized test tasks and are supervisedby a test manager.The tests are recorded on videoby a technical assistant who operates two videocameras. One of the cameras is focused on the faceand hands of the participant, the other one on thecomputer screen.The recordings of the two cam-eras are blended together on one screen and re-corded on video. In order to catch the details ofinteraction, a digital screencam records the actionson the screen. In a labor-intensive process, the find-ings of the tests are transcribed and categorized.

As we evaluated an art museum Web site, recruit-ing participants with an interest in art was suggested.Therefore we asked students of the Arts and Sci-ence department and art teachers to participate inour experiments.The teachers were chosen to in-crease the average age.The number of participantswas decided according to Virzi (1992, 468), whosuggests at least 15 participants. In our study, 17users participated. Five of them were teachers; 12were students of the Arts and Science department.Seven participants were male, ten female.The young-est user was 19, the oldest 48, the average age be-ing 27.

It takes some time and effort to design the testtask scenario for the user test of a large informa-tional Web site (cf. Kantner & Rosenbaum 1997,I54).The test tasks should be as representative aspossible of the use to which the system will be putin the field and at the same time small enough to becompleted in a reasonable time frame, but not sosmall eas to be trivial (Nielsen 1993, 1859.The testscenario, which had been discussed with the clientof our case study, consisted of nine tasks that rep-resented potential usability problems.Table 2 showsa selection of the potential problems and the testtasks.



video camera recordingthe actions on the screen

participant

working place ofthe participant

thinking alouds

computer screenand keyboard

mixin consolefor the ideocameras

test manager

micro-phone

video recordeand tv forrecording

video camera recor-ding the actions ofthe participant

technicalassistant

Figure 2: A sketch of the usability lab

Potential Usability Problem Selection of the test tasks(abbreviated)

Links are hidden in graphical design (images). I) Go from the splash screen to the corepage.

Insufficiently linked information, theexhibition is not linked to the opening hours.

3) Look for the opening hours of a futureexhibition.

Insufficiently linked information, theexhibition is not linked to the service sectionwhere tours are offered.

4) Look for guided tours to the currentexhibition.

All pages are titled the same. There are noindividual title-tags on the different pages.

7) Use bookmarks to go back to certainpage.

Table 2:A selection from the test tasks of the user tests

5.1



The user tests revealed that all assumed problemswere usability problems for test users.The findingswere grouped into several categories, e.g. problemshandling the splash screen (task 1), insufficient useof links between related information (tasks 3 and4) etc., and illustrated by lively quotes taken fromthe test protocols. For example, after finishing task1 a participant stated:"The first screen only showsa headline, a picture and an address but no link.(break) I click on the picture. It works!" Anotherremarked:"As an Internet beginner I honestly havea problem to get to the next page. I'm a little help-less because I prefer big arrows and buttons thatsay 'next page.' But I made it." While struggling withtask 4, the patience of a participant snapped and hesaid:"Now I would try to get in touch with some-one and write an email, if they offer a guided tour,because all the searching takes far too long:' Tryingto use the bookmarks in task 7, a participant said:"Usually I would use the bookmarks. (clicks onbookmarks) Well, now I see that all the bookmarkshave the same name:'Another participant used sar-casm:"That's really funny, because I have five book-marks all named Saarland Museum Saarbrucken.Thathelps a lot. Great! (laughs):'

Quotes like these make usability problems comealive and show the reactions of actual users in realsituations. From our experience,these reactions aremuch more convincing than dry statements of ex-perts no matter how profound these statementsmight be and no matter on what kind of theorythey are based.

The test time and performance of the users wasinfluenced by their computer literacy, measured bydata collected in a questionnaire.The average testtime was about 20 minutes and corresponds withthe expected test time derived from three pre-tests.The duration of the test shows some connectionbetween the computer literacy level and the amountof time needed to complete the tasks.The shortesttest of 14 minutes was conducted with a user whohad used the Web for two years or longer and sev-eral times a week. The longest test of 30 minuteswas conducted with a participant with little com-puter literacy. Further analysis did not seem neces-sary to us because the duration of a test is alsoinfluenced by other factors like interest in the sub-ject and the medium and the method of thinkingaloud.

4. Practical ExperiencesWith the Dif-ferent Methods

In our case study the theoretical foundations of theheuristic evaluation were the Heuristics forWeb Com-munication.At the time of our evaluation, these heu-ristics were brand new. Only a little practical ex-perience existed in applying these heuristics.There-fore we thought it useful to contrast the HeuristicsforWeb Communication with another heuristic evalu-ation tool, Keevil's Usability Index.According to Keevil(1998, 271), the usability index is a "measure, ex-pressed as a per cent, of how closely the featuresof a Web site match generally accepted usabilityguidelines." The Usability Index consists of five cat-egories (Keevil 1998, 273):

Finding the information: Can you find the infor-mation you want?Understanding the information:After you find theinformation, can you understand it?Supporting user tasks: Does the information helpyou perform a task?Evaluating the technical accuracy: Is the technicalinformation complete?Presenting the information:Does the informationlook like a quality product?

Judging from the 203 questions, Keevil's Usability In-dex seems focused on commercial Web sites. But itseemed adequate to use it for the Web site of acultural heritage institution, because Keevil (1998,275) points out that the Usability Index is generallyapplicable: "Information Developers can use thechecklist to measure how easy it is to find, under-stand,and use information displayed on aWeb site:'

In our case study, fifteen students (one of the six-teen didn't hand in the index) used the Usability In-dex to evaluate the Web site of the Saarland Mu-seum.The following table (Table 3) shows the re-sults with the categories 'N/A' (not applicable),`Yes','No', the sum and the usability in percent, which iscalculated from the total number of 'Yes' answersdivided by the total numbers of 'Yes' and 'No' an-swers.

Table 3 shows a wide range of usability for theSaarland Museum expressed in per cent. The re-sults reached from 29% to 55%, the arithmetic meanbeing 47%.The deviation in the percentage of us-



Usability Index NIA Yes No Sum Usabilityin percent

Evaluator I 96 57 47 200 55 %Evaluator 2 134 19 47 200 29 %

Evaluator 3 71 67 62 200 52 %

Evaluator 4 71 67 62 200 52 %

Evaluator 5 81 51 68 200 43 %

Evaluator 6 62 69 69 200 50 %Evaluator 7 96 48 53 197 48 %Evaluator 8 64 61 74 199 45 %

Evaluator 9 77 60 62 199 49 %Evaluator 10 77 60 62 199 49 %Evaluator I I 77 60 62 199 49 %

Evaluator 12 81 61 58 200 51 %

Evaluator 13 66 57 53 176 52 %Evaluator 14 69 53 81 203 40 %Evaluator I S 88 40 72 200 36 %

Table 3: Results of Keevil's Usability Index forthe Saarland Museum Web Site

ability is remarkable. It is due to the bias of theevaluators in interpreting the questions. One rea-son is that certain sets of questions from the Us-ability Index did not fit for theWeb site under evalu-ation. In this case, some evaluators chose to votefor 'N/A', while others voted for `No'.This explainsthe big differences between the 'N/A' votes and the'No' votes and in the overall usability because theusability is calculated from the total number of yesanswers divided by the total numbers of yes and noanswers. Another reason is that the checklist, likeevery checklist, is open to interpretation. Keevil(1998,275) was aware of this problem and tried toreduce it by only allowing the answers 'N/A',`Yes',and 'No'. But still, there is considerable room forinterpretation.Apart from the problem of interpre-tation, the Usability Index has another disadvantage.It gives a number in per cent that indicates the us-ability and some hints on usability problems thatcan be derived from the 203 questions, but it hardlyidentifies concrete usability problems.

The identification of concrete usability problemsand suggestions on how to improve usability arethe practical advantages of the Heuristics for WebCommunication.The heuristics are not simply check-lists that can be answered by 'N/A',`Yes', and 'No'.They are guidelines designed as statements andquestions that guide the evaluator to identify con-crete usability problems by asking guiding questionslike,"Can you decipher all of the elements in thedisplay easily? If not, consider making them larger,"or "Which of the organization's values should beemphasized?" or giving hints like, "Make sure the

most important links appear high enough on thepage to be visible without scrolling, regardless ofthe resolution of the user's monitor. When pagesmust scroll, provide visual cues to encourage usersto scroll down to links that are below the scrollline:' By contrasting these established usability prin-ciples with theWeb site under evaluation, the evalu-ator or information designer can decide if usabilityproblems exist, what kind they are, and how theycan be removed. This is the big advantage of theheuristics.

The disadvantage of the Heuristics for Web Commu-nication is that they are very detailed and compli-cated compared with general heuristics like the onesof Molich & Nielsen. Molich & Nielsen suggest ninebasic items of usability (Molich & Nielsen 1990, 338)

(Table 4).

I Use simple and natural language

2 Speak the user's language

3 Minimize the user's memory load4 Be consistent5 Provide feedback6 Provide clearly marked exits7 Provide shortcuts8 Provide helpful error messages9 Prevent errors

Table 4: Molich & Nielsen's nine basic itemsof usability

Although it is possible to do a successful evaluationwith these nine basic items of usability, evaluatorsmight need more guidance in the evaluation pro-cess, as is offered by the Heuristics forWeb Communi-cation.These heuristics support the evaluators byproviding a structured "guided tour" for the evalu-ation process that takes into account both the bigpicture and important details.They help the evalua-tors to consider all substantial usability issues andto focus on the important points. Moreover theygenerate a profound impression of the overall qualityof a Web site. This makes the Heuristics for WebCommunication a valuable tool forWeb usability en-gineering.

The four content-oriented heuristics (the fifth heu-ristic was not applied due to access restrictions to



log files) are very different as far as their ease ofapplication and the level of background knowledgeare concerned.We found that the heuristic Display-ing information on the Web, the Heuristic forWeb Navi-gation, and the heuristic Text Comprehension and theWeb can be successfully applied if the evaluatorshave an average level of knowledge in informationdesign and Web design.The evaluators in our casestudy, all graduate students of information science,had no difficulties in applying them. The heuristicRole Playing on theWeb requires some special knowl-edge of hypertext theory as it is based on the quitecomplicated author-reader relationship in hypertext(Michalak & Coney I 993).Although it is very inter-esting and provides promising results, the evalua-tors in our case study had some difficulties in ap-plying it.

As expected from the research literature (Nielsen1992, 3780, the heuristics detected a great numberof so-called minor usability problems- no disad-vantage at all because user testing is not an adequatemeans to identify such minor problems.Minor prob-lems were, for example, inconsistent use of linkcolors, no text messages for graphic links, compli-cated sentences, deficits in page structure and or-ganization, lack of informative titles, meaninglessanimation, flaws in the author-reader relationship,etc. Although real usability problems, such minorproblems are not observable in user testing, be-cause average users do not realize that these kindsof deficiencies cause problems because they lackthe background knowledge in information designand Web design.

The user tests in the usability lab were very labor-intensive for several reasons: the technical equip-ment had to be arranged, the test scenario had tobe designed and tested, participants had to be re-cruited, and tests had to be conducted with twoexperimenters who had to be present all the timeto supervise the participants and the technical equip-ment. The analysis and evaluation of the test datawas also time-consuming because the data had tobe transcribed and categorized.The big advantageof this method was that the recordings, especiallythe screencam files, showed cursor movements thathelp to identify problems in navigation and orienta-tion. This is especially helpful when discussing thefindings and suggestions for the redesign with theclient.A simplified method of thinking aloud testing,

54

in which the experimenters simply take notes oftheir observations, is less labor-intensive than vid-eotaping and transcribing the tests. But from ourexperience, it is difficult for one or two experiment-ers to follow the course of the test and take notesat the same time if the test consists of more thansome basic functions. Therefore video taping orscreencam recording is essential.An alternative totranscribing whole test sessions would be to tran-scribe only the most important sequences of a test.

An important point we noticed when comparinganswers in the questionnaires with courses of thetests is that answers about the test experience areoften not very reliable.The answers about the sat-isfaction with the Web site often did not corre-spond with the actual experience of the participantswho were observed during the test. For example,several participants stated that they had no prob-lems with navigation and orientation although theyhad had serious problems during the test.The rea-sons for this gap between behaviour and statementscannot be discussed here. From our experience itis important to remark that a questionnaire alonecannot provide reliable results.This is not new butconfirms the phenomenon that impressions of theirown behaviour and the behaviour in the situationof social reality show a certain deviation. Despitethis fact, from our experience questionnaires orinterviews are necessary to give the participantsthe opportunity to comment on the course of thetest.Test users appreciate this opportunity, and theresults can be used to derive additional informa-tion about the acceptance of the Web site.

5. Conclusion

The evaluation method used in usability engineer-ing depends on the subject that is evaluated andthe goals of the evaluation.Although the combina-tion of heuristic evaluation and user testing pro-vides good results, it is costly as far as time andresources are concerned.With respect to the cost-benefit ration, in many cases the heuristic evalua-tion is sufficient to detect a reasonable number ofminor and major usability problems.

In our case study, the Heuristics forWeb Communica-tion proved to be applicable tools for heuristic evalu-ation. The heuristics support a structured evalua-tion and help both to find and to solve usability


problems. In contrast to checklists, they give theevaluators some scope for interpretation while of-fering guidance at the same time.The drawback ofthe heuristics is that they cannot be successfullyapplied by novice evaluators. The evaluators needsome background knowledge in Web design andevaluation. The heuristics were helpful in pointingout critical points in the Web site that were evalu-ated in the user test.Compared to user testing, theheuristic evaluation was less labor-intensive. Nev-ertheless, user testing is a very valuable tool forusability engineering because actual users give animpression of how theWeb site will be used in prac-tice. Moreover, actual users might have problemswith features that experts do not realize are prob-lematic.The focus on the actual users and the vividand expressive statements they give justifies themuch higher expense in certain cases. From ourexperience, the combination of both heuristic evalu-ation with the Heuristics forWeb Communication anduser testing with thinking aloud is a very usefulmethod of usability engineering.

References

Cleary,Y. (2000),An Examination of the Impact ofSubjective Cultural Issues on the Usability of aLocalized Web Site - The Louvre Museum WebSite. In: Museums and the Web 2000: ConferenceProceedings. CD-ROM documenting the ForthInternational Conference held in Minneapolis,April17-19, 2000. Pittsburgh, PA:Archives & MuseumInformatics.

Garzotto, F. / Matera, M. / Paolini, P. (1998),To Use orNot to Use? Evaluating Usability of Museum WebSites. In: Museums and the Web 1998: ConferenceProceedings. CD-ROM documenting the SecondInternational Conference held in Toronto, Ontario,Canada,April 21-25, 1998. Pittsburgh, PA:Archives& Museum Informatics.

Heuristics for Web Communication. Special Issue of theJournal ofTechnical Communication, 47 (3) August2000.

Harms,I. / Schweibenz,W. (2000), Usability EngineeringMethods for the Web. Results From a UsabilityStudy. In: Informationskompetenz -Basiskompetenz in der Informationsgesellschaft.Proceedings des 7. Internationalen Symposiums fUrInformationswissenschaft (1S12000) Dieburg 8.-10.November 2000. Hrsg. von G. Knorz und R.Kuhlen. (Schriften zur Informationswissenschaft38). Konstanz: UKV. 17-30.

Kanmer, L. / Rosenbaum, S. (1997), Usability Studies ofWWW Sites: Heuristic Evaluation vs. LaboratoryTesting. In: Proceedings of the 15th InternationalConference on Computer Documentation SIGDOC '97:Crossroads in Communication. 19-22 October 1997,Snowbird, UT. New York, NY:ACM Press. 153-160.

Keevil, B. (1998), Measuring the Usability Index of YourWeb Site. In: Proceedings of theCornputerHumanInteraction CHI '98 Conference,April 18-23 1998, Los Angeles, CA. New York, NY:ACM Press. 271-277.Also available online: Internet,URL http://www3.sympatico.calbkeevil/sigdoc98/index.html.Version: 09/98. Consulted: 12/06/00.

Krämker, H. (1999), Die Welt derBenutzerfreundlichkeit. [The World ofUserfriendliness.] In: Hennig, J. /Tjarks-Sobhani, M.(Hrsg.): Verständlichkeit und Numungsfreundlichkeitvon technischer Dokumentation. (tekom Schriftenzur technischen Kommunikation Bd. I ) Lubeck:Schmidt-Romhild. 22-33.

Levi, M. D. / Conrad, F. G. (1996),A Heuristic Evaluationof a World Wide Web Prototype. In: interactions,07/1996: 51-61.

Michalak, S. / Coney, M. (1993), Hypertext and theAuthor/Reader Dialogue. In: Proceedings ofHypertext '93, November 14-18 1993, Seattle,WA.New York, NY:ACM. 174-182.

Molich, R. / Nielsen, J. (1990), Improving A Human-computer Dialogue. In: Communications of the ACM,33 (3) 1990: 338-348.

Nielsen, J. (1993), Usability Engineering. Boston:AcademicPress.

Nielsen, J. (1997a), Usability Testing. In: Salvendy, G. (ed.):Handbook of Human Factors and Ergonomics. 2ndedition. New York, NY: John Wiley & Sons. 1543-1568.

Nielsen, J. (1997b), Severity Ratings for UsabilityProblems. Internet, URL http://www.useit.com/papers/heuristic/severityrating.hunI.Version: 01/ I 1/99. Consulted: 12/06/00.

Rajani, R. / Rosenberg, D. (1999), Usable?...Ornot?...Factors Affecting the Usability of Web Sites.In: CMC Computer-Mediated CommunicationMagazine. Internet, URL hap://www.december.com/cmc/mag/1999/jan/rakros.html.Version: 01/99. Consulted: 12/06/00.

Siegel, D. (1997), Creating Killer Web Sites.The Art ofThird-generation Site Design. Indianapolis, IN: Hayden.




Silverstone, R. (1988), Museums and the Media:ATheoretical and Methodological Exploration. In:International Journal of Museum Management andCuratorship,Vol. 7, No. 3, 1988: 231-241.

Schweibenz,W. (1998),The "Virtual Museum": NewPerspectives For Museums to Present Objects andInformation Using the Internet as a KnowledgeBase and Communication System. In: KnowledgeManagement und Kommunikationssysteme.WorkflowManagement, Multimedia, Knowledge Transfer.Proceedings des 6. Internationalen Symposiums fürInformationswissenschaft (ISI '98) Prag, 3.-7.November 1998. Hrsg. von H. H. Zimmermannund V. Schramm,V. (Schriften zurInformationswissenschaft 34). Konstanz: UKV. 185-200.Also available online: Internet, URLhttp://www.philuni-sb.de/fr/infowiss/projekte/virtualmuseum/virtual_museum_15198.htm.Version:11/05/98. Consulted: 12/06/00.

52

Teather, L. /Wilhelm, K. ( I 999),"Web Musing":Evaluating Museums on the Web from LearningTheory to Methodology. In: Museums and the Web1999: Conference Proceedings. CD-ROM document-ing the Third International Conference held inNew Orleans, LA, March 11-14, 1999. Pittsburgh,PA:Archives & Museum Informatics.

Virzi, R.A. (1992), Refining the Test Phase of UsabilityEvaluation: How Many Subjects Is Enough? In:Human Factors, 34 (4) 1992: 457-468.

56

@ Archives & Museum Informatics, 2001

Design for(re-) Use

57

A Re-usable Software Framework for Authoring andManaging Web Exhibitions

Christian Breiteneder, Hubert Platzer, Vienna University ofTechnology, Martin Hitz, University of Klagenfurt,Austria

Abstract

In the course of a Web development project for the Museum für Valkerkunde in Vienna. ViEx, a reuse frameworksupporting authoring and managing hypermedia exhibitions, has been developed.The framework consists of threemajor components: a relational content database, a corresponding browser based editing interface, and a Web pagegenerator which creates the final Web exhibition. Strict separation of content, structure and layout informationpromises ease of maintenance, especially in the context of multiple versions of the same exhibition to cope withdifferent presentation languages and client platform dependencies.

Introduction

Authoring and managing a hypermedia exhibitionis a tedious and time-consuming task requiring aset of skills not readily available in a museum con-text. In order to liberate curators and exhibitiondesigners from the computer-centric tasks as muchas possible and to reduce the resources spent forevery new virtual exhibition, we developed ViEx("virtual exhibition"), a software framework in-tended to introduce systematic reuse into the do-main of Web based exhibitions.

The framework is based on building blocks thatcover all structural parts of a Web document, e.g.title,various paragraph types,special character types,as for example, proper names and foreign wordsand media types like videos, panoramas or images.Some of these types are complex objects them-selves, since they consist of a set of more atomicentities. Layout information is available in form oftemplates that may be nested to define a morecomplex page.

All building blocks and their instances are stored ina relational database, and pages may be rendereddynamically according to their content and layoutdefinition. In addition, the framework comprisesmodules allowing for the evaluation of user satis-faction and behaviour, like, for example, guest bookand a comprehensive statistics module and a spe-cial navigation module allowing for the pictorial rep-resentation and arrangement of users' favoritepages.

5 8

ViEx has been developed and tested within theproject of a virtual exhibition on Bhutan (hap://www.bhutan.at). In principle it has been designedto be platform independent. The current serveris running under Solaris and utilizes PHP andOracle8i. The reusability of the framework is cur-rently being tested in the context of a site forArchaeology@Austria.

The remainder of this paper is organized as follows:In the following section, goals and motivation forthe development of the framework are discussed.In Section 3, the framework itself is presented. Sec-tion 4 puts the framework development into theperspective of the project history and discussessome advantages and drawbacks of the approachtaken. Finally, in Section 5, our current work relatedto ViEx is briefly described.

Goals and Motivation

Creating an exhibition web site for a museum is amajor software development undertaking which mayeasily go beyond the resources and capabilities of asmall or medium-sized museum. For example, thevirtual exhibition project, Bhutan Fortress of theGods, (http://www.bhutan.at, Breiteneder et al.,2001) which gave rise to the development frame-work presented in this paper involved a set of projectspecific roles. From the category of staff and thenumber of physical people assigned to each role(stretched to varying degrees), we can easily con-

Museums and the Web 2001:ProceedingsArchives & Museum Informatics, 2001, p. 55

Breiteneder etal., A Re-usable Software Framework forWeb Exhibitions

Role in Project CategoryProject management Development team I

Text authoring Museum 2Content advisor Museum I

Quality assurance Museum I

Translation Translators 4Screen design Designer I

Page editing Museum I

Picture editing Designer + Museum 2Video editing Development team I

3D modeling Development team I

Panorama modeling Development team I

Software design Development team 3

Software implementation Development team 5Test (functionality, usability) Development team 5System administration (development site) Development team I

Table I:Staffing overview forhttp://www.bhutan.at

clude that - at least as of today - such a project fallsoutside the main competence and self-image of amuseum.

Outsourcing as the obvious solution to the prob-lem of lack of human resources complicates com-munication during the development phase and im-pedes ongoing content maintenance after the de-velopment phase, as most maintenance requests willhave to be directed to the development team which- as a (not unlikely) worst case scenario - may noteven exist as such anymore.

In this situation, employing a reuse framework suchas ViEx (presented in the next section) may resultin the following advantages:

It enables the museum-based project team to-gether with the art designer to elaborate a com-plete specification of the exhibition site within theframework provided.This avoids the need for con-tinuous feedback from the development team, thuseliminating significant portions of the communi-cation overhead and focusing the developmentteam on its main task, namely implementing thesite according to the specifications provided.

It shifts development work to a higher level ofabstraction: instead of "coding" Web pages inHTML, the main tasks consist of editing and struc-turing content liberated from layout questions.

It supports content maintenance in the post-de-velopment phase, i.e., it enables museum staff to

59

consistently modify the exhibition without theneed to resort to the development team.

It simplifies the software development portion ofthe project due to "pre-fabricated" built-in fea-tures which can be used as-is and need not beimplemented from scratch (concept reuse). Forinstance, the interplay between images, pop-uplegends, and zoom-windows for higher resolutionversions of an image is generated by the frame-work.

It allows systematic reuse of content buildingblocks in different contexts within the same vir-tual exhibition (data reuse). Moreover, export in-terfaces can easily be established in order to re-use content material outside the project (e.g.,consider XML exports).

By all the above, it saves development time andcosts.

The main benefit, however, will be the fact that aframework enforces well-structured and consistentbuilding blocks for the Web presentation. Consid-ering again size and complexity of the project teamas sketched in Table I, any site composed of indi-vidual, more or less "hand madeWeb pages is nearlyguaranteed to end up in complete chaos with thetypical symptoms of inconsistent hyperlinks, aber-rations of page layout etc.The corresponding main-tenance overhead - in this case arising already dur-ing the development phase - will consume essentialproject resources. From the authors' experience,such effects are likely to affect even mid-sizedprojects of about 20 to 50 Web pages.

The advantages discussed above will be especiallybeneficial in the case of temporary virtual exhibi-tions where high start-up costs may challenge thefeasibility of the entire project.

In the specific case of a database-supported frame-work like ViEx, the existence of an object databasecontaining descriptions and media files for all ex-hibits needed in the exhibition under constructionwould constitute an important additional asset,as asignificant portion of editorial work could be settledby a straightforward export-import script pre-load-ing the framework's project database.


BEST COPY AVAILABLE


ViEx A Framework forVirtual Exhi-bitions

The ViEx framework consists of three major com-ponents: a content management system, its under-lying content database, and a set of layout templates.In addition, a parser was also implemented to sup-port the transition between the older file systembased framework ViEe (cf. Section 4) to the cur-rent framework operating on a relational database,by transforming Web page files into sets of corre-sponding database entries.This auxiliary componentmay also be considered a "feature" ofViEx becauseit can be seen as an alternative input channel intothe content database.

This section describes theViEx framework in somedetail. We begin with a brief summary of the fea-tures of a ViEx-based exhibition and present themain components ofViEx in subsequent subsections.

Supported Features

Any software framework designed for reusabilitywill be more effective, the better its applicationdomain is understood and the more specific theframework fits its application domain (cf. standardliterature on software reuse, e.g. Biggerstaff & Perlis,I 989).Thus,when employing such a framework, thetrade-off between generic applicability and reusepotential must be carefully considered.To this end,it is important to briefly summarize the features ofa virtual exhibition realized on the basis ofViEx. Fora more in-depth treatment see Bhutan - A VirtualExhibition http://www.bhutan.at.(Breiteneder et al.(200 I ).

The content of a ViEx exhibition is organized in alayout-independent manner. Building blocks may bepages (content aggregates consisting of other build-ing blocks), pictures with descriptive information,video, audio, text (individual text and standard textbuilding blocks), glossary entries (currently two cat-egories: foreign words and proper names), scriptbuilding blocks (e.g., fader texts), and panoramas (pos-

sibly linked to descriptive information).Table 2 sum-marizes major attributes assigned to these buildingblocks. For all text-based information, the respec-tive language is another attribute stored in the da-tabase (not shown in Table 2).

ContentBuildingBlock

Attributes

Page type of page (chapter, subchapter...),template, window title, documenttitle, color code, ...

Picture Image URI_ zoom image URL, title,legend, display size

Video,Audio,Panorama

URL, title, legend, plug-in, size,quality...

Text type (i.e., style class like header,paragraph...), content

Glossaryentry

text, category, short description(»tool tipq), long description (glossarypage)

Scriptbuildingblock

text

Table 2: Content Building Blocks

As navigation mechanisms,ViEx supports manuallydefined ad hoc hyperlinks, systematic hierarchic links(path to the current page, children emanating fromthe current page etc.) for navigating through theexhibition according to the underlying book meta-phor, local links to navigate within a page, a site mapcontaining thumbnails of the upper levels of the hi-erarchy, a hierarchy browser (available in three fla-vors: pull-down-menus, table of contents, tree con-trols), a search facility, an exhibit index, pop-up win-dows displaying descriptive information linked topictures, and zoom windows displaying pictures at ahigher resolution. In addition, there are two differ-ent tour concepts available: a slide show consistingof a predefined sequence of pictures with legendswhich are continuously presented (fading over eachother), and a guided tour taking a sequential paththrough the exhibition, briefly explaining its differ-ent parts. Last but not least, registered users getaccess to a feature called lookmarks enabling themto create their own catalog of the exhibition by dis-tributing thumbnails of pages of interest in a threedimensional space and grouping them according tosubjective organizing principles. (The three dimen-sional arrangement and the visual cues based onthumbnails are deemed to provide a much betterrecognition / retrieval rate compared to ordinarybookmarking.)

The book metaphor mentioned above serves as themajor structuring mechanism inViEx. Its strict hier-archic structure is deemed to reduce the likelihood



refersTo

Part

(ordered)

Chapter

{ordered}

Sub Chapter

HomePage Entry Page

A lNavigation Page

Web Page

4 1

(ordered)

.,#__.Content Page

Section

4{ordered

Sub Section

{ordered}

SubSubSection

Aux Window

4

Zoom Animation Site Map

Fig. Relationships between ViEx web pages (in UML)

for a visitor to the virtual exhibition to get "lost inhyperspace", especially since visitors are assumedto be able to relatively quickly construct a mentalmodel of the Web presentation due to their ac-quaintance with exhibition catalogues and the like.

The six (at most note the cardinalities indicatedby * in the UML diagram in Fig.1) hierarchy levelsrepresent the main information-bearing contentpages. Besides them, a few other, secondary vari-ants of the abstract concept »web page« exist forrather technical purposes.

ViEx supports multiple coexisting versions of thesame exhibition. Specifically, three versioning dimen-sions can be distinguished:

Static vs. dynamic pages: During the presentationdesign phase, pages generated on demand areusually preferred to static pages, because changesin the »look & feek of the site are immediatelymade visible on all available pages for personsworking on the design prototype.When the de-sign has stabilized,one may want to switch to staticpages either for performance reasons or in or-der to prepare an off-line version of the exhibi-tion on CD-ROM. Dynamic ViEx pages are codedin PHR

Different technological levels:Taking into accountthe diversity of client platforms now in use andtheir mutual incompatibilities, it is necessary tocustomize aWeb presentation to several levels ofbrowser technology.As a minimum requirement,a"high tech" version with contemporary ("cool")features and a low tech" version suited to old orlow-end browsers should be provided.

Language of presentation.

Content Database

The content database is implemented using a rela-tional database management system and containsboth the content and structure of the virtual exhibi-tion, but excludes low-level layout information.Thisseparation represents a standard design goal forcomplex Web sites, because it guarantees consis-tent presentation of the content throughout thewholeWeb site.The database oriented approach ofViEx pushes this design goal to an extreme:As nolow-level formatting information can be entered intothe content database, the separation of content andlayout follows as a consequence.

61


BESTCOPYAVAILAP E


The data model underlying the content database isgiven as a UML diagram in Fig. 2. As can be seenfrom Fig. 2; the core concept of the content data-base is the abstract entity building block which rep-resents a unit of information to be presented andmanipulated as a whole.The specialization hierar-chy emanating from classBuildingBlock distinguishesbetween atomic entities and complex entities rep-resented by containers consisting of other buildingblocks.Atomic entities may be:

homogeneous portions of text (with no inter-mingled markup) in various paragraph types andoptional special character distinction (e.g. forproper names and foreign words),

pictures consisting of a small image, a caption, amore detailed description and a reference to alarger copy

other media like videos and panoramas treated ina similar way (not shown in the figure), and

various kinds of hyperlinks.

A content page is associated with an externallystored layout template divided into so-called »slots«(class Content in Fig. 2) which are to be filled bybuilding blocks. Slots may in turn be subdivided into»subslots«,where the layout of each subslot is againdefined by a layout template. The hierarchy speci-fied in Fig.1 is resolved as a recursive parent-childrelationship between content page entities. For eachcontent page, some layout information in the formof attribute-value pairs (such as color schemes ormain titles) can be specified and is propagated downthe page hierarchy by the page generator (hencethe class name InheritanceData).Thus, every page inthe hierarchy takes this information from higher level

BuildingBlockExtras

B8No [1]Attribute[1]ValueLanguage [1]

BuikfingBlock

1..n

Content

Pageld [1.2]BBNo [1]SlotNamePosition [2]

0..1 SubSlots

InhentanceData

Pageld [1]Attribute [1]ValueLanguage [1]

PagetYletaData

Pageld [1)MOName [11bind [2]

SlotName [1]Pageld 111Template

XLink

Beld [1]URL

1..n

8810 [1]BBNo[2]Language [2]Category

..1

targetBEI

Link8810 [1]

ILink &inkBEild [1] EIBIci [1]

TargetPageld Picture8BldTargetBeld Status

targat

refersTo

ContentPage

..n

kletaData

MDId [1]NIDContent [1]Language [1j

fxor).

1

Pageld [1]TemplateTypeParentldPosition

child In

parent

0..1

BuildingelockkletaData

1..n 0..1

BBNo [1]MOName [1)tfiDld [2]

IV*

0..

ca0..leg nd

Container

BBlci [1]Type

-411

0..1

ContainerConts

BBId [1]ContainerBBId [2)Position [2] SpecialTextInfo

SpecialText (1)Language [1)Type

TextBEI

121810 [1]

CharactersType

VariantsExplanationExplanationDetailsFacsimile

Picture

BBId [1]Caption88NoLegendBBNoHeightWidthZoomURLStclURLThumbnailURLForm

0.1

Generalizations andassociations are shown forclarification purposes onty. -all "Inheritecr attributes andreference attributes areexplicitly specified

Fig. 2: Data model of the content database (in UML with database key designators in brackets)


BEST COPY AVAILABLE

59

pages, but it can also be assigned corresponding dataof its own which overrides the »inherited« default.

When the exhibition is generated from the con-tent database, the resulting page format is deter-mined by

the position within the page hierarchy at whichthe page resides which affects the inheritancemechanism explained before and the generationof structural navigation links,

the layout template assigned to the page,

templates possibly assigned to the slots of thetemplate,

a cascading style sheet (which is dynamically se-lected depending on the client's platform),

the formatting strategy encoded into the genera-tor for each type of building block,

as well as additional layout information assignedto selected building blocks (classBuildingBlockExtras).

Front end (input and editing interface) and backend (pagegenerator) of the content database to-gether form the content management system de-scribed at the end of this section.

Layout Templates

Layout templates reside in the file system (outsidethe content database) and define the principal lay-out of each type of page. Changing a layout tem-plate modifies the look & feel of the exhibition with-out modifying its content and logical structure.

Templates are »stub pages« with fixed HTML, in-cludes etc. combined with special markups desig-nating slots to be filled from the content database.The syntax of these markups is

<! #BeginEditable "Slot Id" >

and is taken from Macro Media Dreamweaver inorder to allow employing Dreamweaver as a com-fortable template editor.The generator replaces thismarkup with the building blocks selected from the

6 3

database via slot identification Slotld and the identi-fication of the page under construction.

Besides the content oriented slots mentioned, thereare also so-called "technical slots" with predefinedsemantics which are used to designate the positionof standard page elements like headings and naviga-tion blocks within the page.

For instance, a typical template taken from http://www.bhutan.at contains the following slots:

a) Technical slots

Page title: The browser's window title. Contentexplicitly given in the content database.

Page meta-information: HTML meta information (in-visible). Content explicitly given.

Script: Includes client side scripting files (e.g.,JavaScript) with functions taking care of CSS se-lection according to the client platform, windowmanagement, and pop-up handling. Content gen-erated.

Document title:Text or image. Content explicitlygiven.

Color scheme: Color code or image. Content ex-plicitly given or inherited from ancestor pages.

Path: Navigation element consisting of titles of allancestor pages. Content implicitly given (takenfrom ancestor pages).

Navigation (»nav«, cf. the table below): One of twopossible types of hierarchical navigation menudepending on the type of page. Content implicitlygiven.

Pop ups: Place holder for pop-ups used as imagelegends. Must be last element in the template.Content generated.

b) True content slots

The standard layout for »Bhutan Fortress of theGods« is based on the following HTML table (withslot names):

row I -left row 1 -center

row 1 -right

row 1

row2-left row2-center

row2-right

row2


row3-right

row3


row4-right

nav

Table 3: Content slots

All these slots may be assigned simple text buildingblocks, containers holding nested building blocks,or subslots with corresponding subtemplates.Thislatter feature enhances the range of application of atemplate:Instead of designing completely new tem-plates, many variations can be incorporated into oneand the same template by means of the subtemplateconcept.

For technical reasons,ViEx only supports one levelof subtemplates; i.e., subtemplates nested withinsubtemplates are not possible. However,for, the vir-tual exhibition »Bhutan Fortress of the Gods« thisrestriction did not impose any undue design limita-tions.

It should be noted that all content which dependson the page type only, i.e., which is identical for allpages using the same template, might also be keptin the template itself (and not in the content data-base). However, such an approach may jeopardizeimportant features of the framework such as cre-ating a presentation in a different language etc.

Content Management System

The content management system (CMS) consistsof the content database input and editing interfacesand the page generator. It allows creation, mainte-nance and deletion of pages. In its current version,it supports ODBC, Oracle's OCI and MySQL data-bases, thus covering all major production platforms.

The database input and editing interface is browserbased and provides features to manipulate content,structure and layout of the web presentation un-

der construction by modifying respective databasecontents. Fig. 3 shows a screenshot of the hierar-chy editor based on tree controls and an editingmenu. This component is used to navigate and/ormodify the hierarchic structure of the exhibition.From here, the main content edit window may bereached. In Fig. 4, building blocks I, 2, and 3 of Con-tainer 1 I (itself an aggregate building block of type»baui(<) of page »lnfrastruktur Baudenkmalera ofa German web page are being edited.The structureof the edit window changes dynamically dependingon the type of content edited.

a Explorer - Contenthianage..._ NEE

Arch test cms v220 HomeCHTeitl hubertC)Teit 2 hubertC3Teit Martin

Arch CMS Demoutturgeschichte

er NorcitSonstiges edit

properties-----

deletecopy

structuredata

move pathmove page

add child

Fig. 3: CMS: Hierarchy editor

Fig. 4 illustrates not only the look & feel of the userinterface for content entry but also the main disad-vantage of the required pseudo-atomicity of thecontent database:As there is no formatting allowedwithin text building blocks, the granularity of textbuilding blocks is somewhat smaller than mere re-use considerations would require. For instance, astext portion "Bauperiode 2" is used as a hyperlinkto page titled "Das Bad des Auxiliarkastells vonCarnuntum", it must be entered as a separate build-


BEST COPY AVAILABLE

61


aecht)Andoerk Infteettuktm-Baudenkmiie- Micros ig El

uPdat

C doge 8)_.,1_.'

sl Das APOOYIERIUM dont alsWON idetaum und utokletOraufbeoshoung odbeend doBadetundg ands. Ea keen. vd, beim 0/ update

arnuntiney Kotellbad ab j C delete

Fro3 Bsupenode 2

Des Bad des Auxiliarkostells von Cornuntum

use Totbaustain as lir. I rtind I .Link Icontain., I Tonle Inope html I

gr-73:1 [ Job petted, luerden.

Internet

51. update

C desta 5221:Lind

weat.

C dew.

Fig. 4: Editing a page of Archeology@Austria

ing block (no.2), although from a semantic perspec-tive, building blocks 1,2, and 3 together form a logi-cal unit (i.e., a paragraph).

As mentioned before, it is possible to associate somelayout information will different types of buildingblocks which will be taken into consideration bythe page generator.The corresponding CMS mod-ule (style editor) is shown in Fig. 5.

1 0..... %.... F.... Lt. . a'i '.. "' 4 a) .......le4,,,,.... . ......2.1/5...s.d... ,t1 0,....

Dv. Lp, vu we I SLAM Vsrmn

bine 1....gr...r-- F F .WZ*M12?2,$I&j.*gQ. A.11

k.....Fi F i 1 .triI. F P-777 I . 1 141

F F-7 I...-..- . 1,.. . mk ri...--,, ..I. Ir g-- fr....31,00E.0..tel......E......11nr rrem.Vadiddrogyelild.y..x. f.r,

F. T---1.....,............-.......-,..,...0- 1,...,.......um .1.41

le., r -.--1,.....,,...,..P......enyvZcan:Win.'Stnoarr8904181 IML.091.11.... I.., *Al.... r.r.---.-- ,,...

. Om . f-r14*,Fig. 5: Editing styles associated with

building block types

Columns "tag" and "vars" define the HTML elementto be generated for building blocks of type "type"."vars" contains variable references ("$ 1","$2", etc.)which are to be substituted by corresponding data-base attributes listed in column "db-vars". These

62

attributes are found elsewhere in the database. (Col-umn »pos« is relevant for nested blocks only.)

Last but not least, a preview window allows theuser to immediately see the outcome of the pagegenerator run (Fig. 6). In such a window, special iconsoffer links back to specific edit windows (edit pic-ture, edit container, edit text building block, editstyles, edit template slot. Cf. the bracketed icons inFig. 6 clicking on the one with the tool tip dis-played would open the picture editing window tomodify the presentation parameters of the imageof Guru Rinpoche).

TIVrMU...4a 01 in Mu, (1.

Math..... M.o....M.. :O.... Pr.ItreA

11:11 0:11

Ots bend ab...1*. dars 1Mot

mb, es 4** Lein palt:....en:

et. Mum lat

([3]IM MUM Mda. ad. NO

1

I .4400.. PSI bonorrscH....4

IIbed., eltss 8.14Mssr. met datErricte.M..1. Mmelreroontuo wad.. tune. 0:

moan.. io.e.

gr)

loomouieder.opmeopocomej*.ekonareele;eOWPAP.OILI t troo.

Fig. 6: Preview of generator run

The page generator represents the back end of thecontent management system, producing dynamicWeb pages (a static version can be generated ondemand in a post-processing step) according to theformatting rules given earlier. It is responsible for ahomogeneous treatment of the different kinds ofbuilding blocks (e.g., generating the code necessaryfor the interplay between images, captions, corre-sponding pop-up legends and zoomwindows), andconsistent navigation scheme, both within a pageand within the web site (for deep hierarchies, man-aging the complex navigation menus is in itself anontrivial task which is hard to accomplish manu-ally, especially when the structure of the hierarchyis changed).

The generator is also capable of creating differentversions of an exhibition as discussed in Section 3.For example,"Bhutan Fortress of the Gods" (http://www.bhutan.at) exists in German and English, bothin a "high tech" and "low tech" variant.

BEST COPY AVAILABLE65



Project History and Reflection

The project which gave rise to the development ofthe framework presented in this paper begun as aresearch project to explore different approachesto virtual exhibitions, funded by the Austrian Fed-eral Ministryof Education, Science and Culture(Breiteneder et al., 2000). However,as a by product,a production release of a virtual variant of a well-known and internationally successful exhibition ofthe Austrian Museum furVolkerkunde on Bhutan hadto be developed. In this situation, the project part-ners opted for a dual approach:A conventional de-velopment track based on manually created dynamicWeb pages was chosen in parallel to the develop-ment of the more advanced database based infra-structure presented in this paper (ViExOB or ViEx,for short).The first track was needed to guaranteea working product, while the second track was aimedtowards a reuse environment for future virtual ex-hibitions.

During the development of the Web site along theconventional track, strong emphasis was laid on awell structured project and software design suitedfor a mid-sized project.Thus, after a short explora-tion phase, a set of tools (PHP3 as scripting lan-guage, homesite and BBedit as web editors, m4 asmacro processor to introduce parametric text sub-stitution etc.) to be employed alongside with cor-responding strict coding and naming conventionswere established, and the file system based versionViEx'S of the framework was born.When its designhad matured so far that a successful completion ofthe development project could be foreseen, all struc-tural decisions made were frozen and the databaseoriented approach was started in parallel. From themodeling decision made thus far, the databasescheme was deduced and the page generator wasdeveloped. Meanwhile, conventional developmentof the web site was progressing within the ViEx'sframework, so that a content transfer tool wasneeded to initialize the database content from theweb pages in ViEx's.The development and employ-ment of a respective page analysis tool helped toverify and optimize the database model and alsopointed out some inconsistencies within ViExFswhich were corrected on the fly.Then, the editingfront end was implemented as a working proto-type,and finally the wholeWeb site was loaded intothe database version ofViEx and manually polished.

From this dual development experience,we are nowable to contrast the two approaches taken.

In the file-system based version of the developmentframework, each building block was realized as afile in directory tree with strict localizing and nam-ing conventions. The unit of work was defined asediting such a source file which usually took placeon a client followed by some kind of data transfer(e.g., ftp) to the web server.With a Web-based ed-iting interface, some edits could also be done di-rectly on the Web server.Thus, a maximum of con-tent creation flexibility was achieved,and all file sys-tem tools available on the server (under Unix) couldbe used (e.g.,access right configuration,global searchand replace etc.). In particular, pages systematicallygenerated via PHP3 and m4 could easily be com-bined with some hand crafted individual pages wherenecessary. Also, authoring and configuration man-agement activities could evolve with the complex-ity of the web site under construction.

The disadvantage was that many interfaces had tobe mastered by the editing staff and that there wasno real control whether all conventions establishedby the project team were observed (as the mainportion of the content editing work was done by asingle person with sufficient self-discipline, this as-pect did not create a lot of problems in our project).Moreover, data transfer from the client worksta-tion to the web server was not secured.

In the database version of the framework, contentwas moved from individual files into tables of a re-lational database system. The corresponding con-ceptual model guaranteed clear separation of con-tent, structure and layoutThere is only one editinginterface (cf. the previous section). Modest use ofrelational features allows the employment of a widerange of DBMS from mySQL to Oracle 8. As theediting interface is somewhat more formal than forViEx's, there were fewer accidental changes of con-tent. On the minus side, there is the higher com-plexity of the editing task (due to the atomic na-ture of data), even though content editing perfor-mance seems to constitute a critical success factorfor such a project We also had to realize that latemodifications of the underlying conceptual modelcost a relatively great deal of work (an experiencewhich seems counter-intuitive to classical argumentsin favor of database technology).Thus, the database-



oriented approach seems to be especially promis-ing if:

the structure of the planned exhibition is com-patible to the framework, i.e., it can be mappedonto the conceptual model of the framework,

the exhibition project exceeds a certain complex-ity which allows amortization of setup costs (in-stalling and getting acquainted with the tools, etc.),

all tools are already well-defined and ready forduty at the beginning of the project,

an exhibit database exists which may be linked tothe content database presented here.

Conclusions and Ongoing Work

The primary project goal to develop http://www.bhutan.at to a fully functioningWeb site meet-ing all requirements of a production system has putrestrictions on the development of the prototypeof a reuse framework discussed in this paper. Wewill definitely need at least resources equal to thosespent so far on its completion to be able to use itin full practice and in a wide variety of contexts.Since reuse of software can only be proven by re-use, we have meanwhile tested the framework in afeasibility study in the field of archaeology. Becausethe participating archaeologists had different re-quirements from the ones to be met for http://www.bhutan.at, minor adjustments were necessaryto increase the functionality available.

Currently, we are extending ViEx to be preparedfor a much larger project covering all Austrian ar-chaeological museums and sites in one Web portal.Since the total number of Web pages of this sitewill be a magnitude higher than for Bhutan, we firsthave to evaluate and improve the performance ofViEx for sites with more that 1000 pages.The mostimportant additional extensions ofViEx will cover:

6 7p.

support for XML

support for metadata (RDF and Topic Maps) inorder to move an important step towards thesemantic web

support for multiple concurrent users and

the enhancement of editing features, as for ex-ample the splitting of building blocks.

Moreover, we want to compare available contentmanagement systems and to include some of theconvincing additional features in our system.

References

Breiteneder, C., Hitz, M., Hon, M., & Platzer, H. (2000).Untersuchung innovativer Hypermedia-Methodenzur Gestaltung virtueller Ausstellungen am Beispiel"Bhutan Festung der Gauen http://www.bhutan.at. Project report,Vienna Universityof Technology/ University of Klagenfurt / Univer-sity of Vienna. October 2000.

Breiteneder, C., H itz, M., Plum: H., & Stockingerd.(200 I )."Bhutan A Virtual Exhibition". Museumsand the Web 2001. Archives & Museum Informatics,2001. http://www.archimuse.com/mw2001/

Biggerstaff,T. J. & Perlis, A. J. (1989). Software Reusability(Vol. I: Concepts and Models,Vol.11:Applicationsand Experience).ACM Press.


Design and Analysis ofVirtual Museums

Gilles Falquet, Jacques Guyot, Luka Nerima, University of Geneva,Switzerland and Seongbin Park, Information Sciences Institute, USA

Abstract

Using the same data, which could come from local databases or external sources such as the Web, virtual museumdesigners can build different hyperspaces. It is possible that visitors would find some of them more useful than others.Therefore,virtual museums designers should be equipped with a tool by which various hyperspaces for virtual museumscan be easily designed and examined.

In this paper, we view a virtual museum as a hypertext that consists of nodes and links and show that a databasepublishing tool called Lazy, which generates a hypertext view (i.e., derived hypertext) of a given database, can be usedfor designing virtual museums. The Lazy system consists of a declarative hypertext view specification language, a nodeschema compiler, and a node server that processes node requests. Since the language is purely declarative, it is fairlyeasy to construct and revise hyperspaces for a virtual museum. With this tool it becomes possible to adopt aniterative design methodology.

Given a database for a virtual museum, we first construct a hypertext using the procedure (Falquet & al., 1999) calledan initial structure. We then proceed to analyze the initial structure and examine possible refinement operations thatcan enhance the usability of the created hypertext. For that purpose, we use a simple graph-based analysis and weshow kinds of analysis that can be done using the graph-based approach. Once the structure is refined using therefinement operations, we apply grammar-based formalism (Park, 1998) to the refined structure to see whether wecan obtain a simpler grammar that can generate the same hyperspace. Our goal is to explore various analysis techniqueson the hypertext and give insights into designing a good hyperspace using the analysis results.

Introduction

In this paper, we consider the issues that arise inthe design and implementation of virtual museums.A virtual museum is defined as "a logically relatedcollection of elements composed in a variety ofmedia, and, because of its capacity to provide con-nectedness and the various points of access avail-able, [it] lends itself to transcending traditionalmethods of communicating with the user; it has noreal place or space, and dissemination of its con-tents are theoretically unbounded" (Andrews,1996). According to this definition, it is reasonableto view the structure of a virtual museum as ahypertext since each element in a hypertext canbe connected to others. However, it is well knowthat large hypertexts, in particular large Web sites,are very difficult to manage. Thus it has becomecommon practice to store data in a database andto use some mechanism to automatically producethe hypertext.

In this paper, we will study the construction of vir-tual museums on top of databases.The content ofa database can represent either real-world entities(e.g. existing works of art, or existing museums) orvirtual entities (e.g.virtual exhibitions that exist only

on theWeb). When constructing a virtual museum,it is important to provide visitors to the virtualmuseum with a well-designed hyperspace so thatthey do not get lost while navigating inside the space.Although"lost in hyperspace" is a well-known symp-tom (Conklin, 1987), in our case, we have an under-lying database that is structured by a databaseschema and we can avoid the symptom to someextent.

Given database contents, different hypertext struc-tures can be created for the virtual museum, andvisitors will find some structures more accessiblethan others. Therefore, the structure should becarefully designed, and it is important for virtualmuseum designers to easily define nodes and linksthat form hyperspaces and see what the results looklike (examine different hyperspaces freely). In or-der to help designers who build virtual museums,we present a database publishing tool called Lazy(Falquet & aL, 1998), by which different hyperspacescan be easily constructed. Using the Lazy system,designers can construct the web site for a virtualmuseum, and once the web site is created, they cananalyze the structure of the virtual museum.



BEST COPY AVAILABLE

Falquet et al, Design and Analysis of Virtual Museums

Some of the related works are as follows: a visualgrammar-based formalism was introduced in orderto analyze hypertext structures in (Costagliola &al., 2000), and an implementation in Prolog for ana-lyzing hypertext that contains conditional linkagewas sketched in [HT99 paper]. Recently, databasepublishing has attracted interest and some of thetechniques are described in (Entin & al., I 998;Toyama& al., I 998). There are several approaches to cre-ate hypertexts from databases. In the proceduralapproach, the hyperspace designer must write pro-grams (CGI programs in C or Perl, Java servlets,PHP scripts, etc.).These programs are generally largesince the code must contain both tags and program-ming constructs.They are therefore difficult to read,and the hyperspace structure is hidden by the pro-gramming constructs. For these reasons, such a codeis tough to maintain and update.The dynamic docu-ment approach consists in extending some docu-ment mark-up language (such as HTML) with spe-cific tags for database querying, result processingand formatting, etc. See, for instance, Cold Fusion(www.allaire.com).These tags introduce proceduralparts into the document description.The declara-tive approach (Fernandez & al., 1998), (Atzeni & al.,1998), (Falquet & al., 1998) consists in specifying ahypertext structure and specifying how to build thehypertext elements from the database content. It isconceptually simple and tends to be closer to theinformation designer's conceptual level. Our workis also related to hypertext design methodologiessuch as RMM (lsakowitz, 1995) and HDM (Garzotto,

66

1993). It also takes into account adaptive featuressuch as content adaptation, a well-known adapta-tion technique in the field of adaptive hypermedia(De Bra, 1998).

The paper is organized as follows: in the next sec-tion, we will introduce the concept of hypertextview with examples constructed with Lazy. Then,we propose a design process for hypertext viewsand show its application in the design of a virtualmuseum. We proceed to explain possible analysison the hypertext. Finally, we conclude the paperwith discussions about our approach for the designand analysis of a virtual museum.

Constructing Hypertext Views withLazy

This section presents the Lazy hypertext view speci-fication language.The presentation is rather infor-mal and based on examples from the sphere of vir-tual museums.

A hypertext view is a set of nodes and links thatrepresent (a part of) the contents of a database. Inthe declarative approach, the hypertext components(nodes and links) are derived from the databasecontent (relation tuples) according to a hypertextview specification, as shown in Figure I.

A hypertext view specification consists of a set ofnode schemas that specify the collection from which

Database(relations)

Hypertext viewspecification

NflHypertext view(nodes + links)

Fig.! : Generating hypertext views from a database and a hypertext view specification.

6 9@ Archives & Museum Informatics, 2001

the node's content is to be drawn; the selectionand ordering criteria; the elements that form thenode content; and links to other nodes. A nodedefinition takes the following form:

node <node-name> ( <parameters>pre <element-list>items <element-list>post <element-list>

from <collection>, ...

selected by <expression>ordered by <expression>

An element of an element-list is an expression ofthe form, '<' type '>"(' <element-list> T or a<simple-expression>. In the items part, a simpleexpression may involve literal constants (string, in-teger, etc.), attribute names, parameter names, op-erators and functions.An atomic element, specifiedby a simple expression, represents basic documentdata (CDATA in XML terms). In the pre and postpart, attribute name may only appear within aggre-gate functions like minO, maxO, sum0, etc.

work

Although the node specification language is genericenough to support relational as well as object-ori-ented databases, in this paper we will only considerthe relational case since most existing databasesare relational. In this case, the <collection> of a nodeis a relation (or the cartesian product of severalrelations) and each node item will represent a se-lected tuple of that relation.

Example

Throughout this paper we will use the same data-base schema, shown in Figure 2, that represents apart of a virtual museum database. Note that thisstructure is close to a (simplified) real museumdatabase schema. In fact, some of the relations (e.g.works) could be existing relations of a museum in-formation systems, while others (e.g. exhibitions)could represent real as well as virtual objects.

exhibition

(exno title, desc, isvirtual, organize)

ex-content

Cwork, exhibition, comment)

museum

Enna name, location, URL

E./no, author, title, c_date, height, width, picture )-41--------------cwork, owner, acquisition )

artist

name, birthdate, deathdate

ownership

art_cnty

(artist, country, activity)

Fig. 2: Museum database schema

Consider the following node definition:

node Artists_after[date]items(name, "(born ", birthdate, ")

href Works_by[ano] ("works")

from artistsselected by birthdate >= dateorder by name

This is intended to present lists of artists born af-ter a given date. The content of an instanceArtist_after[d] of this node is computed as follows:

all the tuples t of table artist that satisfy t.birthdate>= d are selected

an item is generated for each selected tuple, itcontains an element of type that is made ofthe artist's name, the text "( born ", the artist'sbirthdate, and the text")".

This element also contains a reference link (works)to a node works_bgano] which is intended to dis-play the list of works of this artist.This node is de-fined as follows:

node Works_by[artist]items

('cimg src=picture>()),<p(<b(title), " ", c_date,

 (),support," ", height," x

",width

from workselected by author = artist

Each selected work will be displayed as twoparagrahs (), the first one showing an image ofthe work and the second one giving textual infor-mation (title, creation date, etc.). Figure 3 showsthe content of an instance Artist_aftel I 9001 and aninstance Works_by[...] that can be reached by fol-lowing an href link.

meb.u.e.bileuthriett.albritlen G

7*--Char. Mak,' ,1.13 IMS1 LWOW

Gemges.Pene Sara tlaarn 13591 (Iota

Tsar= Marsh 023111 357351135w

Ks. schweErg Wam 1=7) f.:31111

Pablo P.s1.1.11 135I11.7331:11

POIT &wad Nan Mil lmakil

P3, Mona= (born ors Ivarta

Ram Eltlaray tom 16351

Vox. visa Gael (bran 13531 Walt

togel

19357)°dm emvatO M ortml1AM

Fig. 3: Two node instance generated with theLazy system.

The node definition language supports three kindsof links, which are "reference", "expand in place",and "include". A reference link creates an activeelement whose action (when activated by a mouseclick) consists in jumping to (opening) the referredlink.A link specification refers to a node through itsidentity (schema name together with parametervalues). An inclusion link creates a compound-com-ponent relationship between two nodes.The con-tent of the included node is a part of the content ofthe parent node.With inclusion links one can con-struct arbitrarily complex nodes, for instance torepresent complex structured documents. Figure 4shows an instance of the following node schemathat includes three other node instances (Coun-tries,Work_list, and Contemporary (defined in sec-tion 3.2)):

node Artist_ext[id]items

<112( name, " (", birthdate, "-deathdate, ") "),

include Countries[id],<h4("Some works"),<blockguote>(include

Work_list[id]) ,<114("Contemporary with:<blockguote(include

Contemporary[birthdate, deathdate])from artistselected by ano=id

") ,

Vincent van Gogh (1853-1890)lived in Netherlands

moved to France

Some works

The Starry Night.1889 [descriptionj

The Siesta Dec 1889-Jan 1890 [description]

The Iris May, 1889 [descnotionl

Road with Cypress and Star 12-15 May, 1890 [description]

Contemporary with:

Pierre Bonnard Vincent on Gogh Georges-Pierre Seurat Claudeidenet

Fig. 4:An node instance which includes othernode instances

Finally, an expand-in-place link is an inclusion linkthat defers the inclusion until the user activates thelink. The content of a node with expand-in-placelinks will thus depend on user actions taken so far.

7168 ©Archives & Museum Informatics, 2001 BEST COPYAVAII,i3


The hypertext view generation system is composedof :

a node compiler that checks the syntax of nodedefinitions and stores the node definitions (in acoded form) in the data dictionary

a node server (a Java servlet) that receives noderequests from clients' (browsers); loads the appro-priate node definitions; executes database queriesto build the node contents; and sends the resultingWeb pages to the clients.

The Web site development cycle consists in writingor editing source files that contain node schemas;compiling the definitions; and viewing (testing) thenewly defined nodes in a Web browser. Since thesystem is dynamic, once a node definition has beenmodified and recompiled, the new version is imme-diately available to the clients (there is no site gen-eration phase).

Every page that is viewed by a client is an instance ofa node schema; thus any design problem can bereadily located (as opposed to procedural approachesin which the same procedure may be used to man-age several different Web pages).

Design

In this section, we explain how one can construct aninitial hypertext structure that reflects the structureof a given database and how that initial structure canbe modified through refinement operations. We alsopresent techniques that hypertext designers can usein order to implement an adaptive feature with theLazy system.

Designing efficient and effective hyperspaces is a dif-ficult task, probably because there are an extremelylarge number of paths that user can follow. It is thusdifficult to ensure that the users will be able to reachany information node, that they will not get lost ordisoriented in the hyperspace, that any informationcan be reached within a reasonable amount of time/number of clicks, etc. Since we are starting from anexisting database, we already have a conceptualschema, declared by the relation schemes and theintegrity constraints such as foreign key constraints.This schema shows the type of entities that are be-ing considered and some semantic relationships (ma-

terialized by foreign key constraints) betweenthese entities. However, a database schema is notsufficient to create good hyperspaces. Databasedesign and hypertext design do not have the sameobjectives. If we rely on the database schema atthe semantic level, it will be possible to create ahypertext structure that is efficient for readingand navigating. Our design method for hypertextviews proceeds in two phases: I ) define a firsthypertext structure based on the database schema;2) refine this structure by applying various opera-tions to the specifications of nodes and links.

Initial structure

For the construction of an initial structure, weassume that the database schema is given and fixed.One obtains the initial structure by defining a nodeschema for each relation of the data base.An in-stance of such a schema is intended to representa single tuple of the relation. The node schemahas a single parameter that represents an objectof the class (in the relational case it is a set ofparameters that forms a key value). The contentsof the node items are formed of all the collection'sattributes.

Links are formed by attributes or groups of at-tributes that refer to other relations (foreign keys).For instance, the initial node schema correspond-ing to the relation

is

relation work(wno, title, date,author, _)

node Work[w]items a wno, title, date, author,

href Artist[author]from work selected by wno = w

This structure accurately represents the contentsof the database, i.e. the graph of all node instancesand possible links is isomorphic to the graph ofthe database objects connected through the ref-erence attributes. However, this structure is notcompletely navigable, i.e., due to theunidirectionality of links, it is not always possibleto reach any node from any other node. Thusreverse links must be added.

For example, if a node N has a reference (hrefM[r ]) to a node M, we add a link from M to N

7r)©Archives & Museum Informatics, 2001 69


which corresponds to the traversal in the oppositedirection. To carry out this operation, an interme-diate node schema is defined, and this intermediatenode plays the same role as selection menus in sys-tems that support links with multiple ends.

Note that the initial structure together with re-verse links yields a fully navigable view of the data-base.This means that if two objects o I and o2 areconnected (directly or indirectly) in the database,there exists some path in the hypertext to go fromthe representation of o I to the representation ofo2 (and vice versa).

Ftefinernent operations

Refinement operations are intended to improve thenavigability (or the legibility) of a given hypertextview.We list below some of these operations.

Link composition (short cuts)

One way to reduce the number of navigation stepsin the hypertext view is to create "shortcut" links.This consists in combining two (or more) links intoa new one.This is particularly useful to increase thenavigability of the initial structure that typically con-tains nodes of the form:

node A...

from T

node B

from T

[...]href B[key_attribute]...

[k]

href C[x]selected by key_attribute . k

where key_attribute is a key of T, In this case, anyelement of B (e.g. the link to C) can be incorpo-rated into A to suppress a navigation step throughB.

Inclusions

This operation consists in changing a reference linkinto an inclusion link. It allows us, for example, torepresent complex entities in the form of only onenode (including sub-nodes).This operation is par-ticularly interesting when the link has semantics ofthe type "part-of "or "compound-of". It is also away of reducing the number of reference links inthe hypertext and thus shortening navigation paths.

Summarization

When a node represents a large object having manyattributes, it may be desirable to derive a "summa-rized" node by removing certain attributes of theinitial definition.This summarized node will have alink to the complete node. It also should be de-cided for each link that leads to the initial node if itis necessary to "redirect" it towards the summa-rized node.

Adding computed links

The database schema usually represents relation-ships between entities through foreign key con-straints (or referential constraints). However, someinteresting relationships are not represented directlyin the database schema. For instance, the relation"contemporary" between artists is not represented,but it can be computed since we know the birthand death dates of the artists. Links correspondingto such derived relationships can be created in thehypertext schema using diverse schemes. For in-stance,a relationship "contemporary" between art-ists can be implemented by creating a node

node Contemporary[abirth, adeath]item name, href Artist[ano]

from artistselectedy by deathdate > abirth+15 orbdeath > birthdate+15

and adding a new link in Artist

node Artist[id]items ...href Contemporary[birthdate,

deathdate] ("contemporary artists")

Widening

The widening of a node consists of weakening itsselection condition. As a consequence, other ob-jects will be shown in the node. This is a way tocontextualize information by presenting it togetherwith related information. For example, a paintingcould be presented together with other paintingsof the same period or of the same region.

Previewing

Previewing makes it possible to see part of the con-tents of a referred node without having to traverse



the link. The objective is to avoid navigation to anode whose contents do not correspond to theinformation we are seeking.This operation consistsof creating a summarized node (in general with onlya few attributes),as in the derivation operation,andadding to the initial reference link by including thesummarized node.

Building indices and entry points

An index structure is a set of nodes that allows us,by successive selections starting from a root node,to reach a particular node.A simple, concrete, caseis the creation of an index on an attribute A. Thisrequires the creation of a two node schemas: 1) aroot node presenting all the possible values of A; 2)a node presenting a list of all the objects having thesame value for attribute A. One can generalize thisstructure to create indices on several levels whereeach level corresponds to a different attribute.Thetraversal from the root downwards amounts to fix-ing an attribute value at each stage.

Creation of linear paths

This operation creates links that make it possibleto traverse all the node instances of a schema in aprescribed order (guided tour).

Content_of_exh

Designing Adaptive Nodes and Links

Adapativeness in hypermedia systems consistsmainly in taking into account the user's profile whendeciding on what information to display, how to dis-play it, and how to react to user actions. In the con-text of hypertext views, this means that the con-tent of a node, and its links, should be generatedaccording to a user profile. It can be implementedin a straightforward way, provided

some "profile" relation contains suitable informa-tion about the user profiles,

a global variable USER exists that stores the username (in the current implementation it is repre-sented by a supplementary parameter in eachnode schema).

For instance, the following node schema displaysinformation about a particular work of art. Itincludes a Details node that will present moredetailed information, but only if the user profilehas detail_level > 2.

Exhibition_index

exno, title

Exhibition

"contenr [exno] title, desc, Organizer

I[exhibition], work, comment

Work

Works_in_museum

L[owner] work I."

[wno], autho , title, ... picture,

Artist

"orga izer"

"exhibition 5

Museum

Imno, name, location, URL

Owned

[work] owner

"other works by this artisr

Works_by_artist

wno, [author], title

ano, name, birthdate, deathdate

"contempora artists

Contemporary

I[birthdate, deathdate], name

reference finks

[artist], country, type

inclusion links

Fig.5:A hypertext view structure after several refinement steps.

74© Archives & Museum Informatics, 2001 7 1


node Work[n]items title, c_date, include

Details[n],from work selected by wno = n

node Details[n]items width, height, acquired, ...

from work, profileselected by wno = n and

profile.user = USER andprofile.detail_level > 2

If the user has detaillevel + 2 the selection condi-tion will be false for every tuple and thus the nodewill remain empty.With the inclusion mechanism, itis thus possible to create contents and links thatdepend on user profiles or on other contextual in-formation such as time, date, etc. (for instance,forth-coming exhibits could be announced in some nodesduring the weeks before their opening).

Path-awareness is another form of adaptiveness, andconsists of having node contents that depend onthe user navigation path. Current Web browsersoffer a limited path awareness feature that consistsof displaying anchors of previously visited nodes ina particular color.Although very simple, this mecha-nism proves efficient when exploring a new site. Infact, we can think of many situations in which wewould like to have the content of nodes depend onpreviously visited nodes.For instance,we could havean anchor"latest news" in the heading of every node,as long as the "News" node has not been visited.Once is has, this anchor should disappear from allthe nodes.

In order to implement this type of behaviour, weneed some way to refer to the navigation history.The navigation history can be stored in a HISTORYparameter added to every node definition. Eachnode can then add its own identity to HISTORYand pass it forward to the nodes it refers to.Thegeneral schema is thus

node N[_parameters_, HISTORY]

href AnotherNode[ parametersHISTORY + "(N)"]

selected by conditions on attributes

AND condition on HISTORY

72

Analysis

Graph-based Approach

To analyze hypertext structures and see the effectof refinement operations, it is convenient to have acompact graphical representation. Our analysis isdone on the node schema instead of node instances,and graphical representation is smaller than thegenerated hypertext (i.e., the set of node instancesand links among them). Based on the analysis ofthe initial structure for the virtual museum, we canapply appropriate refinement operations.

Once we have a graphical representation of thenode schemas, these are possible analyses that wecan do on the initial structure.

I ) Identify links that do not exist in the initial struc-ture, but might be helpful if we created them.Notice that those links that exist in the initialstructure are directly come from the associationsof data in the database (e.g., foreign key). Aftersuch links are identified, we can apply appropriaterefinement operations to the initial structure.

2) Check the number of navigation steps betweenthe nodes and determine the semantic proximity(or semantic distance). If they are semanticallyclose, the number of navigation steps should bereasonably small.

3) Explore different types of links between the nodesand select a better type than others; for example,it might be necessary to change "jump" to "in-clude" or "expand in place".

4) Check whether a node is reachable from a givennode (i.e., accessibility analysis). Notice that theschema connectivity does not ensure that thehypertext itself (the node and link instances) isfully navigable. This depends on how objects areinterrelated in the database. However, knowingproperties of the links (like cardinalities), it is pos-sible to prove the full connectivity of the hypertextview.

5) Compute the maximum number of steps that auser can follow from one node to another. Sincethe user does not see the entire hypertext, it isnot obvious to the user whether a path is the

75


shortest distance from one node to another. In-formation about the longest path that one cantake to reach a destination node can inform usabout how a user might get lost in the hyper-space.

Grammar-based Approach

After we determine the refined structure for a vir-tual museum,we can proceed to represent the nodeschemas in terms of grammar rules for a furtheranalysis. We use grammars in two ways - one forthe inner structure (i.e., the structure of a node)and the other for the outer structure (i.e., the struc-ture of a set of nodes and links). Once node schemasare represented by a grammar, the following analy-ses can be done:

I ) Find a different grammar that is simpler than theoriginal grammar, but generates the same hyper-space. The set of all virtual documents can befound first, and we can find another grammar thatgenerates the same set.

2) Investigate a property of the grammar, such as"inherently ambiguous", and determine the con-nection between that property and the naviga-tional structure of the hypertext

a) Inner structure

The purpose of representing node contents withgrammar rule is to obtain a compact representa-tion of the node contents in terms of documentstructure and semantic content. For a node,

node N[p]items iteml, itemk

the corresponding grammar rule (in BNF) is N ::= (si sk ), where 5,15 (I) empty if itemi is a constant(2) the attribute list al, ..., an if itemi is an expres-sion involving these attributes (3) the non terminalsymbol M is item is an inclusion of the form in-clude M[...].

If we can prove that a node instance will alwayscontain at most one item (this depends on the se-lection predicate), then the iteration indicators { }can be removed from the grammar rule. For ex-ample, the grammar rule corresponding to the nodeschema

node Artists_after[date]items

(name, "(born ", birthdate, ")

expand href Works_by[ano]("works"),

include Biography[ano]href

from artistsselected by birthdate >= dateordered by ...is Artists_after ::= { name birthdate[Works_by] Biography }.

The grammar of a node is the set of rules corre-sponding to this node and all the included and ex-panded nodes. For the above node this could yield(if nodes Works_by and Biography were so defined)

Artists_after ::= { name birthdate[works by] BiographyWorks_Ey ::= { title date supportBiography ::= { country date activity 1

Since no rule in this grammar is recursive, this canbe rewritten as a single rule:

Artists_after ::= { name birthdate [{title date support }] { country dateactivity )

In fact, this grammar shows the structure and se-mantics of a node. It is similar to a basic documenttype definition (DTD) for this node.

b) Outer structure

We explain the construction of grammar produc-tions that describe the outer structure of a givenhypertext using a simple example.

Artist

ano, name, birthdate, deathdate Countries

[artist], country, type

"contempor with"

Contemporary

I[birthdate, deathdate], name

Fig.6: Example

In this figure, we see that there is one href link, oneinclusion link and one expand-in-place link In or-der to describe the outer structure of this hypertext,we find possible paths from each node to others.We see that theArtist node has two different typesof links, one for inclusion (to Countries) and theother for href (to Contemporary). From Contem-porary, there is one expand-in-place link to Artist.The node Countries does not have any link to othernodes.


Let the node Artist be described as two characters,aldl ,the node Contemporary as a2d2,and the nodeCountries as a2d3. (Each ai and di can be consid-ered as "brackets" for each node.) If we attachprime symbols to each of those, that represents anode instance of the node schema. For example,a 1 'cl 1' is a node instance for the node aldl .Then,each of the following products represents each linkin the hypertext.

aldl ->a2d2 ->a3d3 ->

al'd1' 1 a2d2 1 ala3d3d1a2'd2 1 a2a1d1d2a3'd3'

If we change al d 1, a2d2, and a3d3 as x, y, and z,respectively, and introduce a starting symbol S, weget the following products:

S ->xlYlzx -> x' 1 y 1 alz'd1y -> y' 1 a2xd2alz'd1 -> xa2x'd2 -> y

Notice that x', y', and z' is a I 'd a2'd2', and a3'd3',respectively.

Once this first grammar is found, we can analyzethe structure by examining the property of thegrammar. We can also find a simpler grammar thatrepresents the same hyperspace. One usage of thisgrammar formalism would be that one can use it asa site map for a given wWeb site so that users canget an idea of how the hyperspace is structured.

Conclusions and Future Directions

In this paper we presented a language to specifyvirtual museums in the form of hypertext views ofdatabases.Since the language is non-procedural andexplicitly shows the structure of the generated hy-perspace, it is well suited for an iterative design pro-cess.The existence of a hypertext schema makes itpossible to check properties of the hypertext, suchas path lengths or accessibility, without accessingthe hypertext nodes themselves.The developmentprocess we propose consists in starting from aninitial design and then entering an analysis-refine-ment cycle.The structural analysis of the hyperspaceuses graph and grammar formalisms while the re-finement is based on several basic operations.

7 '7

This development process is supported by softwaretools to compile the specifications and to dynami-cally generate Web pages (node instances) accord-ing to the specifications.

In the near future we plan to increase theadaptiveness capabilities of the generated hyper-spaces, in particular the path-awareness.We are alsostarting experiments with a new version of the Lazysystem to make the hypertext views active. Thismeans that users will not only navigate in a virtualmuseum but will also act on this museum by up-dating the information nodes they see. For instance,users could create their own virtual exhibitionswithin the virtual museum, add annotations to ob-jects, etc.

Acknowledgement

The authors would like to thank their colleagues Jean-Pierre Hurni and Claire-Lise Mottaz for their preciouscollaboration.

References

ACM (1995) Special section on hypermedia design.Communications of the ACM Vol. 38, No. 8.

Andrews, J. and Schweibenz,W. ( I 996).The Kress StudyCollection Virtual Museum Project:A NewMedium for Old Masters. Version 03-Dec-I996.http://www.arlisna.org/werner.html

Atzeni, P., Mecca, G., & Merialdo P. (1998). Design andMaintenance of Data-Intensive Web Sites. InProceedings of the EDBT'98 Conference,Valencia (pp.436-450).

Conklin, J. (1987). Hypertext An Introduction andSurvey. IEEE Computer,Vol. 20, No.9, 17-42.

Costagliola, G., Dattolo,A., & Francese, R. (2000).AVisual Grammar-based Approach for the Analysisand Modeling of Hypermedia Structures. In Proc.Multimedia Computing on the World Wide Web, 2000

Dar, S., Entin, G., Geva, S., & Palmon E. (1998), DTL'sDataSpot Database Exploration as Easy asBrowsing the Web, In ACM SIGMOD 98 Proceedings.ACM.



De Bra, R,Adaptive Hypermedia on the Web: Methods,techniques and applications, In Proceedings of theAACEWebNet'98 Conference, pp. 220-225, Orlando,Fl., 1998.

Falquet, G., Guyot, J., & Nerima L. (1998). Language andtools to specify hypertext views on databases. In A.Mendelzon & RAtzeni (Eds) TheWeb and Data-bases, Selected papers from WebD8'98 (pp I 36-151).Berlin, Springer Verlag (LNCS 1590).

Falquet, G., Guyot, J., & Nerima L.,Vanoirbeek C., Rekik,Y. (1999). Des documents virtuels pour lire lesbases de données. In M. Crampes & S. Ranwez(Eds) Documents virtuels personnalises, atelier de I aconférence francophone Interaction Homme-Machine.Montpellier.

Fernandez. M.. Florescu. D.. Kang, J., Levy A., Suciu, D.Catching the boat with Strudel: experience with aweb-site management system, In Proceedings ofSIGMOD Conference 1998.

Garzotto, F., Paolini R, & Schwabe, D. (1993). HDMamodel-based approach to hypertext applicationdesign. ACM Trans on Information Systems, Vol. 11, I26.

Isakowitz,T, Stohr,A., & Balasubramanian, R (1995).RMM: a methodology for structured hypermediadesign. Communications of the ACM 38, 9), 34 44.

Park, S. (1998). Structural properties of Hypertext. In K.Gronbaek, E. Mylonas, E Shipman (Eds)Hypertext'98. Pittsburgh:ACM.

Toyama, M. (1998). SuperSQL:An Extended SQL forDatabase Publishing and Presentation,ACMS1GMOD 98.

7a© Archives & Museum Informatics, 2001 75

Re-Purposing of Content and Digital Delivery Convergence:Implications for Interface Design

Slavko Milekic, M.D., PhD,The University of the Arts, USA

Abstract

With the introduction of the digital medium, an abundance of content previously obtainable through other mediabecame available in digital form. Digital delivery implied the necessity for some kind of content modification dictatedby the specific characteristics of the digital medium.

The situation is further complicated by the convergence of different media used for the delivery of digital/digitizedinformation. The same content has to meet the constraints of different display devices and bandwidth limitations.Furthermore, digital delivery makes possible adaptations of content not only in regard to characteristics of hardwareused to display it, but also in regard to individual user's goals and needs.

In this paper I will present the implications for content re-purposing with the migration from the traditional to thedigital medium as well as the problems that arise from convergence of different types of digital delivery mechanism.

Introduction

The effects of the 'discovery' of the digital mediumare global and profound. I have described some as-pects of these changes elsewhere (Milekic, 1999).In this paper I will analyze two currently observ-able trends connected to the introduction of thedigital medium.These trends are:

a) re-purposing of content previously avail-able through other media for deliverythrough the digital medium;

b) convergence of means (physical devices)used for the delivery, display and manipula-tion of digital content.

Each of these trends creates complex constraintsbut also opens up new possibilities for contentdelivery/exchange. These possibilities can be ex-plored only by modifying the traditional interfacedesign practices of applications created for digitalcontent delivery. However, before proceeding tothe analysis of general effects of these trends, it isadvisable to clarify the consequences and implica-tions of each one separately.

Re-purposing

According to the on-line version of Merriam-Webster dictionary, a literal definition of re-pur-posing would be that it is a new way (from Latin

7o

re-,again) of achieving the same outcome (from Latinproponere /to propose/: something set up as an ob-ject or end to be attained: INTENTION). In mostgeneral terms re-purposing can be described asmanipulations of space, time and meaning within thenew medium.Most often re-purposing involves somekind of compression. Compression of data space (ul-timately corresponding to the physical storage spacenecessary for storing it) in a medium currently con-strained by bandwidth limitations is frequently pro-portional to time compression.However, with band-width limitations resolved, the space compressionwill be relevant only in regard to the characteristicsof a user's display device.This does not mean thatthe need for time compression will disappear. In fu-ture, time compression will not be tied to how longit takes an image to be downloaded to one's harddrive, but to how quickly one is able to use (under-stand) the content. In other words, time compres-sion will be related to the compression of meaning.

In regard to the digital medium, one can distinguishtwo different kinds of re-purposing: a) one that re-fers to re-purposing of content from another me-dium (that of print, audio or video, etc.) for deliveryin the digital medium, and b) re-purposing thatoccurs within the digital medium.These two formsof re-purposing can be labeled as inter-media andintra-medium re-purposing.

Museums and theWeb 2001: ProceedingsArchives & Museum Informatics, 2001, p. 77

Milekic, Re-Purposing of Content...Implications for Interface Design

Inter-media re-purposing

Inter-media re-purposing is historically not a newoccurrence.A clear example is a shift from oral tra-dition to written transmission of myths and folktales. In order for oral content to be successfullytransferred into the new medium, it had to beadapted to the constraints imposed by the newmedium. In this process some of the characteristicsof the content and possible interactions with it werelost or changed, while new ones were gained. Forexample, oral traditions rely on a very undepend-able storage medium human memory. Historicallyit has often been the case that certain folk tales andmyths were confined to the brain of a single indi-vidual, upon whose death they would be irretriev-ably lost. Committing these to print preserved thecontent and made it available for transition to anynumber of (literate) individuals in the future. Theother side of this transition is that some possibili-ties were lost. Human memory may be unreliableas storage medium, but it is inherently flexible andcreative. Storytellers freely added and embellishedthe story so that it suited particular circumstances.This led to the enrichment and the creation of dif-ferent versions of popular myths.Transferring suchmaterial to print meant that it had to be 'edited':that is, redundancies had to be removed, inconsis-tencies corrected, and one had to settle for a single'true' version of the content (it is only much laterthat the value of the diversity of oral interpreta-tions was recognized and is now extensively docu-mented). It is important to note that these changeswere mandated by the characteristics of the me-dium itself.

Transferring (re-purposing) content from any otherto the digital medium will be affected in a similarway by properties of the digital medium. In orderto anticipate the effects this transfer will have oncontent accessibility and usability (and user inter-face design), it is important to identify in advancethe specific characteristics of the new medium. Prop-erties unique to the digital medium are conse-quences of binary encoding of information and, as Iargued elsewhere (Milekic, 1999), at the most gen-eral level they can be labeled as:

reproducibilitytransferabilitypervasivenessmanipulability

5 0

Reproducibility refers to the fact that any contenttransferred into the digital form becomes instanta-neously reproducible.As a matter of fact, the veryact of making it available to our senses (by display-ing it on the monitor screen) involves reproduc-tion. Reproduction in a digital medium is perfect:that is, reproduced code (even if it was reproducedmillions of times) is indistinguishable from the 'origi-nal' one.Thus, the notion of the original in the digi-tal medium is going to be very different from theone used for works in traditional media. Most likely,it will be defined temporally as the time of the firstoccurrence of a certain binary pattern, before digi-tal 'originals' started multiplying.

As instantaneous as reproduction, transferability re-fers to re-creation of a binary code at another lo-cation, with the loss of the 'original' code.The easewith which any portion of binary code can be re-produced, transferred and modified led to the per-vasiveness of the medium.We are already at the levelwhere any bit of information stored in this mediumis simultaneously accessible to anyone with accessto a digital portal.

It is the level of control over manipulations in thismedium that is unlike any other traditional medium.Not only can one manipulate information at thelevel of a single bit, but one can also selectively in-teract with the data specified at a level of a particu-lar pattern. For example, it is possible (and triviallyeasy with modern rendering software) to changeonly those data bits corresponding to a certainshade of red in a digital photograph. Hyper-linkingof data (which may be in different modalities) rep-resents yet another level of manipulability, the oneat the level of meta-structure. Thus, the informa-tion in the digital medium is infinitely personalizable,malleable, modifiable and adaptable. Because of itscreative potential the manipulability of digitally en-coded information has been the driving force be-hind the globalization of digitization.

However, blind use of recognized properties of themedium can also lead to less-than-perfect results.Good examples are early attempts to re-purposetext documents for use on the World Wide Web.Re-purposing was done with a knowledge that themedium allows for hyper-linking of any two pointsin the information space.Although the abundanceof hyper-links made the text seem more usable,their



net effect was not any greater than what would beachieved with an automatic page-turner.

Intra-medium re-purposing

Re-purposing within the digital medium is achievedby modification of certain aspects of the content orits delivery mechanisms. Depending on the motiva-tion for re-purposing of the content, one can dis-tinguish two broad categories which I will label

a) device-centered content adaptationb) user-centered content adaptation

Very often the line between the two is blurred be-cause the distinction is not made between the us-ers who access certain content and the devices thatthey use for this purpose. Needless to say, equationof a user's goals with the characteristics of the de-vice being used to reach it leads to serious flaws ininterface design. For this reason, I will describe thesetwo different categories separately.

a) Device-centered content adaptation

In the field of HC1, re-purposing of the content fo-cused mostly on technological constraints of differ-ent devices used to access the content and not onuser interests, goals. etc. Examples of device-cen-tered content adaptation are provided in an excel-lent paper by Ma, Bedner, Chang, Kuchinsky andZhang (Ma, Bedner, Chang, Kuchinsky and Zhang,2000).They provide five general categories of con-tent adaptation applicable to different data modali-ties within the digital medium: video, image, audioand text.

They label these categories as:

data prioritizationdata transcodingmodality transformationinformation abstractionpurpose classification

Data prioritization is achieved by assigning differentlevels of delivery priority to different content as-pects.These could be related to content modality,for example, by assigning priority to delivery of textprior to the delivery of images or audio and videoinformation. Or it can be achieved by prioritized

progressive delivery within the same media type,for example, by delivering images first in low-reso-lution version, and then filling-in more detail if timeand/or user display device allow.

Data transcoding is the process of converting dataformat. In their paper the authors (Ma et al., 2000)refer to data transcoding only in the context of cli-ent display device characteristics.This would includetranscoding of original images to display-appropri-ate format; for example, G1F-to-JPEG or color-to-grayscale. However, it is important to note that someforms of data transcoding may also provide otherforms of content adaptation, like information ab-straction. This is demonstrated in Figure I whichdepicts the same information (photograph of ayoung woman) transcoded to formats compatiblewith different display and bandwidth characteris-tics.

Notice, however, that image 4, although claiming theleast amount of data space (and thus the leastamount of download time) provides the viewer withmuch more information due to the addition of a"thought bubble" which provides relevant contex-tual information for the interpretation of the im-age.

Figure I. Sequence of images with differentlevels of data (and hence time) compression,ranging from 232 kilobytes for uncompressedimage in T1F format to less than 10 kilobytes

for duo-tone image.

BESTCOPYAVAILABLE


Rc-pu rposing1

:According to thc on-linc version or Merriam-Wcbstcr dictionmy,a literal definition of rc-purposing would bc that it is a ncw way

((rom Latin re-, again) of achieving the same outcome (from Latin; proponcrc /to propose!: something sct up as an object or end to

bc attained: INTENTION), In regard to thc digital medium, onccan distinguish two different kinds of m-purposing: a) onc thatrefers to rc-purposing of content front another medium (that ofprint, audio or video, etc.) for delivery in the digital medium, andb) re-purposing that occurs within the digital medium. These two

forms of rc?purposing can bc labeled as inter-media andintro-medium re-purposing.

!Inter-media re-purposing

f Rs...purposing; Aocordhlg to the ethlinerershm of Merriam-Websto dktionaty. a literali definition of rolluposing would he that it is s new way ffrom Win re... again)f of achieving the same outcome (frons Latin proponere /to psisposd: sotrethim,)iset up wan object or end to be attained: INTENTION). In regard to the digitalmedium. oue can distinguish two different kinds of re.purposinv a)onc that

i refos to ra purposing of content (ron another medium (that of print, audio orI video. eta.) for delivery in the digital medium. and b) re-purposing that ocean' within the digital medium. Thme two Rams of re.purpssing can he labeled atf

intosnualki and huramedium na.puiposing..

InIsnmedia Iv-parr:wing ishiandeally 005a new ceenntatee . A ChM Otatriplcis a shill from oral tradition to written transmission of myths and folk taks. Inorder for mai content to be soccessfully ussis(eued into the new medium, it hadto be adapted to the conseraints imposed by the new medium. In this processsome of the charammistics of the welkin and possible intetactions with itIsere lost or ehangol stile other were gainol.

Inter-media re-purposing is historically not a ncw occurrence. A [.

's<w.

Figure 2-3. Information abstraction by reduction of font size and reduction of white spaceallows the text to fit into the browser window without scrolling and

without any change to the text itself.

Modality transformation is the translation of the in-formation delivered in one modality into anothermodality. Examples include transforming digital videostream to a series of still images, delivering only theaudio component of the video, or closed captiontext. True modality transformation would not becharacterized by loss of information, and thereforeone should distinguish it from modality transfor-mation-based reduction.

Information abstraction refers to the procedures thatreduce bandwidth requirements by delivering ex-tracted information to the user. Ma and his col-leagues (Ma et al., 2000) provide examples such asvideo highlights, video frame and/or rate reduction,image dimension reduction, image data size reduc-tion (by increasing compression rate), stereo-to-mono conversion, text summarization, font size re-duction etc. However, these examples mix bothsemantic and non-semantic information abstraction

reducing image data by using a better compres-sion algorithm does not have to lead to perceptibleloss of information and have an effect on meaning,while text summarization may.

Another difference is that some of the given ex-amples use automatic algorithmic procedures (likeimage compression, thumbnail generation, audiosubsampling) while the other ones depend on hu-man editing (text summarization, video description

80

and such) to achieve data compression.Thus, it maybe useful to divide this category into two separateones: data compression achieved through differentautomatic procedures, and true information abstrac-tion achieved (at least for the time being) throughhuman intervention.

Purpose classification is the least clear category sug-gested in the above-mentioned article (Ma et al.,2000).While it is definitely true that"a typical Webpage contains a lot of information and media ob-jects that are redundant or may not be of interestto a user" by which they refer to advertisements,banners and promotional material, it is often thecase that the very existence of a particular pagedepends on displaying this (redundant) information.Although the authors invoked this category withthe goal of meeting low bandwidth characteristicsof different user devices, they are coming very closeto user-centered content adaptation, by discussingthe purpose and usefulness of ads for a user.

b) User-centered content adaptation

Adapting the content so that it can efficiently ac-commodate users with different needs, skills, back-grounds, knowledge level, etc., is another challengein the research agenda for universal usability(Shneiderman, 1999). It is worth noting that toolsfor user-centered content adaptation focus on dif-

82


BEST COPYAVAILABLE


.t.m. Evor,osmu goo _

Re-purposingActording to the ondine version of httariant.Wettster dictionary, a literaldefinon of iwpurposing would he that it is 3 new way (from Latth no., *lin)

r of achieving the sunetancome (from Latin proponent tto propose: somethingset up as an object or end to he attained: INTENTION} 1nd1nedintd

I medium, one can distinguish toe different kinds of impurposing; a)one thatIrefers to impasposing of content (rom another medium (that of print. audio orI video. Me.) the delivery in the digital medium. and It) rmptuposing that omurs

oithin the tfigital medium. These two fonns of reimposing can be latelal asintennualia- and intramtedium eel:imposing.

Intermordid minoraviag is himoriimlly not a ag01 neetwence. A ear exampleis n shift faint oral tradition to written ilallSatiRsioa ormyths and folk Was. ht

V miter for kVA content to he succiawfully transferred into the new gushiest, it hadto be adapted as the constraints imposal by the new mediwn. In this proemssome of dto characteristics of the oargent and possible interaetions with itmere lost or ehangml utile other owe gained_

ratm ESL 1.7-

Re-purposing

Rein:rooting it a new ony of achieving the same outcome. In regain to thedigital otediunt, one imit distinguish ton dirTeient kinds of repurpasing:at one that refers in napurposiug of commit from matcher :tam (that of

i print, audio or video. me.) for delivery in the digital medium, andrmpurpasing that ecaus within dna digital medium. These two fauns of

re.parposing can be labeled as intepinedia and intra-mediwn reinapasev.

An ovimplo of imersiserlin rr.pulaning is a shin front oral tradition to writtentransmission of myths and folk tales. In this tamest mine of the charactaisties

V of the co:non and possible interactions with it %al; lost or changed %bile othaMYR gainml.

Figure 3-4 Further compression is achieved by condensing the textand presenting it in a list form.

ferent kinds of problems compared to the device-centered content adaptation. In this case the prob-lem is often not the network bandwidth but ratherthe cognitive bandwidth.The unique properties of thedigital medium also play a much larger role in user-centered content adaptation by providing the op-portunity for the development of new cognitivetools for knowledge sharing and discovery. For ex-ample, the potential of the digital medium to log allof the interactions that take place in it can be usedfor a qualitatively different level of the preservationof the past. In the following section I will briefly out-line different user populations and possible strate-gies for re-purposing the content to suit their needs.

Adapting content for use by different age groups(from preliterate children to adults) has tradition-ally been resolved by creating a new genre; for ex-ample, children's books, movies etc.This approachhas been implemented successfully on the Web onthe NASA space mission sites that have sectionsdesigned for K- I 2 children. Ideally, the contentshould be made adaptable for different age groups(children, adults) by creating an interface mecha-nism that would enable the user to fine-tune thelevel of complexity of the displayed content.

The content should also be made available to fit auser's interest and immediate goals. Relative failureof algorithm-based search engines to satisfy user

needs indicates that this challenge has not beensatisfactorily resolved. More successful Web por-tals, like Yahoo, owe their success to the fact thatsegmenting of the content to fit different users' in-terests has been done by humans. However, recentadvances of "data mining" techniques (also knownas "knowledge discovery in databases" - KDD) holdpromise that in very near future one will be able to"build knowledge" by searching vast amounts of textavailable on the World Wide Web or be able tosearch through video materials in the same way onesearches text today. Other examples of making thecontent of other media (TV) more adaptive by trans-ferring it into the digital form include ReplayTV andTIVo, devices which allow users automated record-ing and replay of preferred shows (and even learnfrom viewing behavior), slow motion replay, pauseand resume of live TV broadcast.

On the side of content providers, re-purposing isnecessary for achieving different goals. The samecontent has to be presented and interacted with ina different way depending whether it is going to beused for teaching/learning, information, or as a database.This idea is still in its infancy, although somelarge content providers (NASA, NY Times, as wellas some museum sites) are making an effort topresent part of their content in form suitable foruse in education.

83B EST COPY AVAILABLE



Re-purposing of content should also address theissues involving different cultural expectations.Cur-rently, English language and Western cultural mod-els dominate the World Wide Web. Issues of cul-ture-specific content delivery have potential broadeconomic implications, and there are already someservices that provide automatic translations ofWebpages into different languages (AltaVista). However,the quality of automatic translation is severely lim-ited, and commercial Web sites catering to usersbelonging to different cultures resorted to the cre-ation of multiple, human translated and accultur-ated mirror sites.

Adaptation of content for users with special needsis a large and looming issue that has not been satis-factorily resolved. It is worth noting that answeringthe needs of these populations has historically beenthe path to improvement in general usability or ledto the introduction of appliances with global im-pact (like the telephone, tape recorder, etc.).Tech-nologies like continuous speech recognition, screenreaders and haptic-enabled interaction devices (likethe iFeela mouse) will play a major role in this area.

Convergence

The term convergence has several meanings(Merriam-Webster), of which two are especiallyapplicable to the topic of this paper:

I . the act of converging and especially moving to-ward union or uniformity;

2. independent development of similar characters(as of bodily structure or cultural traits) oftenassociated with similarity of habits or environ-ment

The two above definitions illustrate both the prob-lems and the state of affairs in the arena of digitalconvergence. Wireless telephones, PCs, televisionand various other Internet appliances are compet-ing with each other in taking each other's function.One can get the weather report on the display ofmobile phone, watch TV on a computer screen, andinteractively shop over TV. This, in itself, can onlybenefit the end user if there is true uniformity be-tween the platforms. However, this is not the case.The number of competing standards and architec-tures is staggering.

There are three major aspects of digital conver-gence (Forman and Saint John, 2000):

convergence of content (audio, video and data);convergence of platforms (PC,TV, Internet appli-ances, game machines);

convergence of distribution (how the content getsto a user)

What is evident today is the user-demand-drivenconvergence of content the use of re-writableCD ROMs is increasing because of the possibilityof storing large amounts of digital data that are com-mon household items - like digital photographs,music files in MP3 format, multimedia presentations,etc.That the trend is going in this direction is alsoevident from the fact that DVD players/recorders(which can store even larger amounts of data) arethe most successful electronic product in recentyears.

Convergence of delivery platforms is a muchthornier issue that has not even been resolved forthe shift from analog to digital television. We arewitnessing a kind of parallel evolution similar to thesecond definition of convergence quoted above:dif-ferent platforms are developing similar characteris-tics, and the 'selection forces' are user preferencesand the characteristics of the medium itself.

The questions of delivery of digital content are alsofar from being resolved.Whether the content willreach us via fiber-optic cable, broadcast or satelliteis not clear. Although it may seem that the ques-tions of the delivery method are irrelevant, for theend user they have enormous implications for thefuture of power and control over the media.

Design Implications

While it may seem pointless to speak about inter-face design in light of chaos and unresolved issuesin the domain of digital convergence (further com-plicated by lack of clarity regarding the characteris-tics of the digital medium), there are some generaldesign guidelines that seem evident. In simple terms,the design has to be device-aware and user-aware.This means that the content has to be scalable(useable with devices with different display charac-teristics,storage space, processor power) and adap-tive (useable by users of different abilities, goals,



backgrounds and cultures). Use of intelligent algo-rithms and adoption of common hardware stan-dards can resolve scalability of the content. Cre-ation of truly adaptive content is still a huge chal-lenge and currently can be resolved only by humaneffort in creating adequate redundant content rep-resentations that suit different needs. On the otherhand, adaptive delivery of content (and 'awareness'of user actions) is part of the digital medium po-tential, and will eventually determine the ways inwhich the medium is used.

Ultimately, delivery in the digital medium (and useof the term digital medium) will include any end-user delivery method where the content, at somepoint, was encoded in binary form in order to usesome of the advantages of the digital medium (forexample, reproducibility). This would include aprinted page (from a printer), or an email messageretrieved with a mobile phone via text-to-speechsoftware.

Theoretically, the need for re-purposing should dis-appear if the medium is used to its potential:that is,the delivery of practically limitless accumulated con-tent becomes infinitely adaptive (and user adapt-able) to meet any individual's need, knowledge levelor cultural background. Of course, we are far fromthis ideal, and in the meantime, we still have to de-velop efficient ways of inter- and intra-medium re-purposing of content

References

AkaVista search engine, on-line.http://www.altavista.com

Forman, P., Saint John, R.W, (2000) "Creating Conver-gence", Scientific American, November issue, on-linehttp://www.sciam.com/2000/ I I 00issue/I 100stjohn.html

Logitech iFeela mouse.on-line: http://www.logitech.com/cf/products-productoverview.cfm/79

Ma,W, Bedner,L, Chang, G., Kuchinsky,A., Zhang, H.(2000) "A Framework for Adaptive ContentDelivery in Heterogeneous Network Environ-ments", presented at Multimedia Computing andNetworking 2000 (MMCNO0), January 24-26, 2000,San Jose, California, on-line http://www.cooltown.hp.com/papers/adcon/MMCN2000.htm

Merriam-Webster on-line dictionaryhttp://www.m-w.com/home.htm

Milekic, S. (1999) Emerging tools & techniques of digitalmedia: history log and multiple futures, inBearman, D.,Trant, J. (eds.) Cultural HeritageInformatics:Selected papers from ICHIM'99, Archivesand Museum Informatics, Pittsburgh

NASA educational sites on-line:http://www.grc.nasa.gov/WWW/K-12/John/K-12_Teacher_Resources.html

New York Times Daily Lesson Plan on-line: http://www.nytimes.com/learning/teachers/lessons/

ReplayTV on-line: http://www.replaytv.com/home.htm

Shneiderman, B., (1999) Universal Usability:PushingHuman-Computer Interaction Research to EmpowerEvery Citizen, Institute for Systems ResearchTechnical Report.T.R. 99-72, on-linehttp://www.isr.umd.edu

TIVo on-line: http://www.tivo.com


New ContentNew Uses

86

A New Way of Making Cultural Information ResourcesVisible on the Web: Museums and the Open Archives

Initiative

John Perkins, CIMI Consortium, Canada

Abstract

Museums hold enormous amounts of information in collections management systems and publish academic andscholarly research in print journals, exhibition catalogues, virtual museum presentations, and community publications.Much of this rich content is unavailable to web search engines or otherwise gets lost in the vastness of the WorldWideWeb.The Open Archives Initiative (0Al) has developed an easily implemented protocol to enable data providersto expose their information and service providers to access and use it. The CIMI Consortium is working with theOAI to make it possible for museums to enhance the availability of their research resources, allowing them to bediscovered in Web-space by the specialist audiences for which they are intended or by service providers who collect,distribute or in other ways provide access. By building on the OAI protocol, Dublin Core, and museum communityXML developments, significant advancements can be made in exposing museum information resources. This paperintroduces the OAI and its protocol, explores its potential relevance to museums, presents CIMI's work as an alphatester of OAI, and looks ahead to future developments.

Introduction

The ubiquity of the Web and success of popularsearch engines have fueled an expectation for quick,easy, and successful results in the quest for infor-mation and knowledge. Increasingly, scholars, stu-dents and other explorers are turning to the Webfor their research needs and relying less often ontraditional research sources. Museums have im-mensely rich information resources in publications,research papers, exhibition catalogues, virtual mu-seums, databases, and intranets, but access to muchinformation of value about the kinds of materialsmuseums hold is rarely available through web searchengines. Internet search engines only reach staticHTML web pages, but much of what museums haveis opaque to the indexers because it is in databases,dynamically generated, or in some other non-HTMLform.These resources constitute what is becomingknown as the hidden Web, estimated to contain400-550 times more content than the commonlydefined Web. (BrightPlanet 2001)

If this problem alone were solved and all the hid-den web resources were suddenly available for in-dexing, the difficulty of finding reliable, useful, pre-cise information would be seriously compounded,not alleviated. One way to address this is throughcollecting and indexing metadata records, ratherthan indexing the entire contents of HTML pages,thereby providing greater possibilities for precision.

This is essentially the traditional library approachof creating descriptive metadata and building unioncatalogues. However, library catalogues are expen-sive to maintain and in theWeb world, both difficultto find and hard to search across.

As separate approaches, it seems neither the oldlibrary methods nor the new Internet approach isserving researchers and scholars particularly well.(CLIR 2001)

A particularly promising solution is to explore theutility of combining the best of traditional libraryand museum techniques, such as creating descrip-tive metadata records in catalogues, with the bestof new Internet techniques like large scale, machineharvesting of information. It is possible to considerthis because of new developments in Web work-able technical protocols, the uptake of XML as away to package and transfer information, and thedevelopment of international standards for describ-ing museum metadata content.

The Open Archives Initiative

The Open Archives Initiative, OA1, (http://www.openarchives.org) develops and promotestechnical protocols and standards, collectively calledthe OA1 technical framework, to facilitate access to

,57Museums and theWeb 2001:Proceedings


BEST COPY AVAILABLE

Perkins, Museums and the Open Archives Initiative

scholarly research information on the Web. It isbased on the premise that a simple, easily imple-mented technical framework can allow holders ofinformation to create repositories of metadata de-scribing their resources that in turn can be harvestedand made available for further processing or use.(0A1 Protocol 2001)

The OA1 technical framework describes how reposi-tories of metadata about information resources areconstructed. Repositories are essentially networkaccessible servers offered by data providers. A re-pository makes available via a simple protocol recordsthat contain metadata about its items (content).Arepository may, optionally, organize its items intosets corresponding to its collections or othergroups, thus allowing clients to harvest metadatarecords selectively.

A record is an XML encoded byte-stream that servesas a packaging mechanism for harvested metadata.The OAI protocol mandates the use of unqualifiedDublin Core as the common record for discovery.(Dublin Core 2001) It also allows community-spe-cific metadata sets described by XML SCHEMASfor more detailed description based on the asser-tion that both simple metadata for interoperabilityand cross-domain discovery as well as a methodfor conveying richer community-specific descrip-tions are needed.

All OA1 repositories must recognize a set of re-quests or verbs carried in http POST or Get meth-ods that allow access to the metadata records. It isthrough these commands that metadata is harvestedand transferred.

One design criteria of the OA1 technical frameworkof particular relevance to individual communitiessuch as the museum community is the notion ofextension packages. Not only does the protocol al-low a community to expose its own metadataschema, but it also allows other extensions such asunique collection level metadata or, if deemed nec-essary, rights metadata.The OA1 protocol doesn'tplace limits on the number of allowable metadatasets, but does specify that their data formats bedescribable by an XML Schema.

In order to federate distributed repositories, theOA1 has established a registry service available

through the OA1 web page to provide a list of pub-licly available repositories and to provide a mecha-nism for conformance testing. (0A1 Registry 2001)

The potential of OAI technical framework is in pro-viding the enabling technology for the federating ofdistributed information resources and their discov-ery and use.The power of the OA1 technical frame-work is in its simplicity and ease of implementation.

Describing Information Resources forDiscovery: Dublin Core and XML

While the OA1 protocol defines new technical stan-dards for repositories and the machine-to-machinedialogue between data providers and harvesters, itdraws on the established international standardDublin Core for the mandatory metadata recordformat. (Dublin Core 2001) The Dublin Coremetadata set was developed specifically to allow asimple and easy-to-use description of informationresources for their discovery.The utility of DublinCore was corroborated by CIMI in its Dublin CoreMetadataTestbed that explored the use of unquali-fied Dublin Core for discovery of museum re-sources, both at a coarse grain level and at a moredetailed, complex level.At the higher, coarse grainlevel, the Dublin Core is effective both for discov-ery of resources and as a means for museums tointeroperate with other communities in a net-worked environment such as theWorldWideWeb.(CIM1 I 999a)

To go beyond simple discovery and interoperation,the OA1 anticipated, through inclusion of the exten-sion packages concept, that in addition to a coremetadata format, individual communities ofimplementers would require additional descriptiveformats.Again, this need was borne out in the CIMIDublin Core testbed findings where it was con-cluded that extending the Dublin Core to handlecommunity-specific needs was problematic. (CIMI1999a)

Alternatives need to be found to extending or quali-fying Dublin Core to facilitate the more completedescriptions needed by the museum community.TheOAI addresses this by allowing support for parallelmetadata sets.For museums, this could conceivablyinclude record structures such as SPECTRUM (rich



museum object information), CIMI (public access),AMICO (art museum images), MIDIS (monumentsand built environment),OBJECT ID (loss and theft),and RLG Inc.'s CM1 (Cultural materials).

The challenge is that each community of OA1implementers must agree on what metadata for-mats are needed beyond the core, and must pro-vide XML SCHEMAS for each of them. Once this isaccomplished, the metadata foundations will be inplace for use of the OAI protocol.

Early in the development of the OA1, CIMI recog-nized it had a number of features that could helpsignificantly advance access to museum information.First and perhaps most importantly, the OA1 pro-tocol was simple and appeared to be easy to imple-ment using tools and skills (Webservers, http, JAVA,PERL, CGI etc.) within the easy reach of museums.Secondly, it relied on the Dublin Core as a metadataformat for the simple discovery of information re-sources within and between communities.This for-mat was proven workable for museums, and thereexists a guide to best practice for its use. (CIMII 999b) Finally, the OA1 mandated XML for packag-ing richer metadata sets and transferring records.XML is a standard that is gaining wide acceptancein museums, and XML SCHEMAS exist or are inthe process of being created for many of the com-munity standards mentioned above.

CIMI's test of CIAI V. I .0

Because of the perceived potential of OAI for mu-seums, CIMI participated as a pre-release tester ofthe OA1 protocol. (0AI AlphaTest 2001) As part ofthe test, we built a generic 0A1-compliant reposi-tory. (CIMI OA1 Repository 2001).The repositoryarchitecture shown in Figure I uses a layered ap-proach, standardized APIs, a generic http interface,and interchangeable components. This allowsimplementers the use of different back-end data-bases, webservers, or XML generators and mini-mizes hard-wired coding.

The repository took a skilled JAVA programmertwo weeks elapsed time to build. This period in-cluded both an orientation to CIMI and the OAI aswell as reading and understanding the protocol, andthen building the application.The development pro-cess started with designing a JAVA API for the re-

Apache Web Server

Apache JServ

CIMI OAI-HTTP Servlet

CIMI OAI Repository API

Reference Implementationof API Using MySQL

Figure I: CIMI OAI Repository LayeredArchitecture

pository and a JAVA servlet to interface betweenhttp/OAI protocol layers and the repository. Thereference repository was written using MySQL andJDBC.

The CIMI reference application serves Dublin Corerecords from an Apache Webserver generated bythe earlier CIMI testbed from the MySQL database.Because of the modularity inherent in the architec-ture, the Repository could be layered on top of anyODBC-compliant database, be served from otherservers, and make use of different XML generators.

Looking Ahead

The initial evaluation demonstrated that the OA1protocol is indeed simple to build. CIMI has limitedtechnical resources and skills but was nonethelessable to successfully build an OA1 repository thatappears to be useful. Based on the positive experi-ence as an alpha implementer, CIMI plans to con-tinue explorations of the OA1 protocol and researchits use by museums.

One way is by making the code for the CIMI re-pository and its associated explanatory materialsavailable for downloading from the CIMI Website.(CIMI Publications 2001) We hope museums willtake advantage of its availability to install, experi-

8 9@Archives & Museum Informatics, 2001

BEST COPY AVAILABLE

89

Perkins, Museums and the Open Archives Initiative

ment with and use the protocol.We hope to com-pile and report the experiences of these ad hoc tests.

CIMI is also interested in conducting a more for-mal, large-scale test of the OA1 for museums as aCIMI testbed.As part of this work, we propose us-ing OA1 V. I .x in combination with scoped exten-sions and other applications necessary for aggrega-tion processes (e.g. editorial control, content man-agement and enhancement, registry) to harvest andcollect museum metadata from cultural memoryorganizations. It will focus on materials that docu-ment culture and civilizations, including museumobjects, art, images, and related materials.We willstructure this as a CIMI testbed, inviting participa-tion from a group of interested members.We ex-pect respondents to include national museum or-ganizations, individual museums, commercial enter-prises, and museum system vendors. Once under-way, the project will run 12-18 months in concertwith projects in other communities and the OAItest period.

The purpose of the research is to explore how aspecific community of users can use the OA1 pro-tocol. Part of this is to investigate what agreementsusers need to make within the protocol frameworkitself (e.g. additional metadata sets), and part is toidentify any extensions or modification required tomake the framework additionally useful. Our testbedwill give museums a place to expose their metadataand promote their institutions, test the 0A1 proto-col for utility in describing non-bibliographic re-sources, and could provide a rich resource of cul-tural metadata leading eventually to the materialsthemselves and the institutions offering them.

It is one thing to test the technical viability of theOA1 protocol by implementing the protocol at atechnical level, but another to imagine and deter-mine useful services that might be built on it.Wehave imagined a number of scenarios that could betested.

We imagine, for example, that services like AMOL(Australian Museums Online), AMICO, the Cana-dian Digital Museum, or RLG Inc.'s Cultural Materi-als Initiative might want to add a feature to "searchfor more like this" in collections or repositoriesnot under their direct control.We imagine that in-dividual museums or groups of museums all using

the same collections management system mightmake use of the repository for internal operationalneeds, for scholarly access, as well as for supple-menting information services they provide publicly.We imagine that commercial services such as AskArt(http://www.askart.com) - a directory of AmericanArtists - or Virtualology (http://www.virtualology.com) - a virtual education project - would find theresource attractive and useful.We imagine that aneasy-to-use protocol might be attractive to salesand auction houses, encouraging them to make use-ful research information resources available (suchas those now manually compiled). We expect na-tional service providers like the UK JISC higher edu-cation information services to have an interest inusing museum repositories.We know that the op-erators of the new Internet top-level-domain formuseums (MusDoma) are extremely interested inproviding directory-like services that would includesearch access to our harvested cultural materialsmetadata.We also imagine that harvesting exhibi-tion catalogues and museum publications from li-brary catalogues, artist biographies, museological lit-erature from A&I services, and sales records fromauction houses is of interest to museum research-ers. These all are the kinds of services that mightemerge once the OAI is widely deployed in themuseum community.

Regardless of the services developed, there will bea number of issues relating to widespread adoptionof the OA1 protocol in the museum community.Weforesee a need for our community to test hypoth-eses, assertions, and issues such as:

the utility of the Dublin Core for meeting infor-mation requirements of service providers andconsumers;

the functionality of the OAI protocol as a basisfor a harvesting service, including issues of hier-archical descriptions, scalability, required exten-sions, presentation and partitioning;

community extensions required for the OAI andDC in order to provide useful metadata withinand between communities;

requirements and practices for content manage-ment, metadata enhancement, and editorial con-trol



aggregation, integration, access and presentationof bibliographic, textual, multiple media, and ob-ject metadata

the need, scope and services of a registry

access control and rights

mechanisms, processes for paths to underlyingcontent

business model for sustainability

Conclusion

Both CIMI and many of our members have signifi-cant experience in the metadata harvesting busi-ness. It is this experience that motivates us to ex-plore the OAI protocol as an enabling technologyto facilitate access to resources by making it easierfor museums to expose and collect metadata.TheOA1 protocol in concert with a museum testbedseems a logical and sensible research initiative thatwill bring us closer to making the rich informationresources museums hold more widely available toresearchers and other users.

References

Bright Planet 200I:A description of the Deep Web is athttp://www.completeplanetcom/Tutorials/DeepWeb/summary03.asp

CIMI 1999a: CIM1's report of the DC testbed http://www.cimi.org/publications.html#dc_2

CIM1 I 999b: CIMI's Guide to Best Practice for the useof Dublin Core in museums is available from http://www.cimi.org/public_docs/meta_bestprac_v1_1_210400.pdf

CIMI OA1 Repository 2001: OA1 compliant harvesterscan access the CIMI Repository at: http://www.cimi.org/oai/OAl_test.html.There is a humanviewable interface page there.

CIMI OA1 2001:General information about CIM1'swork on the 0A1 can be found at hap://www.cimi.org/oai/index.html

CIMI Publications 200I:The CIMI OA1 Repository codeand associated explanatory documents can befound in the Metadata Harvesting section of thePublications area on the CIMI Website at http://www.cimi.org/publications.html

CUR 200I:Vision paper for new Web searchingmethods at http://www.clir.org/digliblarchitec-tures/vision.htm

Dublin Core 2001: Detailed information about theDublin Core, its activities and metadata sets canbe found at http://www.dublincore.org

0A1 Registry 2001:At the time of writing the OA1registry was still being constructed. By the time ofpublication access to the online registrationshould be at http://www.openarchives.org

0A1Alpha Test 200I:A listing of participants and theirexperiences is found at: http://www.openarchives.org/OAISC/alpha-testing-press-release.htm

OA1 Protocol 200I:The current version of the OA1protocol document is available at hap://www.openarchives.org/OAIIopenarchivesprotocol.htm

Acknowledgements

The author would like to thank Carl Lagoze for hiscontributions to the OA1 sections of this paper andHenry Stern for developing the CIM1 Repository.Thanks are also due to the Open Archives Initiative, theDigital Library Federation, the Mellon Foundation, andthe CIMI members for enabling the author to partici-pate in the development of CIM1'sthinking on the useof 0A1 in museums.

91

O Archives & Museum Informatics, 2001 9 I

Using Interactive Broadband Multicasting in a MuseumLifelong Learning Program

Leonard Steinbach,The Cleveland Museum of Art, USA

Abstract

The Cleveland Museum of Art has embarked on an innovative approach for delivering high quality video-on-demandand live interactive cultural programming, along with Web-based complementary material, to seniors in assisted livingresidence facilities, community-based centers, and disabled persons in their homes.The project is made possible inpart by a grant from the Technology Opportunity Program [TOP], National Telecommunications and InformationAdministration, U.S. Department of Commerce totaling more than $500,000.The purpose of the grant is to demonstratehow emerging broadband telecommunications technology can deliver "lifelong learning and the arts" to populationsfor whom direct involvement with cultural institutions would otherwise not be possible.

The approach uses Cisco IP/TV interactive video archive/broadcast servers and broadband multicast technology in acontrolled public infrastructure environment, rather than the closed corporate or campus network environment forwhich it was designed. In addition to describing the program design and operation, this paper analyzes how thisMuseum, whose core competency is not, nor should be, advanced technology development and management, musteredthe expertise to achieve technological innovation in pursuit of programmatic goals. It also focuses on the process ofconvening outside individuals, organizations, and expertise to complement each other to achieve a common goal.

Introduction

Be forewarned!You are about to read a tale of greattechnological, social, and educational aspiration. Ofcommunity groups, corporate interests and federalfunding. Of populations isolated and impaired. Ofexternal evaluation and intradepartmental coordi-nation. It is a work in progress about which theknown facts are provided and our hopes for thefuture are unveiled.

This paper presents an innovative approach for usingbroadband multicast technology to deliver highquality video-on-demand and live interactive cul-tural programming, complemented by Web-basedresources, to seniors in assisted-living residencefacilities, community-based centers, and disabledpersons in their homes. But that is a rather colddescription. We want to change people's livesthrough continuing lifelong learning and the artsand interaction with each other. We strive for noless than arousal of the spirit, though we will defineour goals a bit less spiritedly later on.The technol-ogy we have chosen comprises the best tools wecould find at this time, ones we believe harbingerthe future.The combination of technology, programand process form an experiment which has beenfunded. No doubt the details of technology willchange dramatically over time, but we hope thatthe process and program model we are presentingwill survive, evolve and prove useful to others.

Background

To understand the genesis of the program, it is im-portant to understand its host's core values andexperience.The Cleveland Museum of Art, estab-lished in 1916, is one of the finest encyclopedic artmuseums in the United States and is an integral partof Cleveland's cultural and civic identity.As an insti-tution established "for the benefit of all the peopleforever," the Museum welcomes close to 600,000visitors each year.The Museum also has a long his-tory of extensive public programming in film, music,and dance as well as art-related classes, lectures andfamily programs. In addition, the Museum has strongrelationships with the community and communitygroups through a range of outreach programs andcommunity festivals as well as its teachers schoolservices, and adult continuing education programs.Most recently, in 1998, the Museum became a con-tent provider and remote site for the OhioSchoolNet Program. Ohio SchoolNet provides liveinteractive distance learning programs for studentsin grades K- I 2 throughout the state (http://www.osn.state.oh.us).This is all no quirk of fate.TheBoard ofTrustees had long established the Museum'sstrategic goal of creating "rich and diverse educa-tional and public programs that serve and engagemany different audiences and communities in an in-novative and dynamic fashion."

BEST COPYAVAILABLEMuseums and theWeb 2001:Proceedings


Steinbach, Using Interactive Broadband Multicasting

In addition to its education context, the emergingrole of technology within all areas of the Museumwas recognized by the Board. This resulted in an-other important strategic goal: "to become a na-tional leader in the use of new and emerging tech-nologies to enhance the value to society of themuseum's collections, intellectual initiatives, andother activities:7o accelerate progress in this area,in 1999 the Museum established a new InformationTechnology Division (I.T.), headed by a Chief Infor-mation Officer (CIO) on peer level with other se-nior management and reporting to the Museumdirector. The new CIO (the author) immediatelyestablished a guiding doctrine for technological lead-ership and adventurism:"The core competency ofa museum is not the mastery of diverse, complextechnologies.., it is the creative use of them."Thereis no contradiction here. It is quite possible for aMuseum to achieve its techno-dreams without in-sisting that it find the capital and human resourcesnecessary to become the be-all end-all of complextechnology knowledge and management.This prin-ciple will be echoed throughout this project plan.

This backdrop is important. Without a history ofcommunity involvement, strong educational pro-grams, and a dedication to the use of new and evolv-ing technologies, it is unlikely that the Museum wouldever have conceived of this project nor receivedthe funding and other support needed to make ithappen.

(Technology) Opportunity (Program)Knocks

In the fall of 1999, the Museum became aware ofthe availability of funds for 2000 from the Technol-ogy Opportunities Program (TOP) of the NationalTelecommunications and Information Administra-tion of the U.S. Department of Commerce. Since1994, this has been a highly competitive, merit-based

federal grant program that brings the benefits of anadvanced national information infrastructure tocommunities throughout the United States (http://www.ntia.doc.gov/otiahome/top). The focus of itsfunding may vary from year to year, but for FY2000the intent could not have been more congruent withthe Museum's interests.TOP was "especially inter-ested in projects...using advanced networktechnologies...for example, broadband networks...

93

to deliver immense amounts of data quickly to thedesktop:' In addition, the notice identified "LifelongLearning and the Arts" as an area of interest, par-ticularly "training and instruction to lifelong learn-ers in non-traditional settings such as homes (and)community centers:' With this scope defined, weproceeded to explore whether there was a pro-gram the Museum would want to develop whichwould further its goals. However, it is not our na-ture to contort our goals in order to "follow themoney."

Lurching Toward the Bait

As would be expected, there was significant brain-storming within the Museum, including staff and lead-ers from Information Technology, Development,Education, Curatorial, and Community Outreach.Wehad many answers and ideas, but finally realized thatthe truth lay not in ourselves but in our stars...our friends and colleagues in the community. Weheld a technology/program brainstorming party.Colleagues from SchoolNet,from schools,and fromnearby cultural organizations converged. Represen-tatives of Community groups aggregated. Technol-ogy companies whose help might be needed in speci-fication and installation were there to keep our feetnailed to the ground. (Luckily, we still could reachfor the sky.) Curious curators convened.The publicTV station was tuned in and got turned on.You couldnot have too many cooks in this kitchen.After twohours of guided discussion a general consensusseemed to evolve.We would focus on isolated olderpersons in residential facilities or those who fre-quented community centers, and, if possible, disabledpersons who primarily spend their time at home.We would provide both programming on demandand live interactive programs where they live orcongregate. It felt right and meshed with Museumgoals. However, more precise goals and next stepswould have to be defined.

Before addressing those issues, it is noteworthy thatthis meeting, born of prior excellent relationshipswithin the community, generated more than a con-cept.Our guests informed our thought process, andwe made it clear that we were in this together.Thissetting of stage would prove invaluable as the projectmoved forward.



At this point, the project development processstill had to answer several critical questions:

What did we hope to achieve?

Was there available technology which would letus demonstrate the conceived program at fea-sible cost and effort?

Would residential facilities and community-basedorganizations commit to participating in this pro-gram?

Where would we get the content, both live andarchived, for this project?

How could the Museum inaugurate this level oftechnology without diverting resources, or vio-lating the precept of creatively using technologieswithout investing an inappropriate amount of timeand effort in them?

Proof of Concept/Definition of Goal

Intuitively, it seemed reasonable to expect that olderand impaired persons might benefit from interac-tion with cultural activities, but we needed to con-firm our hunch. With the assistance of MalvinSchechter, a noted journalist, gerontologist, and con-sultant on issues of population and aging, importantresearch findings were brought to our attention.Gerontological research seemed to assert that pro-grams in the arts stimulate cognitive functions, en-hance daily life experience, and trigger memoriesand creativity in older Americans. Research con-firmed that being part of a social network and di-minishing isolation and loneliness have significantimpacts on health and longevity (Rowe & Kahn,I997;Fees & Martin, 1999) . One study concluded that"interactive computing technology opens up accessto levels of communication and personal controlthat impact directly on the quality of life for 'con-fined' individuals. Indeed, the term `confined' losesmuch of its meaning when the world is at one'sfingertips" (McConatha, McConatha & Dermigny,1994).

The Museum also had its own experience in thisarea. For example, the museum has worked withthe Cuyahoga County Board of Mental Retardationand its constituents for about three years in a stu-

dio art program..Work with this population has al-ready resulted in dramatic changes in participants'lives. The simple artistic freedoms like choice ofcolors or how to position the paper have increasedconfidence in participants who have since attemptedother personal achievements. Participants in thisprogram are now showing their artworks in com-munity exhibitions. One individual learned to climbstairs for the first time at the museum, while an-other spoke his name for the first time in his life.

We knew enough to know we were on the righttrack. Our program goal was then articulated asfollows:

To enrich the lives of seniors and disabledadults through excellent arts programmingdelivered to them where they live and con-gregate via broadband technology. Arts-re-lated programs will be delivered to partici-pants at three types of sites: assisted livingresidence facilities, community-based centers,and private homes. Older as well as physi-cally and emotionally disabled adults willbenefit from relevant and rewarding ex-changes with cultural institutions through thenew technologies proposed herein by mak-ing it possible for them, for the first time, toparticipate in a broad range of excellent pro-grams from which they have been histori-cally isolated.

Specifically, we would:

Develop an art program that provides a mecha-nism for intellectually stimulating older adult learn-ers;

Develop an interactive,art-based program that fa-cilitates social engagement among older adultlearners;

Develop a delivery medium which can reduce thesocial isolation often associated with living in se-nior care facilities; and

Develop a replicable model that can be used byother cultural organization seeking to deliver com-parable content to similar populations via broad-band technologies.

94O Archives & Museum Informatics, 2001 95


Most important, our program would explore anddemonstrate how learning makes more meaningfulthe lives of older persons or disabled persons forwhom a visit to the conventional museum and cul-tural institutions and activities would be difficult,inconvenient or impossible.We would hope to learnhow these extramural programs might stimulatesuch audiences to inquire and learn in satisfying ways.

Additionally, we would develop a model and expe-rience base that, by the nature of content and tech-nology, is replicable throughout the country, espe-cially as this technology becomes increasingly avail-able.

The Future is Now: Selecting theTechnology

The technological vision for the project now hadto emerge alongside the more programmatic andcontent-oriented goals.The vision was to establisha multi-point "distance learning-like" environmentwith some level of interactivity, one which wouldinclude archived material, additional Web-like con-tent andWeb access.The technology should be moreopen, less complex and less expensive than tradi-tional video teleconferencing. After all, this is theway Internet is going: more, richer content andgreater interactivity through affordable access to ahigh-speed public infrastructure. Even if it is notthere yet, our charge was to be innovative in ouruse of broadband technology and demonstrate whatthe future may be.

This presented three technical issues: how to cap-ture, digitize, and compress video; how to developand manage a video archive which can provide videoon demand on line; and, how to achieve an infra-structure for getting the content from the sourceto the client.

A special work group was established to discoverwhat existing technology might be right for us.Thissmall group initially included Museum I.T. staff and aconsultant from Keane, Inc., a major national I.T.consulting firm which was engaged in another Mu-seum project He agreed to contribute the assis-tance.

Some prospects were easy to rule out Althoughwe considered cobbling a variety of products andtechnologies together to accomplish our goal, we

9 5

knew this would conflict with our "core compe-tency doctrine" and hobble our chance for success.The best opportunity for Museum support and asuccessful implementation would be as close to aturnkey solution as possible. Environments such asthose available from Real Networks or MicrosoftNetMeeting and others were immediately ruled outThe nature of both content and clients would de-mand close to full screen video of near televisionqual4We wanted our clients to have warm, seem-ingly familiar experiences with both our technol-ogy and our contentAfter all, the program was nottargeted toward highly self-motivated Internet us-ers who would put up with quarter-screen jitters.Our content should be the reason for participating,and our technology should not get in the way.

We knew that we would require high-speed con-nections to our clients. We also realized that wewere looking for an IP multicast solution, particu-larly for our live broadcasts. We decided that theincreasing availability of ADSL (henceforth, DSL),with its emphasis on downstream speed, would bea feasible enabler (Craioveanu, 2000).We did notconsider Internet access services through cable-Nproviders, since this was not available in our area.

IP multicasting is defined as an "efficient means oftransmitting of a single large stream of data to agroup of selected users at the same time on aTCP/IP network such as the Internet" (http://www.techweb.com/encyclopedia). It is also funda-mental to the evolution of the Internet as demandincreases for better quality transmission of videoand audio, and as transmissions become targetedtoward groups rather than broadcast to all. IPmulticast protocols have been under developmentsince the early 1990's.Thus far they have been usedprimarily in corporate and academic environmentswhere network traffic can be carefully engineeredand monitored.We were aware of some public ini-tiatives in this area, such as Mbone, since 1996 acooperative and voluntary experimental virtual net-work within the Internet.This has been succeededby other "bones:' including 6bone and Qbone, andprotocols, like the Internet itself, continue to evolve.But there is no current standard (Wirbel, 2000).

In addition to the protocol issues, IP multicast alsorequires a telecommunications infrastructure whichis"multicast enabled:'That is, it is comprised of rout-

96 ©Archive's & Museum Informatics, 2001


ers and cabling which can handle the traffic de-manded by rich media and which can also filter sig-nals to allow clients to selectively receive transmis-sion with a quality of service that would make theproduct palatable. (This is somewhat simplified, butsufficient for this purpose.) The current accelera-tion of commercial interest in IP Multicast is clearlyevidenced by the IP Multicast Initiative, "a world-wide, multi-vendor forum accelerating the adoptionof IP Multicast stimulating demand of IP Multicastproducts and related services" and significant sourcefor articles, white papers, and conference informa-tion (http://www.ipmulticast.com). More than fiftycompanies now participate in this forum, includingCisco Systems, Real Networks,Yahoo Broadcast,andLucent Technologies. There are even a number ofnational conferences each year focusing on this area.Although there may be some complex and costlyworkarounds, the public infrastructure was notready yet for broadband IP multicast, even thoughwe were.

Welcome to theTurnkey MulticastingSystem

Our product review brought us to Cisco Systems,Inc.'s IPTP/ system. Cisco's IP/TV was chosen overother prospective solutions because it provides amore comprehensive set of features, in an end-to-end turnkey solution, than any other solution wecould find (http://www.cisco.com). It comprisesvideo capture with a choice of compression tech-niques, real-time broadcasting capability, videoarchiving, an Internet-based interface, and full routersupport at all points.As a tool designed for educa-tion and training applications within an academic orcorporate environment, other features, including aninteractive question manager and usage monitoringtools, would also add value. A true broadbandmulticast solution (when used within a networkinfrastucture supporting these features), IP/TV canprovide very high quality full- or near full-screenvideo, something we had identified as a requirement.We also decided that the implementation of a turn-key solution by a single well-known vendor had theadded advantage of one-stop training and support.If our program model of content delivery provedsuccessful, new and more feature-rich productscould simply replace the elements we had chosen.Finally, because we would use Cisco routers, Ciscocould be held responsible for the system's perfor-

mance truly end-to-end.This initial assessment wasfollowed by a product demonstration and tutorialfor our technology team at Cisco's Cleveland of-fice, and additional telephone conferences withCisco's IP/TV specialists. The agreement that ourlocal Cisco sales and technical reps would join ourtechnology team as needed, and that the regionaloffice would provide training and technical assis-tance to help assure project success, clinched thedecision.

The decision was later bolstered by an informativecase description of an IP/TV project at the VirginiaCommunity College System (http://www.so.cc.va.us/vccsit/IPTVproj.htm).

The planned configuration of the Cisco IP/TV sys-tem comprises the following components:

IP/TV 3411 Control Server which centrally managesthe entire IPTP/ system.The Control Server com-municates instructions, such as scheduling infor-mation, available video types, and bandwidth con-siderations to IP/TV Broadcast and Archive Serv-ers, along with program information, to the IP/TVViewer client software at the workstation. Its in-terface allows system administrators to managecontent, servers, and bandwidth, and schedulebroadcasts. It also balances network video loadsand optimizes network performance automatically.This server can be fully controlled remotely fromany, even low speed, IP-connected terminal.

IP/TV 3422/3423 Broadcast Servers capture anddigitize, store, and transmit programs accordingto directions received from the Cisco 3411 Con-trol Server.They are primarily used for multicastinglive or prerecorded programs from devices suchas video cameras,VCRs, DVDs, satellite, or cablefeeds, or from prerecorded Windows Media,AVI,MP3, and MPEG files.The Server must be locatedphysically near the source devices, such as cam-eras, mixers, etc., to which they directly connect.The model differences pertain to the servers' digi-tization format capabilities.

IP/TV 3431 Archive Server provides large-volumestorage capacity for Video on Demand (VoD). Itenables the appropriate stored video programsto be delivered at the right time, to the right audi-ence, whether streaming prerecorded video on a

©Archives & Museum Informai6, 2001 97


scheduled basis or responding to a singular re-quest.This server can be fully controlled remotelyfrom any, even low speed,1P-connected terminal.

IP/TV Viewer is the IP/TV system client-side soft-ware, which communicates with the IP/TV 341 IControl Server to get information about all avail-able programs and display a program listing. It al-lows participants to select and display programs,ask to receive a scheduled program, or choose todisplay a program on demand. (It informs theControl Server of what content is desired, andthe control server then directs the content deliv-ery from the Broadcast or Archive servers.) TheViewer can list both available IP/TV format andWindows Media programs and offers VCR-like

Cleveland Museumof Art

controls, key word searches and program pre-scheduling. [Although IP/TV content could alsobe accessed via an html web page and WindowsMedia Player, this client is required in order toreceive archived content stored in Cisco's pro-prietary RTP format (sort of an encapsulatedMPEG-4).We will require the Viewer on all clientstations in order to benefit from RTP's high qual-ity video compared with the bandwidth neededto receive it].

The configuration of the equipment is describedlater and portrayed in Figure I.

WVIZ / PBS

1P/TV Broadcast Serverand Remote Console

lPflV IPITVControl ArehiveServer Server

IP/TV Broadcast Serverand Remote Console

InternetGateway

Muiticast Router

98

DSL(Virtual Private Circuits)

Figure I: IPITV Broadband Multicast Distribution Network



Supplemental Cisco software which should proveeffective for providing, managing, and monitoringcontent development and delivery include:

Web Presenter,which allows the opening of a videowindow and Web-based slides concurrently.

Web Plug-in, which allows a participant to watchan IPTTV program embedded directly in a Webpage.

Question Manager, which provides a participantsinteractivity by letting participants type in ques-tions on line. A moderator or instructor eitheranswers the questions in real time or archivesthem for follow-up.

Web Info, which allows participants, while watch-ing a program, to click a button and travel to apredefined URL location, typically a Web page,containing additional program information.

SlideCast, which permits participants to simulta-neously see the presenter's PC-generated presen-tation materials in one window and the presenterin the other, as if they were in the same room.

StreamWatch, which gathers participant demo-graphics such as number of participating stations,identities, and viewing times for pre scheduledprograms, and for on-demand programs logs in-formation about which programs participants arewatching and when.

Ticket to Ride

Having tentatively selected the product suite thatshould meet the video capture, storage, manage-ment and distribution requirements of the project,the network infrastructure on which this would ridehad to be identified. As noted earlier, the publicInternet is anything but ready for broadbandmulticast prime time.We enlisted the assistance ofa local major ISP who recognized the benefits ofearly involvement with broadcast multiband.Theirassistance was also encouraged by their ongoingrelationship with the local Cisco office, not to men-tion the prestige and good will associated with as-sisting the Museum. This A-Team team, now com-prised of representatives of Keane, Inc. (consulting),APKnet (ISP), Cisco Systems, Inc.,and Museum tech-

-

nical staff, along with any additional ad hoc expertswe could think of to call or e-mail, set to task. I loveit when a plan comes together (Smith, I 983)... andit did!

The network infrastructure would work as follows:

As a DSL reseller through local telephone provid-ers with direct lines to their central offices, the1SP would create private virtual circuits betweenthemselves and our clients (ISP-to-TelcoPoP-to-client])

The DSL client lines at the 1SP could be then beconnected to a single high speed, broadband,multiprotocol, multimedia router (Cisco7206VXR) with a highly secure, specially config-ured gateway connecting it to the public Internet.

The IP/TV Control and Archive servers would beco-located at the ISP and connected directly tothe high speed router to assure the fastest andmost direct service delivery. Demanded videowould travel directly from the ISP to the client viathe virtual private circuit. The servers could beoperated remotely from any site with an autho-rized IP connection, yet benefit from the powermanagement and physical security inherent in aISP facility.

Live broadcasts would be compressed and fed byBroadcast servers "on location" to the Controlservers at the ISP via dedicated T-1's terminatedat the same high speed router; similarly, video-on-demand archival material could be loaded ontothe archive/media servers at the ISP using the sameBroadcast servers via the same T-I's.

Client locations would require multicast enabledrouters (Cisco 1720) in addition to the DSL mo-dems specified by the provider of the DSL loop.

This configuration establishes a fundamentally closedsystem. The 1SP, which acts as the hub, maintainsend-to-end control of all emanating lines.The projectrelies on a private IP routing scheme for our clientstations and media servers. Remote managementaccess is also protected through access control liststhat permit access only to devices with authorizedIP addresses, in addition to correct passwords.Thegateway to the Internet,also located at theISP, pro-



vides network address translation, thereby mask-ing the identity of the originating workstations andservers. In turn, the gateway is configured to pro-vide anti-spoofing protection and a multi-layeredand firewalled environment to assure a high level ofoverall security for components of our system.Simi-larly, filtering at the gateway prevents our high band-width content from spilling into and degrading ser-vice on the 1SP's Internet backbone. Overall, thanksto an ISP with end-to-end connection control andwilling to provide multicast broadband support fora small population, we expect to successfully emu-late the Internet of the future.

Content Capture, Transmission andDisplay

While the fundamental infrastructure and core com-ponents of this project have been described, thereare key ancillary devices and technologies worthnoting, although elaboration exceeds the scope ofthis paper. For live transmission we will use MPEG-4 (http://www.cseltiempeg/) with its high ratio offull-motion video quality to bandwidth and otherfeatures. For archival storage and video-on-demandwe will use Cisco's RTP format, which is essentiallyencapsulated MPEG-4 requiring IP/TVViewer clientsoftware at the workstation. Content will be digi-tized locally using Broadcast Servers and down-loaded/transmitted via T- I to the Archive Server atthe ISP hub.

Ancillary equipment associated with content pro-duction, distribution, and storage include video cam-eras, microphones, mixers, video editing system withspecial effects, video recorders, document cameras,slide-to-video projectors and others. Some of thesewere already available through our existing distancelearning programs while others were acquired forthis purpose.

Work stations at participant sites will be standardPC's with at least the equivalent of a PI11/600Mhzprocessor; the 1P/TV Viewer uses the processor todecode MPEG-4 and RTP media.Also,128 MB RAMwill be provided.At this time we have not identifiedany specific video or audio requirements beyondthe minimum generally provided with current modelworkstations. Where a personal residence is thelocation of participation, standard small speakersand a 17" monitor will be provided. For group fa-

cilities, a 36" RGB monitor, appropriate speakers,and wireless keyboard and mouse (if desired) willbe provided, along with any special cart or cabinetrequested.

Program Operation and Manage-ment: Partners, Partners, Partners

The technology plan described,although at the coreof the project, is merely a facilitator of the program.The program needs content to distribute and par-ticipants to participate. If one concedes that theMuseum has designed a program that qualifies as"creative use" of technology on behalf of theMuseum's strategic goals, then there still remainsthe question of where the mastery of complex tech-nology should reside. The answer is, with thosewhose core competency is technology. Similarly, al-though the Museum has extensive experience ineducational programming within the scope of itsexpertise, there is more to Lifelong Learning andthe Arts than a single institution can provide. Finally,we needed to identify and enlist community facili-ties and residents who would be willing to join ourexperiment and work toward its success.

To achieve this level of distributed responsibility,participation and interest, three spheres of "Part-ners" have been convened, many of whom had beenat our initial brainstorming meeting:Content Part-ners,Community Partners,andTechnology Partners.

Content Partners is a consortium comprised of theMuseum and additional arts and cultural institutions,including the Cleveland Orchestra, one of the fin-est in the world; the Crawford Auto-Aviation Mu-seum of the Western Reserve Historical Society, arepository for artifacts and archives related to thehistory of Ohio; and, WVIZ/PBS, the area's publicbroadcasting station and active producer of educa-tional television programs.The number of contentpartners may grow as the program matures. Thecombined resources of these organizations will pro-vide a broad and varied menu of programs in theareas of local history, visual arts, jazz and classicalmusic, theater and the applied arts,as well as a rangeof thematic programs about general artistic, culturaland literary issues. We are projecting a minimumarchive of 300 programs by the end of year two, inaddition to a minimum of 12 live broadcasts offeredmonthly. These organizations receive no payment



for their contribution. It is noteworthy that TOPgrant conditions unequivocally stipulate that fundsmay not be used to "produce information content,"although it will support some "creation or conver-sion of content (in order) to utilize information in-frastructure technologies to address real-worldproblems" (TOPApplication Kit, I999).For example,the program is funded to videotape an alreadyplanned lecture series, but cannot pay to create one.Content partner contributions of time and any di-rect expenses incurred on the project's behalf arerecorded and considered in-kind match to the fed-eral cash grant.The Museum's Education and PublicPrograms Division manages the Content Partners.

Community Partner organizations bring a diversepopulation to the program, including minorities andseniors, and those with varying levels of physicaland mental abilities. Organizations include severalassisted living community residential facilities; theCuyahoga County Board of Mental Retardation'straining center; and a college-based special educa-tional program for persons age 50 and over.As theprogram matures, participation may change. Thepersonal commitment of community partner staffis crucial to the success of the program. Commu-nity partners have committed the time and talentsof their on-site staff to facilitate the use of the tech-nology, to integrate the project into their ongoingprogram activities, and to create follow-up activi-ties so that their clients can achieve the maximumbenefit of this experiment.They have also agreed toencourage use of the system for video-on-demandand access to complementary Web resources.Thisincludes making the equipment accessible to clientsfor independent as well as group use.The partnersalso agreed to take part in the evaluations, com-plete surveys and interviews, and send a represen-tative to meet at least quarterly to discuss the pro-gram progress and recommend improvements.These organizations receive no payment for theircontribution.Their contribution of staff time spentin activities, meetings or paperwork uniquely re-lated to the project and any direct expenses incurredon the project's behalf are considered in-kind matchto the federal cash grant.The Museum's Educationand Public Programs Division manages the Com-munity Partners.

Technology Partners comprises technology orga-nizations whose core competencies are congruent

with their projected roles. For example, APKnet,Inc., the major regional 1SP referred to above, willprovide end-to-end network management and con-nectivity services and support to our clients andcontent origination/management points; that is, theMuseum and WVIZ/PBS. They will also co-locatethe Archive and Control servers on their premises.WVIZ/PBS, the area's public television station, willremotely manage the Archive and Control servers,including video-on-demand sources and broadcastscheduling.With an IP/TV Broadcast server on theirpremises, they will also provide their studio as alive broadcast venue as well as download contentto the archive. Cisco Systems, Inc.'s Cleveland of-fice has committed to a special technical supporteffort and has contributed training to assure thatthe project works. Keane, Inc.will help monitor andevaluate the implementation. Museum InformationTechnology staff manage the implementation pro-cess, and will perform all remote site equipmentset-up (except network connectivity devices), pro-vide help desk support for the project, and developa friendlier end-user interface than is provided withthe IP/TV package.The vendor partners are partici-pating on the basis of fee-for-service and/or contri-bution/discount of services. Contributed time andservices from technology partners are accorded avalue, which is considered in-kind match to the fed-eral cash grant. The Information Technology Divi-sion manages the Technology Partners.

Project Evaluation

Intrinsic to any federally funded project, and vitalregardless of auspice, is the ongoing and final for-mal evaluation of project process and outcome.Wehave divided the evaluation into two distinct com-ponents: technology implementation and programeffectiveness.

In order to assess our technology implementation,Keane, Inc. will identify and document all technicalimplementation, training and support issues, prob-lems and resolutions.This record should serve asan important source document for any organiza-tions that would like to replicate in whole or partthe technical program we are modeling. Our Keanerepresentative has been involved with the projectfrom our initial investigations and has an exceptionalunderstanding of what we are trying to achieve.However, since identifying the core technologies,

100@Archives & Museum Informatics, 2001 1 0 1


Keane has not been involved in specification orimplementation, thereby enabling an unbiased viewof our process

In order to assess program effectiveness, we werefortunate to have enlisted the Institute for Innova-tive Learning, Inc., led by noted researcher John H.Falk.This non-profit research and evaluation orga-nization specializes in "free choice" learning andarts-related programs and has extensive experiencein evaluating Museum-based educational programs..Formative and summative evaluations, based on in-tegrated evaluation planning with all program part-ners, will encompass direct observations by evalua-tors,focussed and open-ended interviews,question-naires for caregivers, and case studies as appropri-ate. Interim and final reports will assess the efficacyof our approach and promote the sharing of ourexperience. Ongoing informal feedback will help usnavigate the program.This Institute's involvementbegan at the very onset of the program, even be-fore our first formal Partners meetings were con-vened, and we believe that early involvement ofevaluators is essential to the evaluation process

Budget and Staffing

The formal budget for this program is $1.2 millionover a two-year period. Of this, $545,000 repre-sents the Technology Opportunity Grant financialassistance award.The remainder is matched throughthe in-kind support of our Content,Technology,andCommunity Partners, and the Museum's dedicationof time of existing management and staff, and directexpenses for some equipment and supplies.The fed-eral contribution is primarily used for hardware,software, initial network configuration, and monthlytelecommunications charges.Approximately 10% isapplied to evaluation. Federal dollars also directlyfund 1.5 full-time equivalent positions at the Mu-seum: a full-time Project Coordinator acts as a liai-son between Program and Content providers andthe Museum's project managers, evaluators and oth-ers, performing a variety of functions to assure thatthe program runs smoothly and effectively.A half-time Education Assistant works with all content andcommunity partners, including Museum staff, tomodify or enhance prospective content to bettermeet the needs of older or impaired participants.Avariety of staff from both the Information Technol-

ogy and Education and Public Program Divisionsare involved with this project as needed. TheMuseum's Informaton.Technology Division's HelpDesk also serves as the project's help desk. TheDirector of Education and Public Programs is re-sponsible for program content and working rela-tionships with our community partners.The ChiefInformation Officer, who heads the InformationTechnology Division and is Project Director, is re-sponsible for all aspects of technical implementa-tion all budget management, reporting require-ments, and program evaluation.

Current Project Status

The grant award was announced on October 1,2000. As this paper is prepared in mid-February,2001, following many planning meetings, much ofthe IP/TV equipment has arrived.Technical staffs atWVIZ/PBS and APKnet have been experimentingwith demonstration units for several weeks. Mu-seum technical staff received a briefing and starteddemonstrating 1P/TV features to our Communityand Content partners.This proved a significant en-ergizer for the group, who could finally see whatthey were getting and know that it was for real.

The only big surprise is that there haven't been anybig surprises. Telecommunications lines have alsobeen ordered and are at the brink of hook-up buthere,too, no surprise.... local DSL providers do notreally know if DSL is available at a location until thetime of installation.Although we"pre-qualified" com-munity partner sites last year, as to whether theycould be connected, that turned out to be mean-ingless. Possibly two of our participants may getT-1connections until DSL is really available. Site visitshave been made to all community facilities, in orderto perform facility surveys for equipment selectionand set-up, and to start getting to know residentsand staff (the "joy" part).

Our community partners have already participatedin the baseline stage of the program evaluation,which entailed the on-site collection of baseline dataabout how staff feel about the forthcoming pro-gram, and about the current state of social isola-tion, activity and relationship with the arts of pro-spective participants. Here the feedback from ourevaluator has already proven invaluable.

1 01.72



First, our informal feedback was that our Commu-nity Partners' staffs were extremely enthusiasticabout the project and were already identifying ad-ditional ways of using the technology before ourfirst program even "airs." For example, it was sug-gested that we share special programs currently atone facility with another, or allow residents' talentsto be displayed live from our studios. One facilitywhich has a regularly scheduled discussion of socialand political issues suggested that we open it up formulti-facility participation or view.We are now chal-lenged to harness this energy and level of owner-ship which we never presumed to rely on, nor ex-pected to achieve at start-up.

Second, we have discovered that facility staff areplanning to target our programming toward per-sons who still maintain very active lives, and havestrong longstanding interests in the arts and ourcultural institutions.This may suggest that there isnot much room for great improvements for thesefolk in life satisfaction from our "diminishment ofsocial isolation" or "increased socialization" in thisgroup (or maybe there is?). On the one hand, thispopulation might actively help us evaluate the qual-ity and effectiveness of our programming; on theother hand, this might preclude us from impactingon a more needy population,giving us no more thana short-term Pyrrhic victory. On the third hand, wedo not want to dampen staff enthusiasm or setourselves up for self-fulfilling failure either by sec-ond guessing their judgement or foisting our pro-grams on participants before we have ironed outthe kinks. For now, we believe we have elegantlysolved this trilemma.We will enthusiastically beamour programming toward the audience of the facili-ties' choice, respond to their feedback and analyzeour impact. As we become more experienced atcontent selection and delivery, and facility staff be-come more confident in their use of the technol-ogy and the success with their audience, we mayurge that the audience be expanded to those thusfar deprived of participation.We may even find thatthis occurs naturally over time without our inter-vention.

The relevance of this scenario is not so much theproblems and issues it describes. More importantly,this story illuminates the critical importance of earlyindependent evaluation of program progress, andthoughtfully deliberate intervention (or in this case

lack thereof).Without the early warning, this mayhave become an urgent issue bordering onadversarial. Instead, we share an abundance of lem-onade.

Content partners have also been meeting to de-velop our program schedule,and the wealth of ideasis breathtaking: a special live marimba and organconcert; our Art of the World appreciation coursefor adults, videotaped and parsed into small mor-sels and archived for demand, complemented by livefollow-up conversation from the museum; an histo-rian showing neighborhoods of yesterday and to-day to stir memories, emotions,and intellecta Pan-orama of African American Theater from WVIZ/PBS.These are ideas and plans, and fodder for anotherpaper. One challenge will be to develop the synergyof this program with the existing programs of theContent Partner institutions. Recalling that grantconditions unequivocally stipulate that funds maynot be used for content development, it becomesall the more important to sense what existing pro-grams may be most amenable to re-purposing forthis program audience and technology.Another chal-lenge will be to choose programs selectively andcarefully, learning from those that have proven mostand least effective and popular.Additionally, we mayneed to consciously differentiate or vary program-ming to best match differently skilled members ofour target population. Learning from our success inart classes with impaired populations, we will prob-ably include live and archived activity lessons, withguides for staff.Only one cloud hangs overhead;thatis concern raised by some content partners that intheir enthusiasm for our program goals, they mayhave underestimated the need for their institutions'unreimbursed commitment of time and resources.This may result in a need to attract more contentpartners, or our partners may find that their con-tributions are not so onerous after all.

As we continued to walk through and talk throughthis program, we recognized one more question toconfrontWhy should staff at our Community Part-ner facilities actually believe this could be signifi-cantly beneficial to their residents - more beneficialthan TV or any other leisure activity? Staff attitudein fact might prove key to project success.To helpestablish insurance in this area, the Museum pro-duced a day-long seminar, demonstration, and pizzalunch for all of our Partners (Technology, Commu-



nity, Content). We featured a special film and lec-ture presentation by gerontologist Mal Schechter,who had helped with our original research. Stafffrom the Cuyahoga Community Board of MentalRetardation and Developmental Disabilities dem-onstrated how technology benefits the lives of thepeople they serve.All of us focused on why we be-lieve that this program is not just entertainment;it's life enhancing.A pre-kickoff rally. And it got ravereviews.

Conclusion

This paper has tried to present the evolution of anotion as it becomes an innovation. It has tried torelate the importance of matching an institution'spersona to an almost far-fetched goal. It tries tomake a case for developing sound inter-organiza-tional relationships as a means of achieving projectsthat would otherwise be out of reach.And of coursewe are trying to demonstrate how advanced tele-communications may bring cultural organizationscloser, and with more profound impact, into manymore lives, in the not too distant future. If the pro-cess described seemed too well thought out, toomuch on an even keel, a piece of cake, then I mayhave abused my literary license. Guilty as chargedon bypassing some of the bumps in the road; weare just too busy trying to catch sight of the pot-holes in the darkness ahead. But really, don't youjust love it when a plan comes together!

103

References

Rowe, John W, M.D. & Kahn, Robert L, Ph.D. (1997)Successful Aging. Generations 37:4, 443-440.

Fees, Bronwyn S. & Martin, Peter & Poon, Leonard W(1999) A Model of Loneliness in Older Adults.Journal of Gerontology, Psychological Sciences, July1999, 231-239.

Wirbel, Loring (2000) Russ Hobby Advancing toInternet2 Over "Bones" Electrical Engineering Times,September 27, 2000, 6.

Craioveanu, Radu. (2000).The benefits of IP multicastover DSL. Network World, February 7, 2000.

Smith, Colonel Hannibal (1983) The A-Team. NBCTelevision Network. (1983-87). hap://www.xs4all.n1/jmm/a-team

Acknowledgements

This information was made possible in part by a grantfrom the Technology Opportunity Program [TOP],National Telecommunications and InformationAdministration, U.S. Department of Commerce


Streaming Audio and Video: New Challenges andOpportunities for Museums

Jim Spadaccini, Ideum, USA

Abstract

Streaming audio and video present new challenges and opportunities for museums. Streaming media is easier toauthor and deliver to Internet audiences than ever before; digital video editing is commonplace now that the toolscomputers, digital video cameras, and hard drives are so affordable; the cost of serving video files across theInternet has also significantly decreased.The initial growth of the World Wide Web was driven by the ease of authoringWeb pages, the affordability of the medium, and the potential to reach large audiences.The same conditions are nowin place for streaming audio and video, and we're just beginning to see the widespread use of the technology.

The challenges and opportunities presented to museums today are similar to those that existed when the World WideWeb first became popular. Many questions raised are the same: Can this technology be used to expand the museumexperience? Are there ways to take advantage of the "two-way" nature of the Internet? A few museums have alreadyexperimented with streaming media and some lessons have been learned. However, presentation methods and thetechnology itself continue to evolve. This paper shares experiences with streaming media (focusing on streamingvideo), and discusses some possible new directions for effdctive use of the medium within online museum environments.

Introduction

Ten years ago Apple Computers Inc. releasedQuickTime, a digital video and animation system.At about the same time the World Wide Web wasdeveloped at CERN in Switzerland.These technolo-gies converged in 1995 when the first QuickTimemovies appeared on the Web and when the firststreaming media format (Real Network'sRealAudio) was released. No longer an oddity onthe Web, digital video has now become common-place. Most digital video and audio now developedfor the Web is in a "streaming" format, meaningthat the entire file or large portions of the file don'tneed to be downloaded in order for the users toview or listen to its contents.

Streaming Media Drives the Internet

The growth and reach of streaming media contin-ues to hit all-time highs. According to statistics (http://www.Nielsen-NetRatings.com),35 million Web us-ers accessed streaming content in November of2000 from their homes.This represents a 65 % in-crease from November 1999.The same study foundthat streaming media increased to 36% among allInternet users.This figure compared to a reach of28% of all Internet users the year before. A No-vember 2000 study by MeasureCast (http://www.measurecast.com) and Harris Interactive(http://www.harrisinteractive.com) found these fig-

ures to be even higher, with 57% of online usersfound to be using streaming media weekly.

Citing ease of use, availability of content,and growthof bandwidth, a study by U.S. Bancorp Piper Jaffray(http://www.piperjaffray.com) suggests that stream-ing media will be the next "macro-growth driver"on the Internet.The growth of broadband has beenwell documented. According to Gartner Group'sDataquest Inc. (http://www.dataquest.com), six mil-lion US customers signed up for high-speed Internetaccess in 2000, bringing the total number of broad-band users to 11 million.This represents about 9%of all US users.While this is a small percentage overall,it represents significant growth. Last year, accordingto the Computer Industry Almanac (http://www.c-i-a.com), the figure was a mere 3% of home users.

High-speed Internet access is not essential forstreaming media but it does enhance the user's ex-perience. As detailed in "Creating Online Experi-ences in Broadband Environments" (http://www.archimuse.com/mw2000/papers/spadaccini/spadaccini.html), users get more data throughput,which means better quality video, more frames-per-second, a larger image, better quality audio, and less"lossy" compression.Also, users with a broadbandconnection (56kbps or faster) are 50% more likelyto access streaming media than users with slower

1 0 4Museums and theWeb 2001: Proceedings BESTCOPYAVAILABLE


Spadaccini, Streaming Audio and Video

(56kbps or slower) connections, this according toNielsen-NetRatings (http://www.Nielsen-NetRatings.com).

Easier and Cheaper All the Time

The tools for creating streaming media continue toget easier to use and cheaper to purchase. DV videogear, computers, and software are more affordableand easier to use. Many home users and studentshave begun making their own digital videos.The toolsfor shooting professional-looking digital video havenever been cheaper.A $3,000 camera now has thesame resolution that a $30,000 camera had just fiveyears ago. Desktop video systems are commonplace.The success of the iMac and their free iMovie soft-ware is a good indication of this trend.

The tools for streaming video, server software inparticular, continues to decline in price.This is duein part to competition between three companies:Real Network's Real Media was the standard until inthe last two or three years, Microsoft's WindowsMedia and Apple's streaming Quick Time enteredthe market. Microsoft's server software is now builtintoWindows 2000 and Apple's Quick Time Stream-ing Server is included in the Mac OS X Server. Bothof these products cost less than $1,000.

In addition, the number of ISPs (Internet ServiceProviders) who offer streaming media services con-tinues to expand. Bandwidth costs have declinedand competition has eroded the cost of deliveringstreams to Web visitors. All of these trends takestreaming media out of the experimental realm. Evenwith its limitations (small window size, poor videoquality), streaming media is quickly becoming a vi-able way to reach Internet audiences.

Museum Visitors

Specific information about museum Web visitors isdifficult to find since the medium is still so new.Theonly study examining connection speed and stream-ing media is "Who's Out There? A Pilot User Studyof EducationalWeb Resources by the Science Learn-ing Network (SLN)" (http://www.archimuse.com/mw2000/papers/semper/semper.html ) found that"a significant percentage of visitors (14 to 22%) nowhave access at speeds faster than 56k:' While thisstudy used a small samplingabout 50,000 visitorsover a one-month periodit did study three dif-

ferent Web sites:The Exploratorium, The FranklinInstitute, and Science Museum (NMS1) in London.

About 22% of visitors accessed these museum re-sources from school and about 10% visited the sitesfrom work environments. It's likely that a largerpercentage of these visitors were accessing withhigh-speed connections.This might help explain thediscrepancy between the broadband figures fromthe Gartner Group's Dataquest Inc. study, that found9% of users connecting at 56kbps or greater, andthe figures from the SLN Pilot User Study.

The SLN Pilot User Study also asked visitorswhether they viewed one or more multimedia clips:23% stated they did.While it's difficult to draw toomany conclusions from this one figure taken fromonly three resources (and considering that only twoof the three resources contained streaming media),it's at least an indication that a significant percent-age of museum Web visitors were actively interact-ing with streaming media in 1999. So that from thissomewhat limited data, one could infer that thenumber of Web visitors to museums who can ac-cess streaming media continues to grow (as it cer-tainly has in the general Internet population).

Expanding the Museum Experience

The complex question of how to present stream-ing video remains. Obviously, the answers are sub-jective. Streaming media can be presented eitherlive, or more commonly, on demand. Each type ofstreaming media requires different presentationmethods to effectively reach Web audiences. On-demand video has offered much more variety, whilemost museum Webcasts have been presented invery straightforward ways, with the developers con-centrating on the more complex technical and pro-grammatic aspects of producing and delivering thelive video stream.

On Demand

In early 1997, while working at the Exploratorium,the author developed a resource called "The Sci-ence of Hockey" (http://www.exploratorium.edu/hockey), the first in a series of sport scienceWebsitesdeveloped over the next two years. Permission wasgranted to interview local professional hockey team,the San Jose Sharks. It was perfect timing as the



Shoot i 'How do you tot pock MO oaks m.hole. 1144 Ore of de fed doom= teeasked when Sr begin to ma= Idscrota of takes,. Ts **es

r44Reeffiesergtoodoomeetxtthe... S Ei

Slapracb

de IfHL Mategame 01' depth= roe Oteked%nth a . stas It* Son hueMuds Non arguabte 1110 44-1,

ea ' &freeman

Figure I.The Science of Hockey

first streaming video format, RealVideo, had justbeen released.

There was no question that the video had to bepresented within the context of a text- and image-based site.This had little to do (in the thinking atthe time) with creating a context for the video, butrather the reality that most users would not havethe technical equipment to view the video.The clipsthat appear in the site were thought of as "bonus"materials, available to those who at the time wereon the cutting edge of Internet technology.

The clips were shortfrom 45 seconds to aboutfour minutes. From a technical standpoint, entirevideo interviews could have been presented; insteadclips that matched the topics in each part of thesite were created.The length of the video clips pro-vided flexibility,allowing users to choose which clipsthey wanted to view and in what sequence.

In looking back, this formula worked fairly well.Thereality was, and still is, that the Internet is primarilya text- and image-based medium and thatWeb sitescan be explored in a non-linear fashion.This sameformula was used for many resources on a varietyof topics: "The Science of Baseball" (http://www.exploratorium.edu/baseball),"The Science ofCycling" (http://www.exploratorium.edu/cycling),"Sunspots" (http://www.exploratorium.edu/sunspots),"Frogs" (http://www.explorato riu m.edu/frogs), an d

"The Faultine" (http://www.exploratorium.edu/faultline).These sites used text and graphics to tellthe story, and video and other media to enhance it.This format provides end users with options: theycan read the story, print it out, skim the text, lookat the images, or just view the video clips. Mostnews organizations and variety of other contentsites take similar approaches.

Recently ldeum helped create a site called "Robot-ics: Sensing,ThinkingActing" (http://www.thethech.org/robotics) for the Tech Museum of Innovation. Fourfeatures were created for this site.Three of the fourused streaming media, presented differently in eachfeature. In one section,video was used in much sameway it was done in the Science of Hockey:an articlewith images, video, and other media used togetherto tell the story. For another section audio clipscreate a context for an interactive discussion areafocusing on robotics and ethics. The "Robot Art"feature relies heavily on streaming video, to enablethe artists themselves, and their work, to be theprimary focus.The section contains streaming videointerviews with the artists, images of their work,and QuickTimeVR (360-degree panoramas) of theartists in their studios. The video clips are short,one to three minutes each, and the correspondingtext briefly describes the subject of the clip.Theseclips, along with additional media resources, pro-vide a variety of choices for the Web visitor.

.4 . isCk.


mond

Figure 2. Robot Art

BEST COPYAVAILABLE

107


An excellent site that makes use of on-demand videois the"The Living Room Candidate"(www.ammi.org/livingroomcandidate).The site focuses primarily onpolitical advertisements that ran during presiden-tial campaigns from 1952 to I 996.The ads are pre-sented as separate streaming video clips.Additionalmaterials consist of a short introductory text, alarger program, an educational guide, and thumb-nail images from the videos.The additional materi-als help explain and enhance the video, provide his-torical context, point out themes,and create a richerexperience.

These examples use additional media elements tocomplement streaming video. In the case of "TheScience of Hockey," for example, on-demand videois used within a text- and graphics-based story.Thevideo is not required to tell the story.Whereas both"RobotArt" and "The Living Room Candidate" relymuch more heavily on the user's ability to viewstreaming video. All the examples use short videoclips that provide flexibility in what the user views,and in what sequence. Live Webcasts have not pro-vided the end-user with this sort of flexibility untilnow. Nor has the addition of other media elementstoWebcasting been focused on.

Live Webcasts

In most cases, video-on-demand is just one of a va-riety of media types used to tell a story.AWebcast,

on the other hand, has been viewed as an end initself. Many times little thought is given to how othermedia elements could enhance the Web user's ex-perience. Considering the technical, organizational,and programmatic complexity involved in present-ing live Webcasts it's not surprising. But despite thelimitations of medium (poor video quality and thesmall window size), exploring other ways of im-proving the end-user's experience would seem tobe the next logical step.

Live Webcasts can bring the museum floor to re-mote visitors or remote locations to the museumand Web visitors.At the Exploratorium, the authorwas involved, to one degree or another, in abouttwo dozen Webcasts. To date, the Exploratoriumhas presented over fifty Webcasts.

The 1999 EdipseWebcastQ/wwwexpbratorium.edu/eclipse) was easily the most popular Webcast (toboth Web and museum audiences) theExploratorium ever presented.The ingredients forsuccess largely had to do with the structure of theprogram. First, it presented a unique phenomenon,a total solar eclipse. Second, it presented a live videosignal from an exotic remote location, in this case,Turkey. Promotion efforts and media hype aboutthe eclipse helped drive traffic to the site.The eventitself was an excellent vehicle for exploring the sci-ence and history surrounding eclipses. On the mu-seum floor, the "Webcast" studio and surrounding

Exeloratorium Live Eclipse WebcastTeleconvinterne1Schematic torAugust 11. 1999

030101111sus10041120 GMT

MEW 40.1is...Nun soVOROInesn

Space MOIR 311hropair I I17 dirusWaK0.1....1.11.211..

Pours.Need

11.1%.1Opol Sun 6.11.41.

1111 WON. sr.161.11.Winn.10.11111.1 enbei N. II* fn.

spelblo. WOW /I

Ft,*

1.1.0 =mass1010.1.11.

llatISM bra loom Palen iv* *mksliMasps

lbonlnen en. Amp

On notInnn somanwee

Turkey

%Mir 1.3 wird 04-0WM M. hpimni..........rmur

Wiedwon Irarsiond%Won 0111)

I United States

Illmkasaa. Internet.e=114. 4.1114...11. In 41...................,,14.1111.......... MM. ..I. dee .........1.11..t......1. awn. ea loam nob. 111.......ing....10.a....1.. as.. .10.11.1e.

Itgiontormen1.1.1111C.WISWalon Jade tw1Jpr0

own

Broadcasccom

Figure 3. Exploratorium Eclipse Telecom/Internet Schematic.This schematic demonstrates the technicalchallenges involved in the most complex Webcasts.

107©Archives & Museum Informatics, 2001108 BESTCOPYAVAILABLE


SOJAI ECLIPSE

LIat PESO. F"Ica Covos.

AvthJet- 11,1eu,000 A A,

Fththverf 24Eicertition

carlot,FEATURES

lon,COIthel

Sun.Earin

Ecl Em

RESOURCES

T4;1,11Pnet san 00.100,0

thee th V.e en act ismAny Eclipses kasenEcliplo WO.Shit, Your Stec,.

Coverageexpo tatOdu Aug.( I 1099

ream Europo ana

won... Ansi* s. IA*

Risslindmrt

Eaalwatoesum annatta a fiatRan alums mass tans aml a

In Amara, TUrkff.

!Ms tag, WO Imo Ms0 irttlUdidbbs illagOS 01 0,4 001010 area Pada, Olin an 015 ttu000nall1 Mikw Odytlase.up Minn

-rtk-' 4,41 tot At tTs0000s ...la SAS.,

(limos me Mos, n Packs Ltsnd sod Tin.)ABOUT 1545 etre ; PROJECT PARTNERS

Ms mu thel the h. W.thst ewe.

'.47.". Cf.ethethatethen th 0. a RAW,

Figure 4. Solar Eclipse Page.TheExploratorium's Solar Eclipse Webcast page

hosted by Yahoo! Broadcast.

areas were packed with visitors who by all accountsgreatly enjoyed theWebcast. On theWeb, our visi-tors enjoyed the event, too, but the museum visi-tors a more universally favorable impression.

In I 999, the Exploratorium produced aWebcast onthe Science of Chocolate (http://wwwexploratorium.edu/chocolate) forValentine's Day.The event was hugelypopular on the museum floor.Well over 500 peoplevisited and the museum sold a record number ofmuseum memberships; but on the Web, the eventwas not as popular as anticipated. For obvious rea-sons, meeting scientists, pastry chefs, andchocolatiers, and getting to sample chocolate wasmuch more appealing to museum visitors than toWeb visitors. In addition, museum visitors seemedto enjoy the two-hour format, but Web visitorsdidn't stick around to watch the entire show. Simi-lar criticism was offered from some Web visitorsabout the length of the Eclipse Webcast, which wasalso two hours in length.

Just how long aWebcast should be is a difficult ques-tion to answer.The average duration of a Web visitto the educational resources examined in "Who'sOut There? A Pilot User Study of Educational WebResources" was about eight minutes. Webcastingcan hold visitors' attention for longer periods oftime, but at the Exploratorium, the number of view-ers frequently dropped off during the course of abroadcast.

The Sweet science of CHOCOLATE0 10. sthelees..01114.01, 10, 1.00 IkaPlalertheSweetSaaseeaft.. Ca mot Tn. 61/1In

11-0I-21-71i*tins mast +as mmluted brlltthethatherth Tan Fraartst04maismon as art. scene. anci mama.theth00..

ye Ma.' me host of Watt Coast Lim

Mrs 50005.01notalMo. LC=Nadi la MakA*K00031r

haolai 0,0k OfIto kl0S1

/MCI tkomAMIN/Ws/ wrtMl (MON

Chef :PRIAMRums

040005itaM05.75 Liam MAP

0 Mak+) 0%1449random!

lake *MesaMaSanantan

ElsmotalataLroaa Find oLdWpa 0103 Maka

OnCErtail

Figure 5.The Exploratorium's Sweet Science ofChocolate Webcast Page.

Another factor is how people use their computers.Few people regularly use their computers to viewlengthy movies or TV shows (even when they canfill the entire screen, which DVD players can do).Generally, computers are used in active ways,work-ing, surfing, clicking aroundit's a lot to ask of aviewer, to sit at a computer for two hours and justwatch the screen, particularly when the size of thevideo window is so small.

Most Webcasts have worked very well as a mu-seum event.Visitors to the museum floor can com-municate with each other and with scientists orother guests; they can watch videos or demonstra-tions, and interact in other ways. If the presenta-tion on theWeb could somehow be more dynamic,perhaps by injecting other types of media, or bysomehow involving users more actively in theWebcast, they might stick around longer and theexperience might be richer. In other words, per-haps there is a way to replicate the qualities of themuseum experience for remote Webcast viewers.

In retrospect, the "Webcast pages" for the Eclipseand Chocolate events could have been much moredynamic. Both pages work in the same way (and infact, are similar to most other ExploratoriumWebcast pages).The user clicks on a link launchingan external player that shows the video.These pagestook a great deal of time to develop, making surethat links to "help pages" were easy to find, linking




to other resources, and creating a "look and feel"that enhanced the presentation of the topic. Theproblem with the design is one of function ratherthan one of form.

Both sites contain images and text-based informa-tion that provide background information and cre-ate a context for theWebcastAdditional resourcesgreatly add to the visitor's experience.They couldgenerate anticipation for events yet to come, andprovide some longevity to the sites, bringing visi-tors back well after the excitement of live eventshas dissipated.This approach was developed whileexperimenting with on-demand video.

Unfortunately, there was never the time to developa method to allow visitors to browse these re-sources (or to "push" additional media types) dur-ing a live event. Here's what would happen:When aWebcast viewer wanted to explore these additionalresources, their video window would disappearbehind the browser window.These resources couldbe used before or after the eventbut not during.Unlike sites that use on-demand video, you'll misssome of the liveWebcast program if you try to viewthese additional materials.

Both the Eclipse event and the Sweet Science ofChocolate did allow users to email questions,whichprovided a limited feedback mechanism. A better

method was chat areas, but again these were sepa-rate links that if clicked on sent the user's videowindow behind the browser window. For many visi-tors the Webcast experience was a solitary one,with no interaction with other visitors or with theproducers or guests of the event.

The Webcast Viewer: A PossibleMethod For of Presenting LiveVideo

MostWebcasting examples work in similar ways tothe Exploratorium Webcast pages. The focus hasbeen primarily on the video stream itself. Rarelyhave museums experimented with two-way com-munication such as chat or messaging duringWebcasts. Keeping in mind how most Web visitorsseem to use their computers, I've given somethought to an alternative way of presenting museumWebcasts, one that would make additional materi-als available and that could incorporate user par-ticipation and interaction, all of this while not inter-rupting a continuous program stream.

Some streaming media service providers and Appli-cation Service Providers (ASPs) have or are in theprocess of developing custom interfaces for busi-ness presentations. For example, Presenter.com(http://www.presenter.com) uses a pop-up windowthat allows the user to view a video stream andslide presentation. Obviously, for business presen-

0. Netscape:Webcist VieWer'

museum name & logoevent name

Select Speed028k 056k 0100k 0300k ER::

slide/text screen

Slide I, Slide Z Slide 3

Weesbscaogsit Registered User: Jim Spadaccini513

Send Message

I V /7114,1/ 4.0, 4 VG J.., J4.41 ICU U I. 1/ VC14.4.477...>Fred P.: I liked the presentation last week.>Jean M: I think the Webcast is great. When is the next one? Iheard it was next week.>Sam S.: 1 Ash they'd go back to the remote connection. I

123 wekag want to see more from there.

a

figure 6.The"Weinast Viewer"

1091 10 @Archives & Museum Informatics, 2001

BESTCOPYAVAILABLE


tations, this is a must. Certainly the addition of asimilar feature could help enhance museumWebcasts for Internet audiences.

An interesting real-time messaging system calledSPARC (Space Physics & Aeronomy ResearchCol laboratory (http://sparc- I .si.umich.edu/sparc/central/page/SPARC) deserves mention, too. It en-ables Web visitors to chat on a page in real-time.On the site, others who are visiting the page arevisible, either as guests (with no chat abilities) or asregistered users (with the ability to chat). This issimilar to popular sites that use Java IRC.

Inspired by these interesting examples, the authorhas developed a mock-up of a proposed "WebcastViewer." It incorporates the slide area and a mes-saging feature, along with an embedded video win-dow into a pop-up browser window. Frames areused to present the different features within thewindow.

The embedded video screen can support a varietyof connection speeds.The mock-up incorporates aWindows Media player,but RealMedia or Quick Timecould be used instead.

There is an area for slides or text, a window thatcould support any HTML document.A scrolling fea-ture could also be added. A "zoom" feature couldbe used to allow visitors to examine details of thematerials presented.An "edit" feature would allowusers to set preferences such as image types orsizes.Additional links (which appear as"Slide I" etc.)could permit users to browse slides, or the slidescould be "pushed" automatically as the Webcastproceeded.This could be done with a CGI (Com-mon Gateway Interface) script, Javascript, or Javaapplet.Thumbnails of the images could also be usedwithin the window.

The slide area could provide space for animations,or even applications. Obviously, this would requireadditional resources for development. Flash,Shockwave, or Java technology could be used.Thisroute would require that the end-user have theproper browser or plug-ins to view or interact withthe file, but it could certainly add a new dimensionto a liveWebcast.The real benefit of a slide area formuseum presentations is that artwork or scientific

charts could be shown at an adequate size for visi-tors to properly see the details.

The bottom of the figure contains a Webcast mes-saging frame that allows users to chat with othervisitors and/or museum personnel. Registered us-ers would have chat privileges while unregistereduser could "lurk," that is, observe the chat withoutparticipating.The "edit" feature would allow usersto change their names and/or other information.The chat area could be moderated, although regis-tration would provide a degree of protection fromunwanted graffiti.

The window would serve the important purposeof displaying the number of individuals who aretuned in to theWebcast. Currently, viewing is a soli-tary experience with no indication that others arewatching or viewing the event with you.During smallevents visitors could chat and carry on conversa-tions about the topics presented byWebcast.Withlarge-scale events it would very difficult to carry aconversation, but viewers might be able to influ-ence the direction of the Webcast with their com-ments, which also might be of interest to otheronline visitors.

Technical Issues

The technology exists to create a working versionthe Webcast Viewer presented. (It should be men-tioned that this was not always the case.The "em-bedded" video screen was not well supported onsome platforms and with some older versions ofbrowsers.) The Webcast viewer simply combines anumber of existing technologies in one pop-upbrowser window using frames.

The question of bandwidth is a tougher one to an-swer. Sending additional text and/or images may in-terrupt streams for some users connected at slowerspeeds.Webcast streams bit rates may need to bereduced to allow for more"overhead," insuring thatthe viewer could receive additional media withoutinterrupting the continuous stream. In addition, the"edit" options for both the slide and message fea-tures could contain preferences to limit the size orthe number of times the screens"refresh."Anotherapproach might preload slides or images allowingusers to store them locally before theWebcast be-gins.

110@Archives & Museum Informatics, 2001 1 1 1


The Archive

Most Webcasts are archived, leaving an on-demandclip usually of the entire presentation. With theWebcast Viewer, archives could be shown in muchthe same way the live broadcasts were. Schoolgroups or other organizations could visit the archiveat the same time using the built-in messaging sys-tem to communicate during their visit.

Conclusion

Resources that are flexible and that give visitorschoices in how they interact with streaming mediaseems to work best. Multiple media types providethe user with options on how to use the resourceand work around the technical limitations that manyusers have in viewing streaming media.Webcastingis much more technically and programmatically com-plex than producing on-demand clips.WhileWebcastevents have been very successful on the museumfloor, informal feedback suggests that the experi-ence on the Internet has been mixed.

On the floor, visitors can interact with each otherand can watch a variety of demonstrations that areoccasionally presented simultaneously.As a possiblesolution, the WebcastViewer attempts to incorpo-rate some of the features that make "in person"museum Webcast events so appealing, by creating amultimedia experience and providing the potentialfor user interaction.Some of these qualities are thesame ones that seem to work in presenting on-demand video.The Webcast Viewer could presentusers with choices, makingWebcasts more interac-tive and more compelling.

The reality is that, from a technical standpoint,streaming media is not going be "TV quality" anytime soon. While the promise of broadband willeventually change the situation, it will be years be-fore Webcasting can compete with television and

even then it will not be universally implemented.Recognition of this went into thinking about howtheWebcastViewer might be designed. Incorporat-ing additional media elements and some sort of a"two-way" mechanism will surely enrich the visitorexperience. Otherwise the Internet is just a broad-cast medium and not a very robust one at that.

The Webcast Viewer is that it could be used as a"turn-key" solutionused over and over again tosupport a variety ofWebcasts and topics.ASPs likePresenter.com are trying to do just that, mostly forbusiness applications.The design and technical speci-fications of the WebcastViewer could be modifiedand improved after each broadcast.

Perhaps as more museums begin to experiment withWebcasting, new approaches will be developed. Inthe last two years, several museums have conductedWebcasts, among them the Canadian Museum ofCivilization (www.civilization.ca/membrs/theatre/iqqaip I e.html), the Franklin Institute (http://webevents.broadcast.com/unisys/fi042700/register.asp), the Denver Museum of Natural His-tory (http://event.webcasts.com/dmnh2000/event.html), and the Cleveland Museum of Art(www.clevelandartorg/byzantine/webcast.htm). Itwill be exciting to see what the next few years willbring, as more museums venture into this arena.

The Webcast Viewer is just one way of visualizingwhat could be done to make Webcasting more ap-pealing to museum Web visitors. Any method fordelivering other types of media along with a videostream would help make the experience for onlinevisitors more informative and more exciting.Takingadvantage of the unique two-way nature of theInternet would help make Webcasts a more engag-ing, less solitary experience for remote museumvisitors.The technical elements to make this pos-sible are already out there. Perhaps it's time to be-gin the experiment.

111I I 2 © Archives & Museum Informatics, 2001

PersonalizingWeb

Experiences

112

Co-operation Metaphors for Virtual Museums

Thimoty Barbieri, Paolo Paolini, Dept. of Electronics and ComputerScience, Politecnico di Milano, Italy

Abstract

Cooperation Metaphors are sets of rules to support interaction and collaboration between users who want toexplore complex content and information together.The rules determine how the collaborative community can becreated and managed, how members of the community can operate on their own or can cooperate with othermembers. Different types of situations, tasks and user roles determine different behaviours and therefore needdifferent metaphors.The paper will present the general issue, describing in general what cooperation metaphors areand how they can be defined.A number of examples, related to typical activities for virtual museums (i.e. museumson the Web) will be used in order to give a practical understanding of what cooperation is, or can be.The paper willalso argue that virtual metaphors lack some of the features of real-life cooperation, but, on the other hand, can alsooffer unexpected, powerful and effective possibilities not available in (traditional) real-life. Finally, the paper willdescribe how collaborative activities for virtual museums can be implemented using today's tools and applicationsfor collaboration on the Web; such as Net2gether, Microsoft Research's Virtual Worlds, and the Web Talk series.

Introduction and Motivations

In most cases, users today access the shared spacerepresented by a web site in isolation, with littleawareness, or no awareness at all, of what otherusers are doing"at the same time". In a truly coop-erative environment, by contrast, different users tryto accomplish something together, accessing theshared space simultaneously and trying to collabo-rate to reach their goal. Several different goals maybe the subject of cooperation: teaching/learning,shopping, design, supervising and monitoring, con-sulting, etc. Several different behaviours are associ-ated with the notion of cooperation, but we shoulddistinguish between two extremes:

the cooperating users are peers, with similarknowledge and similar powers.

one or more of the users control the situation,with knowledge or powers superior to those ofthe others, as for example tutors with pupils,shop-clerks and shoppers, tourist guides and tourists,etc.

One basic observation informing our work is thatin the real world most activities are performed, ifpossible, in cooperation. Cooperating while tryingto accomplish something very often appears moreinteresting, more engaging, more amusing or simplymore efficient,

This initial observation is coupled with the fact thatvirtual shared environments have existed for a num-

ber of years. Initially they were text-based (IRC-II,MUDs, MO0s), allowing users to exchange typedmessages, but they have evolved (Blaxxun Commu-nity,Virtual Worlds) introducing 3D graphics, audio,and other advanced features. In these "virtualworlds", as they are usually called, users are repre-sented by virtual objects, called "avatars"; users can"move around", cooperatively interact with the vir-tual world and its objects,"talk" each other, etc.

These worlds allow a great deal of interaction amongusers, but they lack some of the features that wefeel are crucial for an effective collaborative visit toaVirtual Museum:

User interaction seems to be the goal, rather thanthe means to accomplish something. Users amusethemselves by interacting, but they are not tryingto get any specific result out of their experience.Visitors to a Virtual Museum, on the other hand,share the goal of better understanding the contentand the background of the museum.

The amount of information being exchanged amongusers, or being made available by the world, is rela-tively small and loosely organized. Sophisticatedmuseum sites by contrast convey a large body ofsometimes very difficult knowledge.

The cooperation is loose, in the sense that usersare free to behave as they want, with few constraintsor enforced patterns. Our experience, however,

113Museums and theWeb 2001:Proceedings


BEST COPY AVAILABLE

Barbieri and Paolini, Co-operation Metaphors for Virtual Museums

shows that cooperation must be organized and fol-low precise guidelines in order to be effective.

Based on this we have been working on the follow-ing ideas since 1998:

Access to a complex museumWebsite can be mademore interesting, engaging and efficient by allow-ing a certain degree of cooperation among theusers accessing the site at the same time.

The cooperation must follow well organized guide-lines, in order to be effective at reaching a specificgoal (it could be learning, or shopping, etc.).

A 3D virtual world, with avatars representing us-ers "being there", can be an interesting space foran efficient and engaging cooperation.

The information directly represented in the co-operation space (the 3D world) can't reproduceall the information of the Web site - the object ofcooperation. Too much information, in fact, willclutter the cooperation space, making it less use-ful and effective.

The cooperation space must offer enough infor-mation to allow the user to locate and find theobjects on the Web site; once the parts of inter-est are located, there must be gateways allowingthe user to "pass" to the selected portion of theWeb site.

With the above prerequisites, a development envi-ronment,WebTalk-1 (WebTalk, Barbieri, 1999) wasdeveloped and a few applications were developed.The best known of these applications was devel-oped for the Museum of Science and Technology ofMilan. The application, named "Virtual Leonardo"(Paolini, 1999), allowed users to co-operate whileaccessing a site with a number of pages describingdifferent machines invented (designed but neverbuilt) by Leonardo da Vinci.Virtual Leonardo waspresented in previous editions of Museums and theWeb.

Virtual Leonardo allows users to visit the museumtogether, exchange opinions, and interact with theworld and with each other in a number of ways:interactive gateways allow the users to visit the pagesof the Website though standard Web technology.

Our experience with Virtual Leonardo, reinforcedafter collecting usage data and impressions fromthe users (Barbieri 2000), convinced us of the ne-cessity to support the creation of such collabora-tive environments with a theory of design patterns.These are the subject of this paper.

Designing for Collaboration

Within a cooperative environment the goal, for auser, is to accomplish something specific (e.g. to getwork done, to learn something, to buy something,etc.) through cooperating with other users. Theother users can be at the same level of knowledgeand power, or can be more expert and/or havegreater control over the application.A number ofvisitors in a museum, for example, can be at a com-parable level, while a museum guide is a user withmore expertise and with the power to "take" thegroup of visitors around.

When the users interact with each other, they fol-low patterns of interaction that we call "coopera-tion metaphors".The goal of this section is to ana-lyze the general features that cooperation meta-phors must satisfy, in order to be effective, and atthe same time engaging and compelling.The discus-sion that follows is based upon literature and ourexperience gained with WebTalk-I (Barbieri, 1999)and the deployment of applications.

First of all, in a cooperation environment, it is im-portant that all actors share the notion of a com-mon state, bound within time and space.The inten-sity with which this feeling is conveyed to all usersdetermines the level of awareness (Gutwin, 1997)within the environment.The higher the awarenessof the shared state, the better the cooperation be-tween users: thus awareness is the primary effectwe must create to allow users to work and discusstogether. In a virtual space described in three di-mensions, users can experience perceptual stimulithat are more similar to everyday life.By processingand understanding these stimuli, they can interactwith the environment and the other actors, increas-ing their awareness, and creating in turn new stimulifor the others. In a collaborative system, this leadsto the phenomenon of prediction; that is, participantscan anticipate, within a short time, the actions ofthe others, because the number of informationaldetails they have regarding the space and time they

1141 16 ©Archives & Museum Informatics, 2001


share with others (in one word, their awareness ofthe system) can lead them to such conclusions. (Inthis respect,for example, the "slow motion" of ava-tars in a 3D space can be more effective than fastmotion or jumps, since the other users can betterpredict where an avatar is going if it is moving aroundslowly). Other important consequences of aware-ness are the possibility of learning indirectly fromother people's activity, or of gathering informationfrom the state or behaviour of the artefact in theshared environments, or the possibility of using ges-turing or indirect communication forms to supportone's communication (e.g. "Follow me this way","Take this..", etc.) when coordinating multiple ac-tions for which more users are needed.

The problem in conveying the correct form ofawareness is often application-related, meaning thatnot all kinds of collaborative information are alwaysneeded to perform a task within a determined en-vironment.That is why before deploying a collabo-rative system it is important to design in advancethe means by which information will be presentedin three-dimensions (Spatial Patterns), and to deter-mine precisely the rules by which users are able tocooperate with each other, and with the environ-ment, and to enhance their awareness during theiractivity. We call this set of rules cooperation meta-phors.

In our vision,a cooperation metaphor (Barbieri, 2000)is a set of basic rules that describe the differentmodalities of interaction between users and be-tween users and their environment. These rulesencompass various aspects, such as the way userscan gather in groups to talk to each other or navi-gate virtual space, or how visualization of the stateof the artefacts and of the avatars (the figurines whichrepresent the current position of each user in space)is to be performed. By making decisions about allthese aspects, it is possible to define a global de-scription (a metaphor) of the possible forms of col-laboration of users within a system. In the followingsection, we present an abridged list of these smallrules,which can be freely selected to form any meta-phor that is felt suitable to regulate collaborationwithin a certain application.

In addition, the design of the application also callsfor deciding which spatial patterns are to be followed.Such decisions include whether the world will be

represented in an abstract way or by mimicking re-ality in a simplified way (Bridges and Charitos, 1997,Benford and Snowdon, Hearst, and Karadi 97).Moreover, we must decide how to represent, in aconsistent fashion, important visual elements thataid the navigation in space of the users, such as places(Cerulli, 1999), thresholds between one space andthe other, landmarks to assist users in finding theirown routes within the space (Charitos and Ruther-ford, I 997), spatial hyperlinks to jump from one pointof the space to another (Campbell, 1996, Charitos,96), and so forth.We will not investigate this com-plex issue in detail in the present paper.

Our (limited) experience in designing cooperativeapplications has shown that although the basic co-operation metaphors are the same, the way to ac-tually implement them or to assemble them, changesfrom application to application.The role of the de-signer of a collaborative environment, therefore, isnot solely limited to gathering contents and draw-ing 3D geometries, but includes conceiving a con-sistent pattern of cooperation metaphors suitablefor the different application situations; the design-ers, in other words, must describe the rules for thecollaboration between users.

A generic package for supporting cooperation overthe Web, therefore, must not provide a fixed set ofcooperation metaphors, but rather must presentthe designers with a vast range of possibilities thatcan be selected and tuned to the specific needs ofthe application.This functionality is lacking in all theVirtual Communities tools available today, and thatis why we are designing WebTalk-II, a collaborativeenvironment where powerful cooperation meta-phors can be combined and tuned at will.

Cooperation Metaphors for VirtualMuseums

With some oversimplification, we may state thatcollaboration can be represented with two visual-ization paradigms corresponding to 2D and 3D rep-resentations.While both representations are in facttwo-dimensional, since the visualization field is a'flat'computer screen, the main difference between thetwo is that the first visualization paradigm causesthe users to think in term of flat geometric ele-ments and flat structures (windows, icons in win-dows, buttons laid in a bidimensional grid...), while

115©Archives & Museum Informatics, 2001 1 17


the 3D paradigm utilizes perspectival projection toconvey a sense of spatiality and immersion.Very of-ten this paradigm takes advantage of a subjectiveview; that is, the users infer position in space byviewing a perspective representation of immediatesurroundings, but without the ability to see them-selves (exactly as in Real Reality).This is usually theapproach taken in gaming and web-oriented virtualreality. In high-end virtual reality systems that makeuse of complicated devices such as gloves and head-mounted displays, there is an effort to convey muchmore immersion to the user, by tricking the humaneye into seeing depth and space in a synthetic com-puter generated scene.

Most of the collaboration patterns can be imple-mented in either 2D or 3D. In 3D, however, it ispossible to conceive a set of collaborative situa-tions that are impossible to represent in 2D, andwhich can convey a very heightened mutual aware-ness. In the following we will list the most impor-tant collaborative elements, and we specify how theycould be implemented in 2D or 3D. For some ofthese cooperation patterns, it is possible to describesituations that have no match in reality, but can beeffectively used for the purposes of particular ap-plications.We will point out some of these cases.

At the end of this survey, we will describe samplescenarios with which these collaborative sets canbe composed into collaborative behavioural patternsto support a virtual museum application, by usingcommercial or research tools available today.

Distribution of the Shared State

A shared state is a collection of information aboutthe status of each participant in the application. In3D environments, it could be the avatar positionand its movement or idle state, or the status of ashared resource, like the writing on a collaborativewhiteboard. Each user can participate or be iso-lated from the shared state (and in this last case, itcannot be part of the collaboration mechanisms).The sharing of the state works on two levels:

Sharing Space: All the modifications in spaceare shared (movements, changes in the boundingenvironment)

Sharing Time: The unit with which time passes,and current time and speed of the animation, are

common between th e users.It is possible for users to cooperate in a commonenvironment only if they share the same notionof space and time.

Integral Sharing:All users within the system orthe application share the same events.All of themthus have the same notion of the environment ata given moment.

Group Sharing: The shared state can be parti-tioned in sub-groups; that is, several parallel dif-ferent states can exist for the same environment.Each participant can belong only to one of thestate partitions at a time.

Mutual Exclusions

If space and time sharing are not enabled simulta-neously, it is not possible to support collaboration.If space and time sharing are enabled, an Integral orGroup sharing mode has to be specified. If Integralsharing is used, it is not possible to enable Groupsharing and vice-versa.

Enhancement of these collaborative patternsover reality

The possibility of creating applications in which timeis counted differently between users, or in whichfor a same environment there exist several pos-sible states at the same time, has no counterpart inreal reality.The first option can be used to let usersexperiment with artefacts or situations with differ-ent speed and timings.The second option is inter-esting to create multi-path situations in which dif-ferent choices give rise to different events. An ap-plication could model this line of thinking and sup-ply the users with an original way to experimentwith this, using Group Sharing.

Grouping Mechanism

Grouping mechanisms refer to rules set to themanagement of user groups. Groups are a meanstocluster users by interest or activity; each groupshares particular cooperation elements, and behavesfollowing a common pattern. Each group has to beidentified by a unique name, so for example a groupcreated to discuss Modern Painting following a re-nowned critic may be entitled "ModernArtCritique",while a group designed for free browsing around

116I I 8 © Archives & Museum Informatics, 2001


the museum resources can be called"StrollingAroundTheMuseum". Group managementis handled by the following four subcategories.

Group Creation: rules how groups can beformed within the system

Dynamic:it is permissible to create a newgroup, with no restrictions. The user thatrequests the creation of the new group iscalled the owner or creator of the group.

Fixed: the designer of the collaborativeenvironment prepares in advance differentgroups with different names (as at pre-ar-ranged thematic conferences, for example).The owner is the designer himself or any userthat can be determined by a pre-arrangedpassword (see Leadership Management).

Group Protection: defines how a group can beentered by a user

Free: any user can join the group at lei-sure.

Password: the user can join only if heknows the related password.

Fixed Quota:the user can join only if thecurrent number of users in the group doesnot exceed the maximum number requestedby the designer of the application

Banning: it is possible to attach to thegroup a list of particular users who cannotjoin the group under any conditions (theyare banned).

Kicking: other users are allowed to force a userout of the current group. Usually this operation istied to some kind of privileged status.

Group Disbanding:defines a criterion by whicha group has to be disbanded.

Persistent: the group remains active andjoinable even if no users are participating tothis group.

Dynamic: the group and all its rules aredestroyed when the last user leaves.

Forced: a particular user can force thedisbanding of the group, scattering all its cur-rent participants. This can happen only ifthere is the availability of a free group tohost all dispersed participants (usually calledglobal).

Leadership Management: in many collabora-tive situations the users do not cooperate on thesame level, but one or more participants may havespecial abilities (they lead the way, or they showan object, or are the only participants enabled totalk).Leadership models a particular privilege withinthe collaboration, and the mechanism by whichleadership is acquired and passed on has to bespecified.

Forbidden:once leadership is acquired, itcannot be passed on.

Password Protected: leadership can beacquired by providing the proper password.

Op-Passing:a user who holds leadershipcan yield it to another user in the group.

Inheritance: users with leadership cangrant leadership to one or more other us-ers, without yielding their own leadership.

Tokenized: leadership is a token that canbe acquired by the first user who requestsit.There are as many leaders as leadershiptokens available. For someone to acquireleadership, a token must be available. Userswith leadership can yield leadership, therebyfreeing a token.

Creation Based:in free group formationmode, leadership is automatically granted tothe user who creates a new free group.

Mutual Exclusions

A group can be either dynamic or fixed. Similarly, itcan be either free or password protected, but bothtypes can have a fixed quota, banning and kicking.

117@Archives & Museum Informatics, 2001 1 19


Only one type of disbanding can be specified. Lead-ership management can be defined by using moreelements at a time; that is, leadership can be Cre-ation Based and Forbidden, or Creation Based andTokenized.Some of the management elements can-not be used together, e.g. leadership cannot beTokenized and Inherited.


Group collaboration in real reality is usually de-scribed by a Creation Based, Forbidden leadershippattern, with a fixed/fixed quota group creation anddynamic disbanding.There is no complex notion ofleadership management in real groups. In virtualcollaboration, leadership can be spread over thegroup and be used to control enhanced collabora-tion facilities. (See for example visualization andview-point shifting capabilities).

Information Flow

Information flow rules can be attached to any othercooperation element to specify which are the di-rections and modalities in which information canbe transmitted during collaboration.

Pull: information transmits from the informationprovider to the information consumer becausethe consumer explicitly asks the provider for apiece of information (pulls information from theprovider).

Push: information transmits from provider toconsumer, because the provider explicitly pub-lishes new information for the consumer.

Mutual Exclusions

Push and Pull modalities can be used simultaneously.For example, a textual message or information itemcan be pulled or pushed, even if usually the messagetext is pushed to the user when new messages areincoming (as phone calls are pushed to your cellphone). Push and pull are particularly effective invisualization, where it can be decided if one userwants to see from the point of view (through theeyes) of another user, or conversely if one userwants to force other users to see what he sees

113

(pushing his view onto others). Moreover, it couldbe possible to push one's position to others, to forceother users to determined positions in the virtualspace, and so on.


Push and pull are in general common modalities fortextual and vocal communication, but are uniquelyan achievement of virtuality when it comes to visu-alization and physical position.

Visualization

Personal View: the kind of visualization each userhas of the environment in which he is immersed.

Normal: purely subjective, from his owneyes

Proximity:from a camera external to theavatar, but placed near to it

External: from an external camera thatchases the avatar, but at a distance

Bird's View: from a fixed camera placedin an elevated place

External Avatar: from the eyes of an-other avatar (a tour guide or a friend)

Scene View: modifies how other avatarsare represented in the environment

Global: all avatars belonging to all groupsare visualized

Group: only avatars of the same group ofthe user are visualized

Selective:only some particular, selectableavatars are visualized

Mutual Exclusions

Personal views can all be used, or forbidden for par-ticular application-related reasons. However, onlyone kind of view at a time is possible. Scene viewsare all mutual exclusive.

I 20 ©Archives & Museum Informatics, 2001


Enhancement of these collaborative patterns

Visualization elements strongly enhance the dailyexperience of collaboration in reality. In virtual re-alities it is possible to see oneself acting, or to in-carnate other users by seeing from their eyes.Thisyields enormous gaming, demonstration and collabo-ration potential.

Movement

Specifies which kind of movements are allowed tousers in the environment. Movements can be re-stricted for educational or navigational purposeswithin the particular application.

Normal walking: allows users to wander freelyin the environment by contiguous movement(walking)

Near Jump: allows jumping in space next to an-other avatar of choice

Zone jump: allows jumping in space inside a pre-defined landing zone (also known as `teleporting)

Mutual Exclusions:

All movement modalities can be allowedsimultaneously.


The usual pattern in reality is normal walking incontiguous ranges of time and space. In virtualitythe possibility of hyperjumping, not present in real-ity, enables a more effective collaboration.

Self-Representation

Self representation describes the means by whichthe particular state of the user can be represented.These elements can be enabled or disabled accord-ing to their relevance to a particular application.

Emoticons:emoticons are short sentences or ad-jectives, often coupled with symbols, drawings oranimations, that represent the emotional state andfeelings of the user.They are used to better con-vey a meaning, and thus enhance cooperation.Example are: *sm iles, *shrugs, *whistles, *frowns,*yawns,*blows a kiss,*strikes a pose, etc. etc.

Degree of Presence: degree of presence indica-tors indicate the involvement of a user within anenvironment.The representation can be textualor symbolic. Some examples are: busy, do not dis-turb, away, extended away, free for chat, etc. (TheICQ)

Text-based Communication

In most virtual environments, the most effectivemean of communication is still text based: userstype in what they want to say to other users, andread the answers. Some environments also offervocal communication, and there is a pattern aboutgoverning vocal communication that we will notdiscuss here. Still, it is hard to operate in a 3D syn-thetic environment and talk. For a number of rea-sons, typing is still the favored method at the mo-ment.

Free:everybody can talk to all users in all groups,and receive messages from them

Group: talking and listening is limited to withinthe group

Group Restricted: like Group, but only userswith leadership can talk; everybody else just lis-tens (reads).

Whisper: one user talks directly to another user,without being heard by the rest of the group

Private: one user establishes a permanent talk-ing session with another user, like a group com-posed of only two people

Multiple Whisper: like whisper, but with mul-tiple recipients

Mutual Exclusions

Free mode, Group and Group Restricted are inmutual exclusion.The rest of the elements can beused in combination with these first three.


These elements model how talking can happen inreal collaboration sessions; however, multiple whis-per and group restricted elements may be harderto enforce in reality than they are in virtuality.

1 1 9©Archives & Museum Informatics, 2001 121


Using tools and metaphors for creat-ing collaborative museums

We will now present a short survey of some soft-ware tools that enable us to create collaborativevirtual environments for museums, commenting onhow they can be used to enforce collaborationmetaphors in the ways we presented them. Someof these are strictly 3DVirtual Environments (Zydaand Singhal 99), which strive to present a perspec-tival representation of the environment in which allparticipants are immersed. Others simply provideusers with a way to collaborate over the Internetwith usual 2D metaphors. Still, they can be usefullyemployed for museum applications, and they do usesome of the cooperation elements we mentionedearlier.

WebTalk-1

Web Talk-1 (see papers in previous M&Wconferences,(Barbieri 2000.T. Barbieri 99)) allowsyou to draw a 3D environment that represents avirtual exhibition, and publish it on the web. Usersthat connect to the web page are presented withthe 3D representation and can navigate in it to-gether with others, seeing the other users and chat-ting with them.

There are two co-peration metaphors, fixed andimposed by the framework:

Fig. I. In Web Talk-1, collaboration is supportedvia a fixed set of collaboration metaphors,

like free-formed groups, and a guided tour inwhich leadership uses a tokenized

mechanism.

Free Visit to the Museum

Groups can be created and disbanded dy-namically, and they allow free access.lt is thuspossible for users just to form their owngroups, or join any group, and start visitingthe museum freely, like in reality.Text com-munication is grouped, and there is a whis-per mechanism. Movement can be by walk-ing, flying, and zone jumping to importantspots of the museum.Visualization allowedis Normal.

Guided Tour of the Museum

Leadership in groups is tokenized, so anyparticipant can acquire it and become a TourGuide. In this metaphor, other participantscan follow the one who acquired leadership,and pull inside their point of view what theguide is looking at (in simple words, seeingfrom the eyes of the guide), using the Exter-nal Avatar visualization mode. This enablesus to make very interesting virtual guidedtours inside the virtual museum.

WebTalk-11

The goal of theWebTalk-II framework is to providean environment in which it is possible to design anddeploy 3D collaborative environments, acting as away to improve the access to"traditional"Web sites.There is no assumption or constraint about thetechnology used to implement the Web sites, norany implication about the design "philosophy". It isclear, however, that well structured and well behav-ingWeb sites, like the ones that could be obtainedusing a proper design methodology (HDM andW2000, Rohel, 1997, Bochicchio and Paolini, 1998,Bochicchio and Paolini, 1999, Garzotto, 2001), aretargets of specific interest. In addition, there is thefurther requirement that WebTalk-II must be ableto interface to Websites where pages are not stati-cally defined, but dynamically generated (it is in factmuch easier to define gateways to a set of staticallydefined pages, than to a set of pages with numberand content dynamically defined), according to userrequests. The JWEB (Bochicchio et al 1999) envi-ronment is one, of the many available, that doesexactly this.

`" 0


SESTCOPYAVA';.


Consider, for example, the need to coordinate ac-cess to a number of pages, for a museum applica-tion.The user accesses an "index page" that allows"navigation" to the item of interest. In a collabora-tive environment, the equivalent of the index pageis a 3D sub-space, where the users move throughvisualizing "gateways" to the indexed pages.

In "Virtual Leonardo" for example, the index is rep-resented by a virtual "cloister", while each page isrepresented by a virtual room. Each room, in turn,contains an exhibit that can be browsed in a col-laborative fashion with other participants.

Unlike Web Talk-I where you can have only freegroups and pulled visualization from a privilegedgroup user (to form a "guided tour group" meta-phor),WebTalk-II allows you to specify (for any vir-tual exhibition one may design and implement), anypossible combination of the collaboration elementsdiscussed earlier in the paper, choosing the bestway for people to work with each other given theparticular nature of the exhibition the virtual mu-seum is proposing. In other words, you can con-ceive any cooperation metaphor you like, assign it aname, and specify its properties by enabling or dis-abling each of the collaboration elements designedabove.

Net2Gether

Net2Gether (http://www.net2gethercom) is an in-teresting tool for 2D based collaboration overthe Web. It provides a chat window under the usualbrowser window. Chat is organized in groups, andeach group has a leader.The leader is able to push

the Web page being visited into the browsers ofthe other participants. It is thus possible to 'share'the 2D navigation of the museum Web site.The toolalso allows pre-recording of sequences of chat mes-sages and of navigation events within the site, andplays it remotely to the users. In this way it is pos-sible to create various exploration paths within theexhibition proposed in the museum Web site.

Since Net2Gether has no notion of what coopera-tion metaphors are, it is not possible to change theway users collaborate. N2G metaphors are thusfixed.They can be classified as:

Guided Tour

Uses Dynamic group creation, with freeprotection, and the possibility of banning andkicking. Disbanding is dynamic as well. Lead-ership is creation based, and there is an op-possing mechanism that allows other peopleto push web contents to the group. It is in-teresting to note that leadership cannot beinherited, as only one user must be able topush web contents to other users. Move-ment andVisualization elements are not used,since the environment is 2D.Text based com-munication is grouped and there are mecha-nisms for whispering and private communi-cation.

Pre-Recorded Guided Tour

In contrast with the Guided Tour meta-phor, this metaphor requires fixed groups,with a fixed number of participants. The

rist Cooperation settingsC up listGuest

master

New Del I

*Group Motion Relation Chat Audio Visualization Shared State ,!

Prntacvin: 74.1 Kicking Banning Ri Fixed size 5 I ,0) Free (-) Password

Disti untie: (*) Persistent OD,nanic 0 Forced

Canstractirtry 171 Dynamic

Apt*: I I Done I

Fig. 2.The designer is able to define the general access structure, and specify which rules to enact forthe collaborative exploration of the museum (creating cooperation metaphors).

@Archives & Museum InfoUi1,2001

BEST COPY AVAILABLE

123


group is also persistent. Leadership manage-ment is forbidden, since it is the system thatdoes the pushing of the content. Communi-cation can be either Grouped or Group Re-stricted.

Microsoft Virtual Worlds

Microsoft VirtualWorlds (Microsoft Virtual Worlds)is an authoring environment for creating 3D Col-laborative Virtual Environments (http://www.vworlds.org). Even if there is not an explicitway to model general Cooperation Metaphors inthe sense we described in this paper,Virtual Worldsprovides the designer with a high degree of flexibil-ity, allowing them to define, for each object, sets ofattributes and properties describing them. Some ofthe most common cooperation metaphors in vir-tual museums deployed with Virtual Worlds mightbe:

Free Visit

Virtual Worlds does not directly supportthe notion of groups, but divides users inrooms, with the idea that only users in thesame virtual room can talk to each other,like in any real environment. So even if com-munications can be considered Group based,the Groups are all predefined and persis-

tent. Movement can be done by free walk,zone jumping and near jumping to anotheravatar.Visualization is extremely flexible, andcan be of Normal, Proximity, External, Bird'sView type, at the designer's will. It is pos-sible to push and pull visualization to anyExternal Avatar.VVV also has a well designedsystem of Emoticons which allow users toexpress their feelings by modifying the pos-ture and expressions of their own avatarrepresentation.

Piloted Guided Tour

Extends the Free Visit metaphor, with thepossibility of attaching a camera to a par-ticular automated Avatar which can play therole of a semi-intelligent and automated vir-tual guide in the museum.The users can pullin the visualization of this particular cameraand enjoy the ride.

Treasure Hunt

-Virtual Worlds augments collaboration byusing the notion of ownership of objects(abridged from the brief sketch of collabo-rative elements presented in this paper). Itis possible for users to take objects fromthe worlds, giving rise to the possibility oforganizing collaborative games or activities

Fig. 3. An art gallery of a museum can be freely navigated with walking movements in Virtual Worlds(left). Virtual Worlds' Emoticons provide a better way to convey information (right).


BESTCOPYAVA/LAR


like 'treasure hunts', where the guide invitesthe participants to look for a particular ob-ject in the museum, mixing entertainmentwith education.

Conclusions and Future Work

The main point of this paper is that for a large rangeof applications, cooperation among users while vis-iting a web site is a very important improvementover current practices. Museum applications belongto this class for a number of reasons:

They support a (potentially very) large body ofknowledge

The users have strong motivations to try to getthe most out of their "virtual visit"

Exchanging knowledge, comments and ideas withother people is an inherent part of an effective(and amusing) visit to a real museum.Why shouldit not be the same for virtual museums?

We have started analyzing the different componentsfor making cooperation effective, recognizing thatdifferent situations require different organizationsand different actions for cooperation.We have usedthe name "cooperation metaphors" to identify dif-ferent cooperation solutions.Also,we have revieweda number of tools, from the market or our own, inthe light of the needs of implementing sophisticatedcooperation metaphors.

In the future we will work in the following direc-tions:

Building new museum (and non-museum) collabo-rative applications, in order to gain further un-derstanding of the key issues.

Conducting theoretical and empirical researchabout the primitives making up effective coop-eration metaphors, trying to obtain a more com-prehensive list of elementary possibilities and amore complete set of ways of organizing them.

Investigating the correlation between visual rep-resentations and cooperation metaphors.

Completing the development ofWebTalk-II, while,at the same time, keeping a close eye on indus-trial solutions offered by the market.

Among other application areas, (virtual) museumsremain our favorite area of application for thesegeneral ideas.

References

Active Worlds, available at http://www.activeworlds.com

Blaxxun Community, available at_http://www.blaxxun.com

Barbieri T., Paolini R 1999. WebTalk: a 3D Collaborative En-vironment to Access the Web in ProceedingsEUROGRAPHICS '99, Short Papers pg. 111-113,September 99

Barbieri T, Paolini R 2000. Cooperative Visits to WWW Mu-seum Sites a Year Later: Evaluating the effect, in Pro-ceeding Museums&Web2000, Minneapolis (USA),April 2000

Barbieri, 1.1999. Webtalk-ll: un'infrastruttura per lacooperazione sul web. Politecnico di Milano, Thesisfor the Engineering Degree, Milan, Italy.

Barbieri T. 2000. Networked Virtual Environments for theWeb:TheWebTalk-1 andWebTalk-ll Architectures, in Proceed-ings IEEE for Computer Multimedia & Expo 2000(1CM E), New York, USA, July.

Baresi, L., F. Garzotto, and P. Paolini. 2000. "From WebSites to Web Applications: New Issues for Concep-tual Modeling". In Conceptual Modeling for E-Busi-ness and theWeb (ER200 Workshops Proceedings),Lecture Notes in Computer Science 1921, Springer.

Benford, S., D. Snowdon, et al. Visualising and Populatingthe Web: Collaborative Virtual Environments for Brows-ing, Searching and Inhabiting Workspace, in proceed-ings JENC8

Bridges, H.A. and Charitos, D., 1997. The architectural de-sign in virtual environments, R. Junge (ed) CAAD Fu-tures'97, Kluwer Academic Publishers, Dordrecht.

Bochicchio M.A., Paiano R., Paolini R, 1999." JWeb: anInnovative Architecture for Web Applications", inProceedings of ICSC'99

1 2 3



Bochicchio M.A., Paolini R, I 998."An HDM Interpreterfor On-Line Tutorials", in Proceedings of Multime-dia Modeling 1998 (MMM'98), N.Magnenat-Thalmannand D. Thalmann, eds. IEEE Computer Society, LosAlamitos, Ca, USA, pp. 184-190.

Bochicchio M. A., Paiano R., Paolini R, 1999. 1Web: anHDM Environment for fast development of WebApplications" . In Proceedings of IEEE MultimediaComputing and Systems (ICMCS '99),Vo1.2, pp.809-813.

Campbell, D., 1996. Design in Virtual Environments UsingArchitectural Metaphors: a HIT La Gallery. March Dis-sertation, University of Washington.

Cerulli, C. 1999. Exploiting the Potential of 3D navigableVirtual Exhibition Spaces, in proceedings Museums andthe Web 99, New Orleans

Charitos, D., Rutherford, P., 1997. Ways of aiding navigationwithin VRML Worlds, Proceedings of the 6th EuroplAConference, Edinburgh.

Charitos, D., 1996. Defining existential space in virtual envi-ronments, Proceedings of Virtual Reality Worlds '96,February '96, IDG Magazines, Stuttgart.

Franca Garzotto, Paolo Paolini, Luciano Baresi, 2001."Sup-porting Reusable Web Design with HDM-Edit". inProceedings IEEE Int Conf. on System Sciences, Maui(HW, USA), IEEE Press, January 2001-

Gutwin, C., S. Greenberg, 1999.A Framework ofAwarenessfor Small Groups in Shared-Workspace Groupware.Tech-nical Report 99-1, Department of Computer Sci-ence, University of Saskatchewan, Canada.

Hearst, M., C. Karadi, 19997. Cat-a-cone:An Interactive In-terface for Specifying Searches and Viewing RetrievalResults using a Large Category Hierarchy, proceedingsof 20th Annual International AC M/SIG I R Conference,Philadelphia, PA.

ICQ (I-seek-you). Instant Messenger System, http://www.icq.com

MicrosoftVirtualWorlds, available at http://www.vworlds.org

Paolini R, Barbieri T, et al. 1999. Visiting a Museum Together:how to share a visit to a virtual world, in ProceedingsMuseums&Web '99, New Orleans (USA), March 99,pg. 27-35

Rohel, B. et al. 1997. Late Night VRML 2.0 with Java, ZiffDavis.

WebTalk-I Web Site, available at http://webtalk.elet.polimi.it

M. Zyda, S. Singhal, 1999. Networked Virtual Environments,ACM Press.

124

I 26 © Archives & Museum Informatics, 2001

The Hyper Museum Theme Generator System: Ontology-based Internet Support for the Active Use of Digital

Museum Data for Teaching and Presentation

Peter Stuer, Robert Meersman, Steven De Bruyne,STARLabVUB, Belgium

Abstract

Museums have always been, sometimes directly and often indirectly, a key resource of arts and cultural heritageinformation for the classroom educator.TheWeb now offers an ideal way of taking this resource beyond the traditionaltextbook or school visit.While museums around the globe are embracing the web and putting virtual exhibitions,cultural databases and archives on-line, the educator (or user in general) is still facing the daunting task of integratingthis material into an active document, course, curriculum or presentation.This paper reports on the construction ofa personalized theme creation engine as a possible catalyst to the active use in secondary education in Europe ofdigital media published on-line by selected museums.

The HyperMuseumTheme Generator System (TGS) is part of the HyperMuseum system, a European virtual museumportal (Fig. I ). Its function is to assist in the creation of so-called personalized themes.A personalized theme intends toallow the end user to bring together a unique collection of multimedia objects from the HyperMuseum Server, and tocreate a personalized rendering of the perceived and/or recorded relationships between these objects, realized as anew multimedia document (website,PowerPoine presentation, hypertext orWord® document etc.).TheTGS supportsthis creative expression both during the discovery phase, exploring the collection and discovering thematic relationships,as well as the realization phase, the construction of the resulting documents.The primary target audience is non-expertusers mainly from the secondary education community.

This work is partly supported by the Telematics Program (4th Framework) of the European Commission under Projectnr. 3088 (HyperMuseum).

a Hyper Museum Prototype V. 4 - Microsoft Internet Explorer

Fig. The Home Page of the HyperMuseum sitehttp://www.HyperMuseum.com

:

Museums and the Web 201: rraceedings


BEST COPY AVAILABLE

Stuer etal.,The HyperMuseum Theme Generator System

A discovery system is grounded in a liberally linkedontology service. Ontologies, roughly speaking, arecomputer resources describing application domainsin terms of standardized vocabularies, linking andcategorizing those terms, for example as taxono-mies. In our case, multiple linkage forms, selectableby the user, express and enhance the published col-lection of Hyper Museum digital media as a seman-tically linked network according to principles de-veloped in the DOGMA project (http://starlab.vub.ac.be/dogma.htm) (Meersman, I 999a).Relationships between different objects are sug-gested and derived from the metadato that accom-panies each object (as put there by subscribingmuseums), as well as previously published themesand general "background" ontologies. In exploringthis network the user discovers a (new) paththrough the semantic links covering a theme, i.e. anovel set of relations between the objects. At anytime he can export the newly discovered objectsinto the TGS creator tool, where the realization of

Museum

Metainfo DB(e.g. Access File))

Museum, HCS or bird Party HCS

Media Repository(e.g. images, texts etc.

ProfileSchema

ZT

the theme is ongoing.This creation phase concen-trates on expressing and constructing the new re-lationships between the objects. It maintains a levelof abstraction separate from the details of imple-mentation in a specific medium, e.g. as an on-lineWeb presentation.

Overall HyperMuseum Functionality

The overall architecture of the completeHyperMuseum system consists conceptually ofthree different parts:The Museum Data Centers,The HyperMuseum Service Center (HSC) and theHyperMuseum Client (Fig. 2).

The philosophy of the HyperMuseum is that re-quirements for participating museums should beminimal. It was felt that at this experimental stagethe participating cultural institutions as well as mostof the target market would not be inclined to dedi-cate the funds necessary for data conversion to a

RDS

Z39.5D TargelReplica

Replica

ZC

Query

Z39.9) Client N.

Client, HCS or 'bird Party

Web Server

Result

ThXML

Record

Object

TCS

Expert Colltoration

Theme RecordObisct

Client Are

TMS

MulflingualDictionary

TRS

ThemeRepository

TGS

TGS Server

Client Area

Client

TGS

TGS AplicationClent

repositoryWeb Bowser

Fig. 2:Architectural overview of the Hypermuseum system.



specific application.Therefore, existing data systemsshould if possible be reused without a requirementfor conversion or standardization. Museum data canbe replicated to the HSC servers, or be hosted bythe museum itself on an accessible site. Media ob-jects are served through a normal HTTPWeb server,while corresponding data records are consultedthrough a Z39.50 server, with the HSC acting as aZ39.50 client (ANSI/NISO 1995). For each museumdatabase, a Z39.50 profile is developed to map theexisting (meta-)data fields onto a common struc-ture. At the HSC side the retrieved data is con-verted into an XML format, and packaged in a ZIPfile together with its corresponding digital mediafiles.The result of this operation is referred to as anHMRecord file and will become the basic opera-tional unit inside the rest of the HyperMuseum sys-tem.

The first service offered by the HSC is a traditionalmedia consultation service called the Resource Dis-covery System (RDS) adapted from Aquarelle(Michard I 998).The RDS offers aWeb interface tothe virtual HyperMuseum collection, allowing theuser to query for and retrieve information on themedia objects.Several query methods are available,from generic free text search to structured searchin specific contexts.This service is supported by aTerms Management System (TMS), offering assistedaccess to controlled vocabularies and multilingualdictionaries to facilitate intra linguistic retrieval.

The second service, the Theme Collaboration Ser-vice (TCS) is targeted towards the professionalcommunity.This part of the HyperMuseum eventu-ally wants to offer a common workspace for creat-ing themes based on the records in theHyperMuseum collection.This resulting theme filesare once again based on an XML formalism, whichallows the accompanying Theme Repository Sys-tem (TRS) to make a collection of themes availableto the public in a maintainable fashion. Both theResource Discovery System and theTheme Reposi-tory System are accessible with a Web browser athttp://www.HyperMuseum.com (see also Fig.!).

A third service, the Theme Generation Service, isthe focus of this paper and will be explained in de-tail below

Rationale of the Theme GenerationSystem

Experimenting with technology that can enhancethe use of museum digital resources in secondaryeducation is one of the main goals of theHypermuseum project, and the central motivationunderlying the creation of the TGS.We try to pro-vide an alternative to the passive use of museumWeb sites, which we might characterize as educa-tional browsing, and support a more active type ofusage, by letting the user explore and create withthe objects of study.

Museum Web sites tend to fall into two categories.On the one hand there is the database approach tothe holdings.The users are confronted with a search/query-like interface that allows them to get to thevarious assets in the digital collection.This approachhas as an advantage in that it can easily offer accessto the entire digital portfolio of the museum, andknowledgeable exploration can be made highly effi-cient because of the structured nature of the ap-proach. Furthermore, the elements are more oftenpresented without a contextual bias, allowing theusers to appreciate the items from a personal back-ground.

This approach is supported in the HyperMuseumthrough the RDSTTMS.

A second approach is that of the digital exhibition.Selected holdings are presented in a thematic con-text.This approach allows the exhibitor to convey apersonal or institutional vision of a chosen set ofmedia items.The educational and appreciative ben-efits of a professionally presented theme are evi-dent and are supported in the HyperMuseum bythe TCS/TRS.

With the Theme Generator System (TGS) we areexperimenting with a third approach. From a previ-ous project,Web For Schools (http://wfs.vub.ac.be)(Van Assche 1998) we had observed the need andenthusiasm from the secondary school educationalcommunity for uses of digital media in which therewas a constructive component.This hands-on learn-ing approach has become ingrained in the curriculaof several European countries. Students not onlyobserve, but also actively create and experiment

127©Archives & Museum In ormatics, 2001 129

Stuer etal.,The Hyper Museum Theme Generator System

with the educational materials on offer. The maingoal of the TGS is to allow the discovery and therealization of a student's personal theme, based onthe materials offered through the Hyper Museum,that can be used for a transient presentation to hispeers in the classroom, and that can optionally befurther elaborated into a more permanent projectfor sharing with others.

To support this approach, there must be access tothe Hyper Museum resources that allows for thediscovery of a theme. This is clearly different bothfrom the strict catalogue approach and from thepre-packaged thematic approach.

The implementation of theTheme Generator is alsointended as an early example of a practical ontology-based, or rather, in this case, ontology-assisted soft-ware tool.Tools of this kind assist the user in thecreative process by suggesting (or limiting) choicesduring the activities of query, search, design or com-position of desired results by conceptually (seman-tically) linking linguistic elements of these activities(query text, figure captions, documentation, ...) withother documents and elements through "common"thesauri, lexicons, dictionaries, ... that cover thedomain under consideration. Such linguistic re-sources may be quite general (i.e. domain-unspe-cific), as in the case of WordNet (Miller 1990), oralternatively quite specific to a particular domain.The latter kind of resource, now often referred toas a (domain) ontology, at present is much harder tocome by.As an example of such an as yet hypotheti-cal domain ontology, consider a (partial or compre-hensive) listing of all the technical term linkages thatmay plausibly occur in the context of "restorationof medieval paintings". In fact, one of the primarygoals of the DOGMA project (Meersman I 999b)

atVUB STARLab (http://starlab.vub.ac.be) is to de-fine a formalism, method and representation for suchdomain ontologies so they may be stored and inte-grated on an ontology-server and then retrieved, con-suited and "plugged" into applications.

The main purpose - and advantage - of an ontol-ogy-based approach is to make the implementationof an application such as the TGS as independent ofthe data semantics (meaning) as is currently feasible,somewhat abstractly comparable in purpose to theway databases were invented to make applicationsmaximally independent of data structure.While we

certainly cannot claim that the TGS implementa-tion presents a prototypical solution for this, wetrust that it adequately illustrates the underlyingprinciples - as well as offering an already useful toolthat is extensible according to these principles, asdiscussed below in more detail.

Assisting systems with ontologies is not limited tothe application under consideration. E.g. In theDOGMA research context also, library search sys-tems as well as database (reverse-) engineering toolswith ontological support are studied.

Supporting Personal Theme Discov-ery

While the database/catalogue approach to access-ing museum record data is highly efficient in a searchtype of application, it is far from optimal in assistingin the discovery of the relations between the differ-ent elements. The thematic approach is strong inshowing a possible linkage, but forces the user intoconsidering a possibly very interesting set of rela-tions in a pre-packaged format. For thematic dis-covery, we wanted an application, theTheme Hunter,which would support the user in fluidly discoveringthe patterns of possible, but as yet unrealized,themes.

We set out to construct a Web of relationshipsamong the different objects in the HyperMuseum.In such a structure, the user could explore the spaceof relations by browsing this network from pointto point, until a theme would emerge as an idea orpattern.At the same time, the user should be ableto select the objects encountered for use in therealization of the theme.

For finding the possible relationships between thedifferent objects in the HyperMuseum, we startedfrom the textual metadata that is provided with eachof the museum media objects (e.g. physical descrip-tions, materials, caption).We concentrated on thetextual fields in the records. Normalization of thewords by using a stemming algorithm (Porter 1980)gives us a set of relations to words for each media-object. Bringing all the objects with their respectiveword links together in a graph creates a browsablestructure (Fig.3), where one can move from objectto object through the shared word links.

128


/le

sms AC - 1047 AC

Snullbox

Unused Articles

If Unused Article GroUnused Indices

Unused Maps

Unused Tinelines

SnuffboxReaping marine.. GedsPlough / modelgustrafionIlushation

Advert

DrawingTrade token/7proofPhotographPhotographScytheScythe

,Locaficrt

iscstland Dundee

ca. 1910AC

rrt: copper

copper's_nark

copper-base_alloy

copper-bottomed

copper_beech

copper_color

copper_colored

copper_glance

copper_mine

copper_oxide

copper_pyrites

copper_rockfish

copper_sulfate

copper_sulphate

copperbottom

copperhead

copperplate

coppersmith

peacock

malachite

\Graph Wow I List Vie

EkeWss. FINfew allgraph

start rl

IIIMimic Microsof_lt MC-42001 : HM T 4 Eoplorinp -1-11nrdibtlypetrouseu.., Hypermuaeue 91?

Fig. 3: Screenshot of the Theme Generator System User Interface showing the HMPPG (background)with the Theme Hunter (foreground)

While the browsing interface might be moreadapted to the task of theme discovery, with onlythe word relations from the records we have donelittle but facilitate iterative searching of the recorddatabase. In order to get more depth in the pos-sible theme exploration,we add in the system back-ground ontological resources.These can provide forsemantic links between the words that were linkedto the records. Since this operation will providethe context for exploring thematic space, the choiceof resources should be carefully matched with theintended audience and application. For our initialprototype we used an adaptation of the WordNetlexical database (Miller 1990).

The WordNet lexicon is centered around synsets,lists of synonyms that are intended to represent ameaning. Each synset is related to other synsetsthrough semantic links. The types of links includeamong others hyponym,hypernym,meronym,holonymand antonym relations. Since a word can have morethan one meaning, it can be part of more than onesynset.

By extending our graph to include the wordrelations obtained from WordNet, we get asignificantly richer linkage structure in our system.We no longer need to navigate through coincidingwords, but can resolve synonyms and go intogeneralizations and specializations, opposites andhierarchies.This brings a qualitative change in thesupport of theme discovery. In the next section,we will delve deeper into the consequences ofthis type of facility.

A third source of relations between the objectsand the graph can be mined from realized themes.Since the HyperMuseum Theme Repositorystores themes as XML documents, withembedded links to the media records, it isrelatively straightforward to get the context ofthe reference and construct links in our graphbetween the context words and the objects.Thisexpands the number of meaningful links (withinthis context) that exist between an object andthe rest of the graph.

12 9©Archives & Museum Informatics, 2001

BESTCOM/AVAILABLE131

Stuer et.al.,The Hyper Museum Theme Generator System

Supporting Personal Theme Genera-tion

In general, one performs theme discovery with thepurpose of communicating it.While it is certainlyfeasible to use a modern content creation packageto construct multimedia presentations, there aretasks specific in our context, for example generat-ing timelines or maps,that would clearly benefit fromspecialized automated support.This need has alsobeen identified and reported in similar projects suchas (Buchanan 1999). The Theme Generation Sys-tem contains a client side application, theHyper Museum Personal Project Generator(HMPPG), whose primary goal will be the facilita-tion of this aspect.

Since the theme itself should be kept independentof the details of the medium in which is it rendered,we tried to keep a separation between the concep-tual creation of the theme (and in particular theontology it uses), and its instantiation in a particu-lar environment (e.g. as aWeb site or Word docu-ment).While the current prototype only containsgenerators for Web site creation (as sets of HTMLfiles), the implementation is such that the genera-tion of different output types such as e.g.3D virtualexhibitions as described in (Alonzo 2000) can beachieved through the inclusion of a new set of gen-erators.The user constructs a theme based on con-ceptual objects such as groupings or indices that inone medium may be generated as Web pages andnavigational menus, in another as virtual rooms andfloor plans, and in yet another as chapters and atable of contents.

Realizing a theme based on museum objects is typi-cally the creation of a new (thematic) dimension onthe set of these objects.This new dimension doesnot replace the existing time or geographical di-mension, but adds to these more objective refer-ences. Often the theme woven from a string of ob-jects described in the thematic light will benefit frombeing also placeable and navigable along a timeline,or on a geographical map. Indeed, these factual dataare often present in the HyperMuseum records, butthe inclusion of these perspectives with generaltools might be a laborious task.The HMPPG proto-type generator automates the generation oftimelines as HTML clickable maps, and a tool that

130

automates placement of objects on geographicalmaps. These geographical maps can be createdthrough a separately developed tool, GeoMap Edi-tor, which is included in the HyperMuseum TGStoolset. Both these systems are based by default onthe extraction of metadata from the HyperMuseumrecords. However, all project data can also be en-tered and edited manually for each instance, shouldthis be necessary.

All elements in the HMPPG have themselves timeand location meta-data. This makes it possible tocascade these elements into each other in essen-tially limitless combinations.One can easily constructtimelines of maps and vice versa. Changing the or-der in the navigational hierarchy is a simple matterof dragging and dropping items from one place toanother.The system is constantly checking for con-straints (e.g. in order for an element to be placedon a timeline, it needs to have a date or periodassociated), and will prompt for missing data asneeded.This feature makes adding and changing dif-ferent perspectives relatively straightforward.

At all times the user can observe what the elementbeing constructed will look like with the currentlyselected generator.This provides immediate visualfeedback.At any time more data can be fetched fromthe HyperMuseum through the Theme Hunter orthe RDS, or new media files canbe added from thelocal file system. Most popular media types for im-ages, audio and video are supported. The applica-tion can be used by itself to construct presenta-tions.The output produced by the current proto-type Web generator is standard HTML and can befurther processed with other packages.

Architecture of the Theme Genera-tion System

The Theme Generator System (Fig. 4) can be bro-ken down into two major parts. One part is re-sponsible for the theme realization and is referredto as the Hypermuseum Personal Project Genera-tor. The second part is referred to as the ThemeHunter,and is responsible for theme discovery sup-port.While both are very different in their imple-mentation details, they integrate seamlessly fromthe user's point of view, with no noticeable transi-tions.



HSC Server TGS Server

Metadata

Themes

Record Processor

XML

Record

Object

Word Extraction

Geo Map Editor

Theme processor

Mufti Media Objects

0

Wordnet 1.6

ransformer

Other ...

Transformer

HM OntologyServer

Graph constructor

Graph DB

TGS Client

Client, HCS or Third Party

Web Server

WebGenerator

Generator i--

RequestProcessor

cr"

FIMPPG Main Theme Hunter Client

Fig. 4: Overview of the Theme Generator System Architecture

l 3@ Archives & Museum Informatics, 2001 I 3 3


TheTheme Hunter is a classical client-server appli-cation.The centerpiece of this system is a relationaldatabase system which houses the object-ontologygraph described above. Client applications are ei-ther delivering new graph elements, or querying thegraph. For practical purposes, the delivery of theactual multimedia files is handled through a stan-dardWeb server.Since the only point of interface isthe graph in the database and the media files on thewebserver, any application that can deliver graphextensions can be added to the system. This fea-ture, together with the localization of format-de-pendent code in the HMPPG, will make it possiblefor the system to easily migrate to or be extendedwith new data formats.

As seen in the database diagram(Fig 5), the graph isrepresented as a simple set of binary relations be-

tween entities.The entities are typed to distinguishe.g. between words and object references, and as-sociated with a source.This source is itself typed,and in our current system there are three types:the Hypermuseum Records, the HypermuseumThemes and WordNet. Each source type has a setof relations that can be further classified into rela-tion types for easy property attribution. From theTheme Hunter client, the advanced user can con-figure the system to work only on certain sets ofsources,or take into account only certain relations.

On the server side there are three main data deliv-ery modules, each responsible for extracting infor-mation from the different data sources: theHypermuseum Records, the HypermuseumThemes,and the WordNet 1.6 data files. For the WordNetintaker, we start from the DOGMA representation

id

name

sourcetypegloss

id

sourceentitylrelationentity2

id

name

sourcetypetype

Fig 5.:Theme Generator System Graph Database Diagram

1 3 2I 34 ©Archives & Museum informatics, 2001


ofWordNet 1.6, which is in content identical to theWord Net data files, but allows for easier manipula-tion. From this representation we can build the re-quired graph representation through direct SQLmanipulations.We build a semantic net graph cen-tered around the words, as opposed to the synsets.We ambiguate the meaning structure by flatteningthe word-meaning relations into a single node, andtransform the synset objects into explicit synonymrelationships between the member words.

For the Themes and the Records, the first step isextracting the field data that is to be considered forword mining.This operation is made easier by thefact that both use an XML representation, makingthe parsing and validation of the files more conve-nient. From this step the relation between the ob-ject and the possible word set is determined.Textdata is passed through the word extraction pro-cess, in the current prototype a chunking (cuttingthe sentence into pieces) and stemming operation.There are many opportunities here for the applica-tion of advanced Natural Language Processing tech-niques that could improve both the accuracy andthe depth of the text mining operation. The laststage in our current mining pipeline matches thewords against WordNet 1.6 to prune the false re-sults out of the previous pipeline stages.The result-ing words are then inserted into the graph, togetherwith their respective relations to the objects. Itemsrecovered form Theme files are only included if therecord they refer to was previously processed bythe system.

The part of the Theme Hunter that is visible to theend user is conceived as a 3-tier application on thegraph database.The front end is a GUI based appli-cation (Fig. 3) where the user can search for termsto start exploring in the graph, and graphicallybrowse the graph by simple mouse clicks. Wheninteresting objects are discovered, they can bedownloaded for processing with the HMPPG by asimple right click on the image.The Theme Hunterand the HMPPG work completely asynchronously,so realization and discovery can be fully intermixedin the creative process.

Behind the scenes the first tier constructs a seriesof request objects that are passed on to the middletier.The middle tier interprets these requests and

builds database queries out of these. With thesequeries it extracts the information out of the graphdatabase, and the result is transformed into a replyobject and passed back to the first tier, where theresult is unpacked and displayed.The protocol gov-erning this interaction is stateless, ensuring that theprocess is robust to client or server failure and eco-nomically scalable on the server side. Furthermore,both the request and reply objects that govern thefirst-middle tier interaction are serializable, so thatthe actual physical location of the middle tier is eas-ily changeable. In our current setup we placed themiddle tier on the client side, but it can with rela-tively minor efforts be moved to the server sideshould the need arise.

Our Theme Hunter implementation uses theMicrosoft SQL Server v7.0, and has also been testedwith the free Microsoft Data Engine 1.0 for cheaperdeployment scenarios. Since the database opera-tions are handled through SQL over a standardODBC interface, there should be no problem tosubstitute for these choices.The database is comple-mented by a standard Web server, in our systemInternet Information Server 4.0, for the delivery ofthe actual multimedia files to the clients. All codefor this part of the system was written in Java 2except for some of the DB conversion code thatwas written directly in SQL.

Like theTheme Hunter,the Hypermuseum PersonalProject Generator also tries to isolate the depen-dencies on data formats. The HMPPG is a clientside program that is constructed as an output-for-mat-agnostic generation framework. Data input intothe system can be from Hypermuseum record filesthat can originate either from interaction with theRDS or from the Theme Hunter, or from raw mul-timedia data supplied by the user from othersources. Due to its intense local data interaction,this application was developed mainly using theBorland C++ Builder system, since the currentinstantiations of the Java platform did not deliverthe required performance at this point. (TheThemeHunter, being a network-data bound application, isnot critically hindered by this).The dependency onthe Hypermuseum record format is shielded by lo-cating the parsing and extracting code into a sepa-rate program, the FileAcceptor. The responsibilityof the FileAcceptor is to extract from a set of Hyper-

I 133@Archives & Museum Informatics, 2001 I 35


r museum records the relevant metadata and themedia files, and produce these into an internal for-mat on which the rest of the HMPPG will rely. (Sincethere is no user interface involved here, we coulddevelop this program in Java 2).

To achieve data format independence on the gen-eration side, two requirements had to be fulfilled.

First, all the operations on the data, apart from thegenerated representations, had to be implementedas operations on the conceptual structures. Sec-ond, the generators could not directly be linked intothe rest of the program. Instead, a COM (Rogerson1997) interface was defined that specifies whichservices a generator has to provide to the HMPPG.Since the HMPPG acts towards the generators as aCOM client, each generator object implementingthis interface can be added to the one deliveredwith the current prototype.

The current HMPPG generators can work withmost of the standard formats for audio, video andimagery. Every format supported by the WindowsMedia Player and the Internet Explorer controls maybe used in the prototype.The Geo Map file formatfor geographical data is supported through theGeo Map Editor tool that also acts as a COM servertowards the HMPPG for handling Geo Map opera-tions.

We opted for this defensive position with regardsto changing data and output format requirementsbecause of the current absence of stable and domi-nant standards for this field. Since this situation iscurrently receiving quite some interest, we hope tobe able to incorporate the emerging standards intothe system in the future.

Discussion and Results of Initial Ex-perimentation

As stated above, the choice of background onto-logical resources is critical to the intended use ofthe system.The choice of theWordNet lexicon wasactually one of convenience and availability, but firstreactions to the prototype indicate that the choicemay have unexpected benefits.

SinceWordNet has attempted to model the lexicalknowledge of a native speaker of English, its vo-

cabulary is quite broad but non-specific.As a result,the user is generally not confronted with jargonwhich might have been more accurate or meaning-ful to the professional user, but could have alien-ated or derailed our untrained user in the processof making free associations. Secondly, while ourpresent crude word harvesters can extract thewords, they cannot disambiguate them to their in-tended meaning (e.g. it can not distinguish betweenthe use of bank as the place that takes your moneyas opposed to the bank at the side of the river).Rather than an expected disadvantage, this "feature"in the context of our application surprisingly seemsto bepositive, since these mismatches seem to beone of the best sources of theme "triggers". Eventhe fact that our current main view in the browser(there is a graphical main view and a textual view)does not show the relation types has triggered sur-prise, reflection and discovery in the first test whenusers unexpectedly came across links they at firstthought were not supposed to be there (e.g. thelink between "rich" and "tasteful"). For some in-stances at least, less could indeed be more whenselecting background ontologies.

The previous point is not intended to be a dog-matic principle. We could point to the lack ofagentive links (e.g. there is no relation between"baker" and "bread") as an obvious shortcoming ofour current system.

Once the background ontology is in place and thegraph generated, the choice of how much of it todeploy can only be discovered through experimen-tation. In our current Theme Hunter, the advanceduser can include or exclude both types of relationsand sources. Even so, depending on the scenariowe can foresee a need to further reduce the amountof presented information.We are currently investi-gating the introduction of additional retrieval con-straints such as a maximal distance rule (words areonly included in the graph within at least few"hops"of a media item) to bring more focus on findingthemes that will be "realizable" with theHyperMuseum data at hand.These criteria are de-pendent on the specifics of the usage scenario andwill have to be evaluated on a case-by-case basis.

As far as the records themselves are concerned,the HyperMuseum approach of not requiring rigor-ous data standardization undoubtedly has the ad-

1 31I 36 ©Archives & Museum Informatics, 2001


vantage of lowering the threshold for museum par-ticipation.The flip side of this coin is, of course, thatsince there are not yet standard representationalformats for things like dates or geographic loca-tions, the correct interpretation of these becomesdifficult (time) to very difficult (geographic).A com-mon reference to standardized vocabularies suchas the Getty vocabularies (Lanzi 1998) could aug-ment the quality of information extraction by thedifferent tools.

Conclusion

The HyperMuseum Theme Generator System aimsto be a contribution to the exploration of new usesof museum online digital information in,among otherenvironments, the secondary education curricula inEurope. By allowing a novel way of exploring col-lections, supporting both the discovery of new the-matic dimensions by the non-professional as wellas the rapid creation of the presentation of thetheme, we hope to address concrete needs of theintended audience. By localizing the dependencieson specific formats, both on the data delivery sideand the document generation side, we hope to haveprovided a technical platform that can be flexiblyadapted for usage in many environments.

Acknowledgements

This work is supported by the European Commission'sEU Telematics Application Project HYPERMUSEUM(Project nr. 3088).We would also like to acknowledgecontributions by our museum partners in this project,the Oxfordshire County Council museums (Oxford,UK), the Musée Calvet (Avignon, FR) and the GalleriaDegli Uffizi (Firenze,l) . In particular the datasetdeveloped by the National Museum of Scotland(Edinburgh, UK) provided an excellent ontology testingtool.

References

Alonzo, F., Garzotto, F., and Valenti, S. (2000) 3-DTemporal Navigation on the Web: How to Explorea Virtual City along Multiple Historical Perspec-tives., In D. Bearman & J.Trant (Eds.) Museums andthe Web 2000 Proceedings.,Archives & MuseumInformatics, 2000.

ANSI/NISO (1995) Information Retrieval:ApplicationService Definition and Protocol Specification.Available at http://lcweb.loc.gov/z3950/agency/document.html

Buchanan, S. (1999) Unlocking the Treasure ChestUsing SCRAN Tools. In D.Bearman & J.Trant (Eds.)Museums and the web 1999 Proceedings.,Archives& Museum Informatics, 1999.

Lanzi, E. (1998) Introduction to Vocabularies, EnhancingAccess to Cultural Heritage Information. Los Angeles:Getty Trust Publications.

Meersman, R. (1999a) Semantic Ontology Tools inInformation Systems Design. In Z. Ras & M.Zemankova (Eds.) Proceedings of the ISMIS'99Conference. Heidelberg: Springer Verlag.

Meersman, R. ( I 999b).The use of lexicons and othercomputer-linguistic tools in semantics, design andcooperation of database systems. ln:YanchunZhang, M. Rusinkiewicz,Y. Kambayashi (eds.),Proceedings of the Conference on CooperativeDatabase Systems (CODAS'99). Heidelberg: SpringerVerlag.

Michard A., Christoph ides V., Scholl M., Stapleton M.,Sutcliffe D., and Vercoustre A-M. (1998) TheAquarelle Resource Discovery System., Joumal ofComputer Networks and ISDN Systems,30(13):1185-1200.

Miller, G.A., Beckwith R., Felbaum C., Gross D., andMiller K., (1990). Introduction to WordNet :AnOn-line Lexical Database., International Journal ofLexicography, 3, (4), 235 - 244.

Porter, M.(1980) An algorithm for suffix stripping.Program, 14, (3), 130-137.

Rogerson, D. (1997) Inside COM. Redmond: MicrosoftPress.

Van Assche, F (1998) (Ed.). Using the WWW in SecondarySchools. Leuven:Acco.

135© Archives & Museum Informatics, 2001 1 37

Providing Personal Assistance in theSAGRES Virtual Museum

Ana Carolina Bertoletti, Museu de Ciências e Tecnologia - PUCRS,Marcia Cristina Moraes, PPGCC-UFRGS,Antônio Carlos da Rocha

Costa, Escola de lnformitica - UCPEL, Brazil

Abstract

The SAGRES system is an educational environment built on the Web that facilitates the organization of visits tomuseums, presenting museum information bases in a way adapted to the user's characteristics (capacities and preferences).The system determines the group of links appropriate to the user(s) and shows them in a resultant HTML page. Inaddition, SAGRES enables cooperative learning by supporting interaction among users and also among members ofgroups of users.

Users in SAGRES are aided by personal assistants that are software agents, whose purpose is to monitor the visitors'actions, helping them during the navigation. Considering studies in human-computer interaction (Ball et al, 1997)(Koda, 1996) and aiming to provide a friendlier interface for the SAGRES system, the agents have a graphical representationas animate characters. These characters improvise a group of behaviours similar to human behaviours (happiness.satisfaction and vibration), making the interaction more attractive.

In order to evaluate the user's degree of satisfaction with the agents, we developed a questionnaire and obtained somefavorable results. This system is developed in the Museu de Ciências e Tecnologia (MCT) at Pontificia UniversidadeCat6lica do Rio Grande do Sul (PUCRS), Porto Alegre. Brazil.

I . Introduction

Most virtual museums are sites that offer on-linelearning resources, inviting the visitor to investi-gate and explore.The SAGRES system is a virtualmuseum that seeks cooperation between museumand schools, in order to create a new educationalenvironment that provides continuous education,allowing access to information of the museum bythe community at large. Due to the possibility ofusing the system as a distance-teaching tool, and tothe diversity of the school population to be reached,SAGRES was conceived as a flexible and adaptivesystem, able to pay attention to the different needsand situations of its various visitors. Because pre-sentation of information on the Web is achievedthrough hypermedia documents, success of virtualmuseums can be limited by a lack of initial trainingand by users getting 'lost' during navigation, due tothe large number of links available.

In order to overcome these limitations, softwareagents were used to incorporate personal assis-tance to the SAGRES' users.These kinds of agentshave been used for the development of a large num-ber of applications.They can provide assistance tothe user during some systems operations and canexecute tasks on the user's behalf (Genesereth and

Ketchpel, I 994).As personal assistants they can beresponsible for helping in the navigation and opera-tion of the system, overcoming the need for initialtraining. Besides, software agents are quite useful inanalyzing and monitoring the users' actions. Con-sidering this, we believe that agents can help in thesystem operation both at operational and interfacelevel.

Many systems developed for the Internet have aconcern about how to accommodate user needsand preferences and how to provide interfaces thatare more comfortable and easy to manage. Someresearchers are considering software agents as agood option, because they are adaptable, persistent,semi-autonomous (Maes et all., 1999) and can berepresented through animate characters.There arestudies that indicate that interaction with comput-ers unavoidably evokes human social answers. Peopletreat computers as humans even if the interface isnot explicitly anthropomorphic. In this case, thesocial metaphor represented through the presenceof animate characters (similar to real life people)has reduced anxiety associated with the use of com-puters.



BEST COPY AVAILABLE

Bertoletti et.al., Providing Personal Assistance in the SAGRES Virtual Museum

In order to evaluate the degree of user satisfactionrelated to the agents,we conducted a study of agentevaluation methods and, based on this research,developed a questionnaire. We applied this ques-tionnaire to partners of the Computer Club in theMuseu de Ciências e Tecnologia (MCT) and we ob-tained some satisfactory results.

2. SAGRES A Virtual Museum

The SAGRES system (Bertoletti and Costa, 1999)is a virtual museum that seeks to form a partner-ship between museums and schools, and to build anew educational environment going beyond suchpartnership by providing information available in themuseum to the whole community, through theWeb.The system determines the information appropri-ate to the visitor(s) and shows it in a resultant HTMLpage.

The information available in the SAGRES is basedon the physical structure of MCT.The subjects forconsultation are arranged in the system in agree-ment with their actual location in the MCT exhibi-tions area, allowing to the visitor to find the ex-periments in a real visit.

Through the on-line exhibition visitors access theinformation of the physical museum.The use of suchon-line information can happen in several ways(Lewis, 1997): in a set of activities coordinated byteachers in which students participate; in "halls ofactivities", i.e., series of activities with a centraltheme open to any visitor, and by individuals as asource for individual studies. In addition, SAGRESprovides cooperative learning by supporting inter-action among visitors and also among members ofgroups of visitors.

The system supports different groups of users:

A Visitor accomplishes the interaction with thesystem individually. He is included in the systemby a process of self-registering, where he is re-sponsible for defining his profile, informing thesystem of his characteristics, capacities and pref-erences, in order to facilitate the construction ofhis model.

ATeacher (generally identified as a school teacherwhose student groups visit the museum) is re-

1 3

sponsible for the definition of the profile of thegroup. By a profile we understand the set of char-acteristics of the group; that is, the students' back-grounds and preferences, any particular subjectbeing focused in the visit, and the activities.Theteacher is also responsible for registering the stu-dents, as well as accompanying them and evaluat-ing their performance during the visit, throughreports delivered by the system.

A Student is allowed to interchange ideas withcolleagues in his group and to work on the activi-ties and subjects determined by the teacher.

One advantage of SAGRES is that it allows visitorsto plan their visits to the MCT in advance. Thus,when arriving at the museum, the visitor alreadyknows where to find specific experiments. Later, intheir homes, the visitors can get other informationrelated to the experiments visited.

For teachers, SAGRES is a powerful tool to sup-port teaching, because it works as a repository ofinformation from which they can elaborate a libraryof activities (for example, questionnaires for tests)and share them with other teachers.

3. Integrating Personal Assistance inthe SAGRES

Since information is presented on theWeb throughhypermedia documents, virtual museums are lim-ited by the lack of initial training and by users get-ting lost during navigation, due to the large numberof available links (Yamada et all., I 995). Therefore,software agents were used to incorporate personalassistance in the SAGRES system. As personal as-sistants they have the responsibility to help the visi-tors during interaction with the system.These agentsare represented as animated characters.Accordingto Hayes-Roth (1998), the introduction of charac-ters on the Internet introduces the sense thatpeople are interacting with "real characters" in vir-tual worlds.This sense is intensified because char-acters simulate verbal and physical behaviours similarto humans,such as happiness, satisfaction and greet-ing.As people are social beings and interact instinc-tively with each other, the user's interaction withthe interface becomes friendlier.



To support independence between the agents,SAGRES adopted a three-layer development model(Kirtland, I 999): presentation, business and data.Thelayers of data and business meet in the server. Datalayer stores the SAGRES's databases and informa-tion related to the agents' execution scripts andrepertoire of behaviours.The business layer imple-ments the agents' functionalities.The presentationlayer is located in the client and contains theSAGRES's HTML pages and agents' pages. Thismodel is visualized in figure I.

Server DATA LAYER

Agents

.--------- _>_---

SAGRESDatabes

Dataid. state ...

ScriptsDatabases--

BehaviorsDatabase

Server 1 BUSINESS LAYER

SAGRESSystem ,

Modules

Agents

Control Modul e

Client

y

PRESENTATION LAYER

SAGRESPages

......----- '..---"- Browser

t0 1NE crosoft

Agent

User

Figure I:Agents and SAGRES System

Figure 1 shows that agents are structured into twomain modules: internal representation (presentedin the data and business layers) and interface (pre-sented in the presentation layer).The internal rep-resentation module is divided into data and con-trol. Data contains information about the agent suchas agent identification, current agent state (idle, ex-ecution), script (name of the file that contains theexecution's script of the agent), list of neighbours(name of the agents that it will be able to commu-nicate to) and who is communicating (name of the

agent that is communicating at the moment).Thisinformation is stored in a database.All informationin data is updated, according to the process in thecontrol module. The control module contains thepersonal assistants' functions.Assistants have com-mon knowledge about the world that is implicitlyincorporated in the control module.The interfacecontains graphical representation of the agents thatinteract with visitors through the execution scriptincorporated in the HTML pages (Moraes et al.,I 999a). An example of this script is illustrated infigure 2.

1. To present to the visitor (to greet to wave)2. To invite the visitor to make a visit to the

museum (to invite, to dame)3. To wait reply of the visitor (to wait)

3.1 If visitor selects a consultationLooks to presentation agent for content(researcher)D emonstr ates satisfaction (satisfaction,varation)

3.2 Else demonstrates reproof (to refuse, todisapprove)Stimuldes visitor to make a visit (toshMulate, to move)

3.3 If visitor reque ts mural's monitoringMonitors communication intral(inveshgating artist)

3.4 If visitor requets monitoring of documenteditionMonitors edition of document (research,

artist)

Figure 2: Example of an execution script

This example shows that agent's actions are relatedto behaviours (physical and verbal).Each one of thesebehaviours has one to five possibilities of execu-tion. In this way, the agents can improvise behavioursand different possibilities to execute the selectedbehaviour. This characteristic makes agents moreflexible, life-like and believable, increasing the users'satisfaction related to the system operation as wecan see in section 5.

4. Visualizing Personal Assistance inthe SAGRES

In this section we are going to describe the SAGRESsystem, as it operates with personal assistants, byillustrating an example of the visitor' interaction.Figure 3 shows the main page of the system.

1 3 8© Archives & Museum Informatics, 2001 141

Bertoletti et.al., Providing Personal Assistance in the SAGRES Virtual Museum

46.

0C1. equctoo ono. d.

Figure 3: Main Page

After choosing the personal assistant option, thevisitor provides his name and password (figure 4).

Pam ,n10 no SA4,1E7 L'en

,Invente uma senha:

Figure 4: Logon Page

Based on the name and password information thesystem identifies the user's kind:visitor, student and/or teacher (figure 5).

1

..1...aro 2 Apt

Prefonor

Figure 5: Kind of User Page

After the user chooses the visitor module, the sys-tem loads his visitor's models (figure 6).

139

Am* NW 11. Coolmolow. lino*r., I 0 xs. Pt.hiais

r A mglop* I' .... thAvpv,

r Ser., " Penimlo

r n1.1: Ivar I , Fvwpos

r Flod .55 ,..r .P.1.14* I PAN Paewoo

r ...45 Pootmvit

r 4..... Gm ,Pwmpo,

Figure 6: Consultation Options Page

After the user selects a subject, the system builds apresentation page where the user can see the docu-ments about the selected subject and the link forthe communication mural and document edition(figure 7).

C ,JU C.CtlAOn. N.7 IN0

liecemonlmeolem CAN

Figure 7: Information Presentation Page

In some pages there is a button related to the per-sonal assistant, as we can see in figures 6 and 7.Thisbutton activates the assistant, so when the userneeds help, he only has to click on that button.Whenthe user wants to stop the explanation, he clicks onthe button again.The user can activate the assistantwhen the button is green, and he can stop the as-sistant when the button is red.The explanation pro-vided is in accordance with the pages' context, as-sistant actions and behaviours.When the button isnot present, the assistant acts autonomously. Thismeans that the assistant observes the user's actionsand motivates and helps the user when necessary.

BEST COPY AVAILABLE



5. Evaluation of Personal Assistancein the SAGRES System

In order to evaluate the user's satisfaction with thedegree of personal assistance in the SAGRES sys-tem, we researched evaluation approaches for in-telligent agents. Hayes-Roth and Doyle (1998) pointout that works in intelligent agents inherit the evalu-ation approaches of Artificial Intelligence and otherfields of Computer Science such as Human-Com-puter Interfaces. These approaches define the de-sirable qualities of systems. However, as personalassistants are intended to be more similar to people,traditional approaches should be adapted or modi-fied. Hayes-Roth and Doyle (1998) present an ad-aptation of some important approaches for evalua-tion of animated characters.As personal assistantsare represented through animated characters, theseapproaches can be applied.The approaches are:

Reliable becomes variable:to be more human-likeand believable the animate characters must dis-play normal variability in their choice and mannerof executing behaviours

Predictable becomes idiosyncratic: animate char-acters must display roughly predictable patternsof behaviour, punctuated by interesting surprises

Correct becomes appropriate:rather than behav-ing correctly, animate characters must behave ap-propriately, given the roles, their circumstancesand their imperfect human-like natures

Complete becomes effective: animate charactersmust function effectively, albeit sometimes incom-pletely, in the roles they play

Efficient becomes interesting: animate charactersmust forego efficiency in their paths to goals infavour of paths that are intrinsically interesting

Optimal becomes distinctively individual: it is im-portant that animate characters be distinctive in-dividuals whose behaviour surprises, delights orprovokes us in ways that are unique

Considering the ideas presented by Hayes-Roth andDoyle (1998), we developed a questionnaire toevaluate characteristics related to the agents.Thequestions are illustrated in figure 8.

Questions Bad Regular Good VeryGood

Did the personal assistant present differentbehaviors during the execution of particulartasks/

Table Questions Form

Each question has four alternatives as answers:bad,regular, good and very good. Weights of one tofour, for concepts, were associated with answers,as illustrated in tables I and 2.

Concept WeightBad I

Regular 2

Good 3

Very Good 4

Table 2: Concepts Weight

These questions follow the same format presentedin table I, except the last one, where the learnershould choose between "with" or "without".

I-Did the personal assistant present different behaviors during the execution ofparticular tasks?2-Did the personal assistant present appropri atedbehavs or the each task?3-Did the personal assistant was useful to help you in the system operation?4-Did the personal assi stant help you in a personali zed way and did he encourageyou to do your tasks?5-Did the personal assistant disturb you during the presentation of informan on?6-Did the personal assistant help you to concentrate on relevant information?7-Did the information presented by the personal assistant are easy to understand?8- Can you begin and intenupt the action of the personal assistant?9- In the future, if you can choose the presentation with or without the personalassistant, which one you choose?

Figure 8:Agents Questionnaire

The average obtained in the questionnaire was 3.68.We verified that users favour the assistant'sbehaviours during the execution of tasks, obtainingan average 3.875. It's important to say that the otheraspects obtained an average superior to 3.6.So, theassistant reached successfully the approaches pro-posed by Hayes-Roth and Doyle, displaying normalvariability in their choice and manner of executingbehaviours, and being attractive and interesting tousers. Besides, the guide treats visitors in a person-alized way, aiding and motivating them to accom-plish tasks.

We can conclude that visitors are satisfied with theuse of assistants.This satisfaction is explicit, becauseall the visitors answered "with" to the question "Inthe future, if you can choose the presentation with orwithout the personal assistant, which one will youchoose?" In this way we verified that the personalassistant was an important aid to visitors during

14 0BEST COPY AVAILABLE


Bertoletti eta, Providing Personal Assistance in the SAGRES Virtual Museum

the operation of the system, executing its main func-tionality and serving as an incentive for future ac-cess to the system.

6. Conclusions

The SAGRES system makes the interactive museumof PUCRS remotely available to the public visitor,as well as contributing to the improvement of teach-ing and promoting the exchange of informationamong geographically distant visitors.

Personal assistants are used in SAGRES to assistvisitors during a visit They do this by helping visi-tors in the exploration of information and systemsoperation, through a friendly interface where visi-tors interact with animated characters whosebehaviours are similar to human behaviours.A ques-tionnaire administered to users of the virtual guidedemonstrates that users prefer to operate the sys-tem with the virtual guide, because it facilitates navi-gation, offers personalized help and demonstratesinteresting ways to interact with users, achieving itsinitial aims.

References

Ball, G; et al. (1997) Lifelike Computer Characters:ThePersona Project at Microsoft. In: Software Agents.Menlo Park, California:AAA! Press. 1997.

Bertoletti, A. C. & Costa,A. C. R. (1999) SAGRES AVirtual Museum. In:Museums and The Web 1999.USA: New Orleans, Louisiana. 1999.

Bertoletti, A. C. et al. (1999) The SAG RES system:incorporating software agents to assist visits in avirtual museum. In: Fourteenth InternationalSymposium on Computer and Information Sciences -ISCIS'99, 1999. Kusadasi,Turkey.

Genesereth, M.R. & Ketchpel, S.R (1994) SoftwareAgents. Communications of the ACM. NewYork: v.37, n. 7, p. 48-53, July. 1994.

Hayes-Roth, B. (1998) Interacting with AnimatedCharacters: Puppets, Bartenders, and AutoSalespersons. Technical Report (KSL-98-07) -Stanford University.. 1998.

Hayes-Roth, B.; Doyle, R (1998) Animate Charaters. In:Autonomous Agent and Multi-Agent Systems.Netherlands: Kluwer Academic Publishers. 1, 195-230, 1998.

Kirtland, M. (1999) Projetando soluciies baseadas emcornponentes. Rio de Janeiro: Campus, 1999. 413p.

Koda,T. (1996) Agents with Faces:A Study on the Effects ofPersonification of Software Agents. Massachusetts:Program in Media Arts and Sciences, MIT. MasterThesis. http://tomoko.www.media.mit.edu/people-tomoko/thesis.htnil. 1996.

Lewis, L. C. (1997) At home in the Heartland Online:forming a museum/school resource via the WorldWide Web. In: D. Bearman & J.Trant (Eds.)Museums and the Web, Selected papers fromMuseums and the Web97. Pittsburgh:Archives &Museum Informatics.

Maes, R et al. (1999) Agents That Buy and Sell Communi-cations of the ACM, New York, v. 42, n. 3, p.81-91,March 1999.

Moraes, M.C. et al. (1999a) Virtual Guides to AssistVisitors in the SAGRES Virtual Museum. In:International Conference of the Chilean ComputerScience Society - SCCC'99. 1999,Talca: Chile. IEEEPress.

Moraes, M. C. et al. (1999b) The SAG RES virtualmuseum with software agents to stimulate thevisiting of museums. In: WebNet'99. 1999, USA:Hawaii.

Yamada, S. et al (1995) Development and Evaluation ofHypermedia for Museum Education:Validation ofMetrics. NY:Acm Transactions on Computer-humanInteraction. 284-307.1995.

141

'


OPALES:An Environment for Sharing Knowledge amongExperts Working on Multimedia Archives

Henri Betaille, Marc Nanard, Jocelyne Nanard, LIRMM, France

Abstract

INA, the 'Institut National de l'Audiovisuel' keeps records of national TV and radio production as French patrimonialarchives. They are mainly accessed by specialists for research purposes, and by TV producers for inserting archivesegments within new productions. INA and several others partners have initiated an R&D project, OPALES, to developa distributed environment which enhances experts private work on multimedia archives and enables collaborativeknowledge work on the Web.The challenge is to advance knowledge by building digital communities of experts whoadd value to the archival dataset by annotating items.The environment supports users working on multimedia archives,preserves their data in private workspaces, and helps them to share expertise. Each end-user accesses informationwithin a private workspace. Any document (annotation as well as archive) is handled as a private copy which canvirtually be annotated, indexed, linked to other information, edited to be inserted into new documents, and so on.Direct anchoring of annotations within audio or video is supported.

To manage information and knowledge sharing, OPALES introduces the notions of an 'authoring point of view' whichidentifies annotation categories and of a 'reading point of view' which specifies which categories of annotations areader wants to see.Any added piece of information always has an author and an 'authoring point of view.' To enableknowledge sharing, any user can 'export' a point of view to make some part of the elaborated knowledge available toothers. Exporting a point of view consists of indexing it into the shared ontology to enable other experts to retrieveit easily and import it into their workspaces. A 'reading point of view' defines how a document is enhanced byannotations when presented. It is a mix of imported points of view. For instance, a researcher on sociology may'import' ('borrow') the knowledge previously elicited and exported by economists, politicians, ethnologists, and so on,to better understand a document or to improve the relevance of queries. The selected annotations and links aredisplayed with the document.To enable computer activity using shared information, the system provides a mechanismfor handling an extensible ontology, including point of view dependant aspects. It provides support for indexing and forsearching in annotated documents.

The paper presents the features of OPALES, describes the mixing of points of view on video archives, and discussessome issues raised by knowledge sharing among experts.

Introduction

This paper presents a national initiative currentlyunderway in France to develop a new service atINA. It aims to support work on video archives.The expected result is a step towards knowledgecreation through digital communities of expertsexchanging their expertise and making their workavailable to others.Their common work improvesknowledge of the archives, thus leading to addedvalue.

This service relies upon three interdependent andintegrated features:

an improved information retrieval system, withmore precise recall for most users

private workspaces in which users workseamlessly with private documents, archives andannotations of their own or of other users.

a computer supported collaborative environmentwhich enables distant users who do not know eachother, but have the same concerns, to cooperateand to share the result of their work.

The project is intended to apply mainly to videoand audio archives.

The paper first introduces the context of the work,presenting the origin of the video archives and theproject.Then the role of the Web in facilitating dis-tant exploration and annotation of video is pre-sented.The focus is placed first on the features avail-able in private workspaces, then on the sharing ofknowledge among users.The notion of point of viewis described from both the author's and the reader'sside.

142

BESTCOPYAVAILABLE

Museums and theWeb 2001:ProceedingsArchives & Museum Informatics, 2001, p. 145

Nanard, etal., OPALES:An Environment for Sharing Knowledge

Working on Multimedia Archives

Many projects deal with video archives analysis(Chang 97, Houghton 99).Very few rely on humansfor such a task. For instance, the Informedia projectat CMU (Hauptmann 95, 011igschlaeger 99) relieson automated approaches such as language recog-nition and image recognition to produce on lineoverviews of audiovisual documents.The OPALESproject has a quite different purpose and strategy:it aims to enhance the relevance and the level ofdetail that already exist in the indexing, with addedmetadata contributed by users working on thevideo.

Video Archive Storage Policy inFrance

The systematic collection of books by National Li-braries is now routine. Most great nations keep arecord of all of their published information in li-braries, and already a part of this is available in digi-tal libraries. By contrast, most movies,TV programs,and radio productions are still collected only byproducers.As a consequence, this costly collectioneffort is often limited only to documents which havegreat value and a high probability of being reused,either frequently or within a short time. Often thecultural or the patrimonial value of a document isnot the main criterion in deciding on its preserva-tion for the long term. Such behavior is explicablein terms of market economy, but it causes much ofthe history of the audiovisual life of the country tobe lost.Therefore, in the future it will not be acces-sible to researchers studying the evolution of oursociety. This is why France enacted a law in 1992making it mandatory that any information producerdeposit all published production into a specializednational institution: the Bibliotheque Nationale deFrance for printed documents, the Centre Nationalde la Cinematographie for films,and the '1NATheque'of the Institut National de l'Audiovisuel (INA) foraudiovisual production broadcast on TV or radio.

This policy for audiovisual archives is even older,although it was initially limited only to the produc-tion of national TV. INA was created in 1975 tostore the archives of national channels and makeits collection available to producers and research-ers.At the time of its creation, it inherited all thearchives of the earliest national broadcasting com-

pany. It now deals with more than one and a halfbillion hours of TV and radio and more than onebillion still pictures stored on more than fifty milesof shelves. Currently INA has started to convert apart of its collection to digital format; about 300000 hours of radio and 200 000 hours of TV arecompleted, making it now one of the largest re-positories of audio-video archives (Auffret, 2000).Nevertheless, note that INA is just the archivist,not the copyright owner of all the deposited docu-ments. It often just operates as a central clearinghouse between buyers and information owners.Most of its services are related to indexing andsearching for relevant information in this huge quan-tity of audiovisual data.

INA as a Service Provider

In addition to its storage function, INA is also incharge of promoting cultural heritage by proposingmany services to clients. Basically, INA serves as apatrimonial archive. Archives are accessed mainlyby specialists; for instance, by TV producers want-ing to insert archive segments into a new produc-tion: perhaps a brief recall of historic facts withinthe news.Also, it is well known that journalists usearchives to prepare and maintain biographies of mostfamous people, ready to be broadcast within a fewminutes whenever needed. Documentary film se-ries also take advantage of 1NA's archive.

One less visible application concerns the study ofour cultural heritage. Many domain experts (histo-rians, sociologists, economists...), and even teach-ers, novelists and movie producers study audiovi-sual archives for research purposes: to better un-derstand some events, to elicit their relationships,as well as to catch authentic tiny details of a pastway of life in order to produce more realistic sto-ries.These people typically are the target users ofthe new service developed in the OPALES project.It is now a challenge for INA to take advantage oftheWeb to provide better service to these users inorder that the institution, as a whole, benefit fromtheir work.

The OPALES Project

In fall 1999, the French Ministry of the Economyinitiated the OPALES project within the PRIAM na-tional R&D planning program. OPALES is an acro-

4 3



nym for 'Outils pour des Portails AudiovisueLsEducatifs et Scientifiques' (i.e.: 'Tools for Audiovi-sual Portals for Education and Science'). Projectevaluation is scheduled during the fall of 2001.OPALES aims to develop a distributed environmentwhich boosts private expert work on multimediaarchives and provides support for collaborativeknowledge construction on the Web. Several insti-tutions and research labs and an industrial partnerparticipate in its elaboration and evaluation.The fi-nal system is not specifically dedicated to INA:somesimilar institutions dealing with patrimonial archivesare interested in it and collaborate in the project.The MSH `Maison des Sciences de l'Homme' in Paris,the 'Cite des Sciences et de l'Industrie', the CNDP'National Center for Distance Learning', and theBPS 'Program and Service Bank' of the 5th TV Chan-nel, as well as INA, provide both video archives andexpert users to work on them. In the first experi-mental stage of the project, the corpus has beenlimited to copyright free documents in order tomake experimental work cheaper.

The OPALES Project and the Web

OPALES is a private Web portal, open only to reg-istered users. Currently, access is restricted toproject staff. It supports expert users' activities whenworking on multimedia archives, preserves their datain private workspaces, and helps them share exper-tise. Each end-user accesses information within aprivate workspace. Any document (annotation aswell as archive) is handled as a private copy whichcan be virtually annotated, indexed, linked to otherinformation, edited to be inserted into new docu-ments, and so on. Direct anchoring of annotationswithin audio or video is supported. Elicited exper-tise may remain private in one's workspace or beshared. A shared ontology, coupled with indexingand search techniques (Chein 98) based on con-ceptual graphs, is used to handle semantically richannotations.

Overview of Private Workspaces

Exploring video archives on the Web

Until now, end users working on INA's video ar-chives needed to look physically at videotapes, ei-ther in the INA building or in their institution usingpurchased copies. Digital video on the Web and se-

cure transactions now make it possible for the us-ers to work from anywhere and for INA to drasti-cally reduce access costs.A large part of users' workonline consists of querying the archive base and thenexploring retrieved video sequences to decide whichparts are the most relevant. Efficient online workrequires fast response time, something which is notcurrently a strength of video on the Web. There-fore, it has been necessary first to develop specifictools to enable rapid searching of video. Typically,users need to look quickly at video contents andexplore them at variable levels of detail, not simplyto play parts of the videos. Video players are notrelevant for this service because they are designedfor playing, not for exploring. VideoPlayer orQuickTime rely on stored video: they support im-mediate seeking, but on the Web, their use is re-stricted to very short movies. RealPlayer usesstreaming. It allows playing a portion of a video be-fore receiving its entire contents. It is well suitedfor live video on the Web, but does not providereal-time exploration features.Video summaries likeimage albums enable rough overviews but are notsufficient for exploring videos at a detailed level.Aspecial `videoExploree tool (Nanard, 2001) has beendesigned and developed at LIRMM in order to mini-mize information transmission between the serverand the client station.The explorer server quicklydelivers on-demand computed short overviews ofany part of a video at any level of detail.A very simpleinteraction method is used to seek the overviewfrom the client station and select new segments toexplore. This technique enables users quickly toobserve videos at any level of detail and to focus onany relevant part, as precisely as at the frame level.Its external appearance and interaction method isvery similar to that of a video player.

Working in private spaces on the Web

OPALES provides its users with support for privateworkspaces. One of the expectations of the projectis to induce a feeling of ownership in users' work inorder to attract clients, keep them, and make itharder for them to go on working elsewhere. Incontrast to most Web sites and portals which en-able either passive reading of documents or cre-ation of private sites, OPALES supports active read-ing of documents.Active reading consists of directlyworking on documents as if they were private; forinstance, directly annotating or editing them. In an



active reading environment, the reader does notdistinguish between interacting with an archivaldocument and interacting with private notes. Suchannotation features are proposed in many systems,but very few are actually effective.

In OPALES, any document displayed on the user'sscreen is always a virtual private copy.When a userselects a document to be displayed, automaticallythe server brings back both the document itself andall the references to the annotations previously at-tached to this document by the user, plus thoseexplicitly imported, in context, from the sharedknowledge bank.Since annotation contents are firstclass objects, they are handled exactly like otherdocuments. One can annotate annotations at will,thus inducing a private hypermedia structure overthe set of documents.

Users tan also prepare private documents, includ-ing edited segments from archives. For instance,suppose a history teacher prepares a course, in-cluding in it selected relevant archival sequences asillustrations, with private comments added on thesound track and with some minor graphical en-hancements such as indicating names of personsdirectly on images. Such documents are handled inOPALES simply as edit lists dynamically interpretedwhen displayed.They are rebuilt from the archivesat playing time.Thus any annotation of the includedsegments anchored in the archives becomes avail-able also from the private document, including notescreated later.

Users can also flatten documents to use them out-side of the OPALES environment, but in this casethey lose all links to the OPALES environment. Flat-tening a document also triggers the evaluation ofthe copyrights of included archival segments.

Annotating documents and videos

An annotation results from an explicit user action.One 'annotates' a document by linking to it metadatathat we call the 'annotation contents'. In essence,an annotation references the annotated document.Annotations as such are described separately fromthe annotation contents by a RDF descriptor.Thereis no restriction on the nature and contents of theannotation or on the annotated document.Anchor-ing can be done into the annotated document. In-

14

ternally it uses an Xpointer notation (http://www.w3c).Anchoring into video documents inter-nally relies upon a very simple SMIL description(htpp://www.w3c) of archival movies that just takesinto account the actual segmentation that occurswhen indexing the archive in the database. All ofthe precise anchoring is expressed as time-codedsegments, enabling anchors to overlap where thereis suitable enabling stratification (Smith 1992) of an-notations.

It is important to note the asymmetry introducedin OPALES by using hypermedia typed links. By na-ture, annotation links are different from other links.A link to a document is not necessarily an annota-tion link. For instance, it may just be a 'citation' link.Annotation links result from a user explicit annota-tion action which internally creates the RDF de-scriptor and registers it in the database. Navigationlinks are handled in a more classical manner. Forinstance, the annotation contents can be an HTMLdocument with links to other documents; such linksare not considered as defining annotations since theydo not result from an annotation action.

Annotations are objects in the sense of 00 pro-gramming.Specific editors are available for each classof annotation.Among them, on the simplest side, asimple text editor or a XML/HTML editor enablesusers to produce documents; at the opposite ex-treme, a specific NCG 'Nested Conceptual Graph'editor enables users to attach semantically rich andcomputable (Mugnier, 2000) descriptions of docu-ments as indexing annotations.

Overview of Knowledge Sharing inOPALES

Interest in private workspaces would be very lim-ited if they did not communicate. Therefore, themost important aspect of OPALES is its supportfor knowledge sharing among user communities. Itenables any users to export part of their work andto import into their own workspace exported partsof other users' work. The OPALES basic belief isthat the more shared metadata bound to a docu-ment, the greater the value of that document tousers.Typically, exporting and importing annotationsprovide mechanisms for sharing knowledge and thuseliciting shared knowledge



Lesson Learned from the Web

The Web currently is the largest shared informa-tion structure in the world. Studying the evolutionfrom the poorest HTML I to the XML based lan-guage family provides a rich set of lessons.

The most important requirement to enable large-scale collaboration is rigorously to define a simplebut powerful shared language, and then support itsextensibil4 Lack of rigor quickly leads to theTowerof Babel phenomenon. Paradoxically, limited powerof expression of a language combined with a lack ofextensibility produces the same effect. Simplicity,precision and extensibility are required to enablelarge-scale collaboration.

To accommodate this observation in knowledgesharing ,OPALES provides its users with a very pre-cise but extensible ontology (Gruber 1993, Staab2000), and with very simple rules to support itsevolution. An 'archive' ontology is provided to de-fine precisely all of the vocabulary used in the ac-tual indexing of archival documents. But OPALESaims at capturing new expertise that, by nature, doesnot yet belong to this ontology.Therefore, it is nec-essary to provide mechanisms to extend the ontol-ogy at will. This is not an easy task; it is even anextremely risky one.The chosen solution relies on'private and sharable' extensions of the ontology.

Since any extension of the ontology and of the docu-ment indexing remains private, it has no conse-quence on the archive base and cannot introducemessy structures that would be troublesome forother users. On the other hand, if a subgroup ofusers trusts a given set of extensions, they can useit as a dynamically shared extension to the ontol-ogy, just for themselves. The OPALES knowledge-sharing engine enables users to work with dynami-cally extensible ontology and indexing of documents.

The notion of 'point of view' in OPALES is the keyfor dynamically managing private and shared exten-sions, indexing, and annotations, and as a conse-quence, for building and managing digital communi-ties of experts who add value to the archive.

The Central Notion orPoints ofView'in OPALES

Since OPALES is a private workspace strongly dedi-cated to supporting digital communities of experts,the management of dynamically evolving virtual com-munities is a major component of the system.Wehave chosen to permit the free creation of com-munities and free access to them.A user may cre-ate a community just by specifying a concern.Wecall it a 'point of view'.That user just has to write aninformal document to define the concern and toindex it formally in terms of the shared ontology.

In OPALES, any piece of information has a descrip-tor which identifies its 'point of view'. A point ofview is not at all an index of the document, butrather a mark that denotes which category of us-ers might be concerned with this information. Pointsof view have some weak similarities with'newsgroups,' but it would be erroneous to pushthe metaphor too far. For instance, consider anno-tations of a politician's speech: one user may anno-tate it from an 'economist's point of view' evaluat-ing long term consequences;another may do it froma 'rhetoric expert's point of view' discussing thespeech structure; whilst yet another may focus ondetails of hand motion and face expression from a'psychology expert's point of view'. Other psychol-ogy experts could be interested in retrieving docu-ments annotated by colleagues in their virtual com-munity. By declaring an annotation with this pointof view, the annotator locates it within the concernof this virtual community.

Points of view induce an a posteriori classification ofusers based on their stated concerns.Tagging eachpiece of data with the author's point of view implic-itly defines a partition of information spaces. It isstored in the'workspaces' database.There is no needto use a partition since the formal indexing by pointsof view enables the NCG search engine (Chein,1998) automatically to determine the location ofthe point of view, especially its specialization(Mugn ier, 2000)

A point of view is either private or public.A privatepoint of view makes sense only for its owner. It canbe used as a kind of private classification system,

1 4 6



with a personal vocabulary.A point of view becomespublic if its owner 'exports' it, making its descrip-tion and indexing visible to other users.As a conse-quence, public points of view are retrieved like anyother documents, thus enabling users to be awareof declared public points of view that are close totheir concerns.

Using points of view is quite simple.Any editor win-dow is assigned a default point of view which tagsany new document created in this window.The usermay assign to it any other point of view, private aswell as public, retrieved either from a favorite pointof view list, or from a search of existing ones closeto the query. If no point of view matches the query,the user may create one and export it the queryalready is a good base for starting to index the pointof view. It just has to be informally described moreprecisely. Since points of view are attached to thewindow rather than to the user, one may easilyhandle several points of view, private or public, in aworkspace.

Since all pieces of information have a point of view,any extension done to the ontology also has a'pointof view'. Extensions can be private or public. In thiscase, other users can share it. OPALES also sup-ports regulation mechanisms for points of view, es-pecially 'moderator approved points of view' thatare suitable for asserting the consistency of exten-sions done to the ontology.

OPALES points of view have been designed to en-able knowledge sharing.Any displayed document hasa 'reader point of view' that specifies enhancementsby annotations, indexing, ontology extensions, andso on. The readers' point of view is distinct fromthe author's, since they may also enjoy reading in-formation stored with other points of view to get awider understanding of a document; but the read-ers have no reason to place any writings in suchpoints of views.They read from a point of view thatcombines several authorial points of view, but writefrom their own. For instance, let us suppose that asociologist studies the speeches of a politician inorder to write an essay on 'tricks to convincecrowds'.This user would benefit from importing intothe workspace the points of view of the 'psycholo-gist expert' who has analyzed the details of handmotion as well as those of the 'rhetoric expert', but

147

has no reason to write in those points of view ifthe usei's current writing does not concern thesevirtual communities.

Any window has a 'reader's current point of view'that is built as a list of imported points of view.When a document is delivered in this window, thesystem also delivers and displays the document'slist of public annotations that have been created inthe points of view currently included in the reader'spoint of view..This enables users to browse any ofthese annotations and recursively annotate them.

The reader's point of view also acts on the searchengine. First, during the query preparation, it filtersand expands the ontology to help the user choosethe proper vocabulary. Second, the search enginetakes into account both the actual indexing doneon archive and also any public indexing done withthe points of view included in the reader's currentpoint of view. This enables both writing far moreprecise queries which are domain dependent, andretrieving very short segments matching such que-ries, since annotations can be anchored freely andprecisely into the video archives.

Museums keep track of our history and culture.TheWeb can help them make the past available to all.The most important thing is not simply to showremarkable crafts, but to make their impact on ourculture understandable. Museums are like icebergs;the part of collections made visible is rather small.Museums also preserve huge and rich material instorage, but in most cases, it is poorly used anddifficult to study.This typically is the case with hugeamounts of digital information that contains a quitecontinuous record of our social evolution. Payingmuseum staff to exploit these resources efficientlyis far beyond museum budgets. Other solutions areneeded. OPALES is one of the solutions. It relies oncollaborative work on the Web.

Beyond the well-known role of the Web in infor-mation access, a far more important applicationdomain is emerging:collaborative and distance work.The effectiveness ofWeb users' work is now greaterthan most analysts of the last decade could haveimagined. Breaking encryption keys that were sup-posed to require mainframes for centuries of workhas taken months owing to the shared distributed



power of home computers. Dealing with freely dis-tributed collective knowledge is a great challengefor the new century.The Semantic Web project ini-tiated byTim Berners-Lee (Berners-Lee 1998) betson the power and efficiency of freely organized col-laboration. Several other initiatives propose tech-niques for distributed annotation of Web pagesbased on a RDF schema, to improve the efficiencyof search engines (Kahan 2000).0n a smaller scale,OPALES bets on mixing distributed work within acentralized environment.The choice depends on theneed for a well-balanced solution. Letting users dofreely what they want has obviously been a successwith the Web: a reliable structure has slowlyemerged. But in smaller environments, this strategylacks statistical regulation mechanisms,so it is knownto run out of control easily. Constraining users inorder to enforce controlled structures requiresexternal force.lt never works in open environments.The solution is a balance in which users both feelfree but easily find attractive clusters where theirexpertise is recognized and can be cumulated withothers.

The point of view mechanism in OPALES is easy touse. It is sufficiently free to allow everyone to use itat will, without any regulation mechanism. But byits nature, it leads to the formation of virtual usergroups within which knowledge can be elicited in aconsistent manner, relying on small and local exten-sions of a shared ontology. This feature enablespeople working on the same topics to cumulatetheir efforts. Furthermore, the 'reader's currentpoint of view' provides the means to trigger inter-disciplinary work by importing knowledge fromother domains for better understanding.

OPALES currently is a tool for experts, mainly be-cause exploring the storerooms of museums is notyet for end users. But expert work has results whichusually are presented to end users. Currently thecorpus chosen to bootstrap OPALES contains,among others things, a rich collection of documentsabout the history of modern mathematics, espe-cially hundreds of hours of records of work meet-ings of the Bourbaki group.As an example, explor-ing and annotating these historic documents willbenefit the 'archaeology' of science by providing abetter understanding of the evolution of this disci-pline, and will make these enhanced documents ac-cessible to a larger audience.

References

Auffret G. (2000) Structuration de documents audiovisuelset publication electronique, Phd thesis, Université detechnologie de Compiègne, France

Berners-Lee,T. Semantic Web Road map, http://www.w3.org/DesignIssues/Semantic.html, 1998

Chang, S.F. et al.( I 997),VideoQ:An automated content-based video search system using visual cues. InProc. ACM Muttimedia'97, pp. 313-324.

Chein, M., Mugnier, M.L., & Simonet.. (1998), G. NestedGraphs:A Graph-based Knowledge Representa-tion Model with FOL Semantics, in Proc. 6thInternational Conference on Principles of KnowledgeRepresentation and Reasoning (KR'98), pp. 524-534,Morgan Kaufmann Publishers.

Gruber,T.R. ( I 993),Toward principles for the design ofontologies used for knowledge sharing. In NicolaGuarino and Roberto Poli (eds), Formal Ontology inConceptual Analysis and Knowledge Representation,Kluwer. On line: http://ksl-web.stanford.edu/knowledge-sharing/papers/onto-design.rtf.

Hauptmann,A. and Smith, M. (1995),Text, Speech, andVision for Video Segmentation:The InformediaProject, AAAI Fall 1995 Symposium on ComputationalModels for Integrating Language and Vision, 1995. Seealso: http://www.informedia.cs.cmu.edu/

Houghton, R. (1999), Named Faces: Putting Names toFaces. IEEE Intelligent Systems Magazine,Vol. 14, No.5, pp. 45-50.

Kahan, J., Koivunen, M.R., Prud'Hommeaux, E.. & SwickR.R.Annotea: An Open RDF Infratructure for SharedWeb Annotations.

Mugnier, M.L Knowledge Representation and Reason-ings Based on Graph Homomorphism, in Proc. 9thInternational Conference on Conceptual Structures(ICCS 2000).

Nanard, M., Nanard, J., & Payet, D. (2001), Designrationale of a Video Explorer, In ACM conferenceCHI'2001, Seattle,Washington.

011igschlaeger,A., Hauptmann,A. (1999), MultimodalInformation Systems and GIS:The InformediaDigital Video Library, ESR1 User Conference 1999.

,:

1 4 8©Archives & Museum Informatics, 2001 15 1

Nanard,etal.,OPALES:An Environment for Sharing Knowledge

Staab, S., Erdmann, M., Maedche,A., & Decker. S. (2000).An Extensible Approach for Modeling Ontologiesin RDF(S), Workshop on SemanticWeb associated toECDL'2000.

Smith.A., & Davenport, G. ( I 992);The StratificationSystem:A Design Environment for Random AccessVideo. In ACM workshop on Networking andOperating System Support for Digital Audio and Video,San Diego, California.

149


Being aMuseum,Digitally

150

"Beyond museum walls" -A critical analysis of emergingapproaches to museum web-based education

Kevin Sumption, Powerhouse Museum, Australian MuseumsOn-Line, Australia

Abstract

"An effective educational museum might be described as a collection of labels bearing instructions, each of themillustrated by a carefully selected specimen." George Brown Goode, Assistant Secretary at the Smithsonian, 1896(Alexander, 1979).

Although Goode's assertion may look naive by contemporary pedagogical and museological standards, I'm sure manyof us wish that the process of museum education were indeed that easy. However, we know that such assertions areanathema, as both the primacy of the object and our understanding of learning have advanced considerably in the last100 years. Since the arrival of photography in the mid nineteenth century, the interpretive pre-eminence of artefactshas been steadily eroded by a succession of media. More recently, interactive multimedia (IMM) have penetratedhomes as well as museums, transforming visitor expectations as well as the way modern museums approach education.

The most ubiquitous of contemporary 1MM, the Internet, is making steady progress as an interpretive tool withinmuseums. However, its major impact is being felt beyond museum walls. As an outreach agent, the Internet hascaptivated many museums and particularly their educators.As a communication medium, the Internet allows museumeducators to enter the homes and schools of students without their ever needing to visit the museum. So it's notsurprising that some museum education products try to simulate the spatial and social experience of visiting a museum.However, this approach is just one of many resource "types" educators have deployed as they grapple with thepromise and reality of on-line education.

As part of the research team developingAustralian Museums On Line's new Education Gateway, I have spent considerabletime examining why and how museums are using the Internet for education outreach. In this paper I will explore these,as well as the diversity of emerging on-line education expressions. I will also review current research into the uniqueinterface, navigation and content preferences of various learners and, just as important, discuss best practice teachingand learning strategies to help museum educators develop more effective on-line educational resources.

Introduction

For all but eight of its first 100 years, my museum- the Powerhouse Museum - was part of the NewSouth Wales Department of Education. Not sur-prisingly,then, interest in the use of objects in teach-ing here has a long history, dating back to the 1890s,and still remains a critical component of Power-house Museum education practice. However, in thelast 100 years our understanding of the complexinterplay between visitors and objects that givesrise to learning has changed as epistemological andpedagogical research has evolved. From this,a num-ber of educational theories have developed to ex-plain how learning in object-orientated environ-ments such as museums can take place.Today thedirect-experience model is the one widely ac-cepted, particularly in science and technology mu-seums.

Direct-Experience Learning

Direct-experience-based learning divides ourworld into two distinct realms: the physical world

where objects exist and events happen,and our mindswhich are capable of memory and conscious thought.By allowing visitors to encounter real objects, labelsand a variety of media, we immerse them in a physi-cal world where they see, hear, touch, taste, smelland do. It is human nature that visitors should thenseek to understand these 'direct-experiences: Thismeans they must use their minds to find regularitiesand relationships between these new primary expe-riences and existing concepts and principles. In turn,these experiences help visitors develop a range ofskills from observation and reporting to deduction.

However, the direct-experience learning model haslimitations. First, the reader (visitor) must be suffi-ciently stimulated by artefactual encounters to wantto engage cognitively with them. This is a problemfor information technology objects such as micro-chips and personal computers, whose form is oftenuninspiring or masks function (Sumption, 1999). Crib-cally, for contemporary museums, the direct-experi-ence learning model is also restricted to the real



BEST COPYAVAILABLE

Sumption, Emerging Approaches to Museum Web-based Education

museum. The moment we move beyond museumwalls and into the realm of photo-mechanical andelectronic outreach, where the reality of the objectis supplanted by a mediated reality, the fundamentaldynamic of the learning experience is changed.

The Outreach Revolution

As with education, outreach has been an integralpart of many museums' core business. As early as189 I , my own museum opened its fifth satellite cam-pus on the grounds of the Broken Hill TechnicalEducation College.The Broken Hill branch was es-tablished to encourage the use of Australian prod-ucts and Australian materials in the manufacture ofarticles for everyday use. To achieve this, BrokenHill was supplied with a series of taxonomic dis-plays, all carefully selected and labelled by the cen-tral museum's staff. These were then dispatched,complete with hand-crafted cabinets and 2000 ob-jects, to the remote and inhospitable plains of NorthWestern New South Wales.

In the late Nineteenth Century, the limitations ofcommunication media, combined with the unique-ness of the museum concept, meant that nothingless than a complete reproduction of an exhibitwould have sufficed for outreach. But by the turn ofthe century, photography, and the process of itsmass-dissemination - photogravure, had opened upnew outreach opportunities and challenges.

Why Use Digital Objects?

Like contemporary travelling exhibitions, the pains-taking reproduction of displays such as those usedin Broken Hill ensure that museum educators andcurators retain authorial control over exhibitioncontent. However, the process of historical produc-tion, via print and film, is more problematic.As pho-tography and printing physically supplant the realityof the object with a new mediated reality, curatoror educator control is diminished by a set of inter-mediary authors - the camera, the lens, the film, thephotographer and the designer. Also, in the late1990s, when photo-mechanical outreach was itselftransformed through digitisation, a further interme-diary author was added.Through interactive digitaloutreach media, even the reader has become a pow-erful intermediary author. Nowhere is this more

152

evident than in the latest generation of active learn-ing resources now appearing on museum websites.

Since the privatisation of the Internet here in Aus-tralia in 1995, museums have increasingly sought touse the web to service the learning needs of pre-dominantly secondary school students.To do so theyhave harnessed the computational power of mod-ern personal computers via the World Wide Webto create a plethora of learning environments.Criti-cally, these environments have tried to attract andhold students by supplanting the inherent pedagogi-cally weak 'digital object' with more engaging inter-active experiences.As was evident in the first gen-eration of museum web pages, the process of 'ob-ject digitisation' can potentially strip artefacts ofcritical signifiers such as scale, weight, colour andtexture.Without these material signifiers, the digi-tal object is distorted in ways that can have a quali-tative impact on learning. Therefore it is not sur-prising that more recently museums have enhanced,or even supplanted 'digital objects' with interactivemultimedia affordances such as Java applets, stream-ing video and IRC, etc.As I will show, this shift hashad a liberating effect on many museum educatorswho can now concentrate on the development ofeducational products centred not just on the ob-jects, but on the students themselves.

Six 'Active Learning' Typologies

For evidence of this shift, look no further than theeducation winners of the Best of the Web. In 1998the Odyssey website (http://www.cc.emory.edu/CARLOS/ODYSSEY/index.html), utilised museumobject images and stories to 'inform' students ofthe daily life of Middle Eastern and sub-Saharanpeoples. In 2000,the Puppets Action website,(http://www.childrensmuseum.org/artsworkshop/puppetshow.html) equipped students with tools to'create' their own on-line puppet show. This shiftfrom seeing the web as an information medium toseeing it as an active learning environment has con-solidated. Now more than ever, many museums arecreating web-based learning environments whereusers, not educators, are empowered to make de-cisions about the tasks, content, navigation, presen-tation and assessment activities they undertake.Thusa range of active learning web typologies has startedto emerge. These typologies range from mature

156 @Archives & Museum Informatics, 2001


education meta-center products, which continue toutilise digital object images and records, through toimmature expository resources, which have onlytentatively embraced the active learning philoso-phy.What distinguishes one typology from the nextis not just content and enabling technology, but alsothe cognitive strategies each employs.These are thegenerative, problem-based, creative play and exposi-tory methodologies that help learners to build ontheir own knowledge structures (schema) to con-struct new knowledge and understanding. Theseschema are consciously, and sometimes uncon-sciously, deployed by designers and educators withvarying success. So by offering the following typo-logical overview, I hope to provide a snapshot ofcurrent thinking and practice. In turn, I hope thiswill help museum educators understand the com-plexities, pitfalls and challenges of creating 'activelearning' web resources for museums.

I . Education Meta-centers

Of the six museum education typologies I have iden-tified, Education mew-centres, along with Creative play

resources, are still primarily concerned with teach-ing and learning via museum objects and documents.However, as we shall see, best practice in both isnow exemplified by activities which enable users tomanipulate and interrogate both the physical andintellectual content of artefacts.

Education meta-centres are web-based resources thatallow students to explore, investigate, compare andevaluate textual and pictorial information frommuseum collections and archives. Most utilise OnlinePublicAccess Catalogues (OPAC) to make museuminformation accessible via the web. However, themere accessibility of such data doesn't in itself guar-antee learning.Therefore, many developers of data-rich resources, such as the Museum of Victoria'sButterfly site (http://www.museum.vic.gov.au/bioinformatics/butter/), have adopted generativelearning methods to help students convert raw datainto useful knowledge products.

Supporting generative learning

Museum Victoria's Butterfly site allows middle andupper high school students to access a single data-base of 30,000Victorian butterfly records via sevendiscrete searches. Each search is designed for stu-

dents of a particular age and studying a particularsyllabus, and culminates in a series of comparativeactivities that allow students to mentally experi-ment (generate ideas) with various data sets. Stu-dent experimentation and analysis is via a series ofweb-based mapping tools.These allow them to plotand overlay the distribution data of butterfly spe-cies with Victoria's vegetation,altitude, river and zoo-geographic information.The result is a unique sche-matic of species and topological data which stu-dents can endlessly modify and interrogate.This al-lows them to develop learning outcomes builtaround their own interests, and in turn can resultin deeper conceptual understanding and affinity withthe subject.

2. Creative Play Resources

Like nearly all active learning resources, educationmeta-centers are primarily designed for secondaryand tertiary students. Currently there are few mu-seums developing active learning resources for un-der eights.This is surprising as most museum edu-cators agree that young children have just as muchright to make sense of their world with digital me-dia as do older children.Although small in number,there are a few examples of what I call Creative playresources for under eights.The Museum of ModernArts' Art Safari (artsafari.moma.org/) is exemplary.Art Safari uses paintings and sculpture from MOMA'scollection to help children develop their observa-tional skills.The site includes creative activities whichencourage students to examine and creatively re-spond to the forms, stories and characters inRousseau, Kahlo, Rivera and Picasso artworks.

Playing to learn

Early childhood learning research points to the valueof activities which stimulate and support creativeplay. Creative play is an important part of youngchildren's learning and describes those experiencesthat encourage them to explore and test ideasthrough the process of making. Most project-basedwebsites offer some degree of creative play if theyallow students to create original written or oralcompositions. However, for young children, draw-ing and painting-based activities are often the mosteffective. Art Safari supports this kind of creativeplay by helping children draw a portrait of a pet ora fantastic animal using a downloadable digital pal-



ette. Ordinarily, young children find this kind of ac-tivity very stimulating, particularly as it allows themto explore subjects that are close to their hearts.Pets, parents, brothers and sisters, etc., are all ofdeep personal interest to young children. Thesesubjects, when combined with creative play, excitechildren to keep on investigating, listening, talking,designing, constructing, asking questions, reading andwriting. (Downes, et al, 1999).

3. Electronic Field Trips

While Creative play resources are ideal for helpingdevelop observational and hand- eye coordinationskills, Electronic field trips are best suited for teach-ing complex cyclic, structural or relational conceptsby immersing students in credible microworlds.Credible microworlds convincingly 'transport' stu-dents through time to historical sites; through spaceto the museum; or through their imagination to fic-titious and fantastic worlds.The effectiveness of thistransportation is in part dependent on audio-visualmaterials and interactive technologies to 'realisti-cally' simulate an experience or environment. Bethey imaginary planets, tropical rainforest or civilwar battlefields,all these microworlds must not onlycapture the attention of students, but also engagethem in stimulating activities for long periods oftime.To achieve this, many electronic field trips utiliseanchored instruction techniques.

Anchored instruction

Anchored instruction activities utilise sets of inter-connected problems to encourage users to acti-vate their reasoning, deductive and investigative fac-ulties.An excellent example is Brookfiekl Zoo's GoWild field trip (http://www.brookfieldzoo.org). GoWild challenges students to solve the problem offinding a safe way through the Ituri forest of centralAfrica.To help, developers have provided four ava-tar guides who drive the narrative by bombardingusers with questions.As with all anchored instruc-tion activities, Go Wild challenges are based on realproblems like using a compass and map, makingbarkcloth or choosing suitable forest foods to eat.The realistic nature of these problems is impera-tive to encourage students to take ownership oftheir dilemma, and in turn, responsibility for activelyovercoming it.

4.Video Conferencing

Like electronic field trips, video conferencing activi-ties utilise the connectivity of the web to transportstudents and teachers beyond their classroom orhome. However, unlike current electronic field trips,video conferencing can utilise synchronousaffordance technologies to connect students andteachers to the real museum's staff and exhibits, inreal time. Here the microworlds created are notsimulated, but utilise the performances and 'scen-ery' of staff and exhibits. Ultimately it is the photo-genic nature of these displays, together with theaffability and open-endedness of the student-pre-senter dialogue, which can determine the level ofmeaningful engagement.

Unfortunately video conferencing is bandwidth hun-gry. Therefore, museums need ideally to have ac-cess to high-speed networks like that of the NewJersey Department of Education's Distance Learn-ing Network.This state-wide ISDN/ATM networkallows schools from across New Jersey to link tothe Liberty Science Center's (LSC) 250 exhibits.Currently, access to LSC exhibits is via a series ofE-connection programs, e.g. the aquatic ecosystemspackage.This video conferencing activity links a classof upper primary students studying river ecosys-tems to a biologist based in the LSC's Hudson Riveraquarium.After a brief introduction,a camera scanstanks of toadfish, eels and spider crabs.A series ofcarefully choreographed activities follow, prompt-ing students to ask questions about the habitat andbehaviour of each species. Quickly a dialogue be-gins as the Biologist moves from presenter to con-versation facilitator. As well as conversational abil-ity, the presenter needs in-depth knowledge to cre-ate an active learning experience built around open-ended dialogue.

The nature of digitally based, open-ended experi-ence is complex and ideally needs to be sufficientlystimulating to children to elicit many and varied re-sponses and questions.To achieve this, research hasdemonstrated that digital resources need to be so-phisticated enough to ( I) encourage children torespond in thoughtful ways, (2) offer response tochildren's answers, (3) offer variations that are child-controlled, and finally (4) cater to each child's abil-ity level, cognitive development and computer skills,



as well as culture/s and language/s (Downes, et al.,1999). For all these reasons, we can see why humanfacilitation of genuinely open-ended learning expe-riences is still so necessary, and is a learning at-tribute of video conferencing unmatched by anyother active learning typology.

5. Museum Learning Networks

Museum learning networks have been around, ininternet terms, for a relatively long time. But sur-prisingly, they have yet to flourish, and indeed eventoday most museums are preoccupied with devel-oping resources for students working independently,either from home or within a classroom, rather thanworking in teams.This is understandable as the com-plexities of building resources to accommodatemore than one simultaneous user can be immense.

Learning networks like the Science Learning Net-work (http://sIn.org/) or STEM (http://www.nmslac.uk/education/databased3/stem/stem/stemintro.asp) characteristically provide aggregatedaccess to an array of field trip and expository re-sources themed around a single discipline like sci-ence, contemporary art or the environment.Twelvescience centres from across Asia, Europe and theAmericas provide content and manageThe ScienceLearning Network (http://www.sln.org/), establishedin I996.Within SLN, students can conduct a seriesof genetic, solar, aviation and oceanographic experi-ments or participate in one of three collaborativeprojects, including Monarchs and Migration (http://www.sci.mus.mn.us/sIn/monarchs).

Learning through collaboration

Conceptually Monarchs and Migration uses the natu-ral phenomenon of migration to encourage an in-ternational community to build a rich textual, artis-tic and photographic snapshot of Monarch Butter-fly migration. Classes in the USA, Mexico, NewZealand and Australia use galleries and tracking stud-ies to share observations and photographs of mi-grating Monarch Butterffies.Whilst offering excel-lent content and tools, Monarchs and Migration, likemany learning network activities, is limited in the'genuinely' collaborative opportunities it offers.These are ideally activities that foster cooperativeproblem solving through continuous communica-

tion.This kind of collaboration is best facilitated viareal meetings, or chat rooms and listserves, withinwhich the process of creating a joint product isnegotiated. These products can be anything fromweb-based scrapbooks and photo-albums to reportsand diaries, but their design and structure need tobe carefully considered to encourage mutual re-spect and trust amongst collaborators. An excel-lent example is the National Museum of Scienceand Industry's STEM project.

Users as authors

The STEM project is a learning network with a dif-ference. Instead of museum educators and cura-tors producing content, students and teachers do.Students and teachers who visit one of the NationalMuseum of Science and Industry's three museumsare invited to compete with one another to createa web resource based around their visit. Invariablythose teachers and students who take up the chal-lenge work in teams, both real and virtual. In thisway STEM students create products that exploreissues, ideas and problems that have roots in theirunique experiences, interests and opinions. As inany active learning experience, given the opportu-nity to control content directly, students are muchmore likely to be internally motivated to exploreissues and ideas close to their hearts.The bonus isthat these very same issues are much more likelyto be of interest to the resources' intended audi-ence other students.

6. Expository Teaching Resources

Often found on science and technology museumwebsites, Expository teaching resources are only nowbeginning to embrace active learning principles. Pri-marily developed for children twelve years old oryounger, Expository resources ordinarily are com-prised of a series of highly structured computer-directed activities through which a set of principles,laws or theories is presented (expository).

The Australian National Science and TechnologyCentre's Funzone is typical. Funzone (http://www.questacon.edu.au/fun_zone.html) features aseries of Macromedia-based puzzles, activities andlaboratories where students can conduct web-basedexperiments.One of these asks students to observe



a series of animated lines moving across their screen.After 60 seconds the lines stop moving; however,students observe a motion-after-effect.That is, thelines `seem' to continue to move across the screenbecause the brain is so accustomed to the move-ment, it `imagines' they are still moving.

Addressing metacognition problems

Interaction with other students and with theMacromedia animation itself has been deliberatelyconstrained in Funzone.While this may seem at oddswith active learning principles which traditionallyencourage open-ended learning experiences, these`closed' learning resources are not without merit.Potentially they can alleviate some of themetacognition problems often encountered byyoung children. Metacognition problems can occurwhen people lose sight of, or can no longer accu-rately monitor, the success or failure of their ownlearning because activities are not stimulating, orare too complex. A number of studies have indi-cated that poor metacognition can also arise whenstudents fail to develop clear-cut learning goals be-cause they are given total control of their learningexperience. However, metacognition problems canbe addressed through careful navigation and inter-face design coupled with rigorous remedial evalua-tion.While this type of work is often time consum-ing, it is undoubtedly worthwhile as the vast major-ity of younger children, like older children, still pre-fer to control their learning experiences.

Therefore the dilemma for Expository resource de-velopers is to find a compromise capable of accom-modating both the normal student's desire to ex-plore (active learning) and a specific student'smetacognition problems.While there are no easyanswers, one possible approach is to try to enrich'experiments' with layers of additional activity. Inthe case of Funzone, students could have been of-fered extension exercises with different colouredlines or lines moving at different speeds availableunder student control.This kind of 'value adding' toexpository resources has the potential to lessenthe likelihood of metacognition problems, as it sup-ports limited student exploration and at the sametime embraces active learning. Critically, this layer-

156

ing also has the potential to expand student under-standing, as it more clearly defines the limitationswithin which phenomena such as motion-after-ef-fect operate. An excellent example of this kind ofExpository layering is the National Museum ofAustralia's Ann Fibian website (http://www.nma.gov.au/education/ann/index.htm). Herestudents investigate soil erosion and deforestationphenomenon via a series of expository activitieshaving a carefully set mix of computer and learnercontrolled activity.

Do They Work?

As these six typologies demonstrate, there is nosuch thing as a sure-fire way to guarantee learningor teaching, only methods or devices (pedagogic ortechnical) which together facilitate the individual'sendeavour (B. Whalley, 1995). Indeed, the deploy-ment of any one of the cognitive strategies I've dis-cussed is not in itself a guarantee students will evenpay attention.The best museum educators and de-signers can do is to carefully consider these cogni-tive strategies, along with pedagogically appropri-ate content, design and affordance strategies. Buteven then we need to recognise that learning is stilla process beyond the total control of museums,andis predominantly invested in the individual learner.So even if the mix has been carefully prepared,workcan still be undone by the relatively unknown im-pact of users' unique cognitive, environmental andcultural characteristics.That is why I wish to closethis paper with a plea for further research, particu-larly remedial and summative evaluation. It is notuntil we have a better understanding of real out-comes, against those intended, that museums willbe able to advance the practice of active learningproduct development. However, I'm mindful thatactive learning environments are expensive to de-velop, as are evaluation studies. What's more, I'veyet to come across a workable evaluation frame-work. So in a year when AMOL will be launching itsown museum learning gateway, I'm very eager totalk to museums or universities interested in shar-ing costs and expertise to develop appropriateevaluation frameworks for active learning resources.



References

Alexander, E. (1979) Museums in Motion,An Introductionto The History and Functions of Museums. AmericanAssociation for State and Local History, Nashville.

Downes,T,Arthur, L., Beecher, B,. Kemp, L ( I 999).Appropriate EdNa Services for Children EightYearsandYounger. Education Network Australia,http://www.edna.edu.au/EdNA

Haugland, S.W. (1992). Effects of computer software onpreschool children's developmental gains., Journalof Computing in Childhood Education, 3 (1), I 5-30.Sumption, K, ( I 999). Hall of Mirrors:TheDilemmas of Presenting Information TechnologyCulture Through Information Technologylnteractives and Artefacts. Selected papers fromICHIM 1999: Pittsburgh:Archives and MuseumInformatics.

Whalley, B.(1995). Teaching and Learning on the Internet.Active learning Journal. Issue 2. Institute forlearning and Teaching in Higher Education, UK.

157

@ Archives & Museum Informatics, 2001 I 6 I

1000 years of the Olympic Games:Treasures of AncientGreece: Digital Reconstruction at the Home of the Gods

Sarah Kenderdine, Powerhouse Museum,Australia

Abstract

This paper results from a recently completed project that augmented an exhibition at the Powerhouse Museum,Sydney, entitled 1000 Years of the Olympic Games: treasures of ancient Greece (July 18 November 18, 2000).ThisExhibition offered an opportunity to supplement the traditional visitor experience with the introduction of 'virtualreality components.These include a 3D digital reconstruction of the archaeological site of Olympia in 200 BC: thelaser scan of the famous statue of Zeus from the archaeological museum in Athens: and the construction of a large-scale Web (http://www.phm.gov.au/ancient_greek_olympics) that combines the Exhibition components with a host ofother information sources.

The paper reflects on the theoretical concerns companion to the processes of digital reconstructions using historicaland archaeological data sets. From this position it is possible to examine the curatorial and technical decisions madein creating the works.An outline of the web architecture and design is presented.A statistical analysis of the web sinceits launch is also examined.

as part of

AN EXHIBITION ORGANIZED AND LENT BY THE HELLENIC MINISTRY OF CULTURE,ATHENSAS A CONTRIBUTION TO THE CELEBRATION OF THE SYDNEY 2000 OLYMPIC AND PARALYMPIC GAMES.

DEVELOPED BY THE POWERHOUSE MUSEUMIN CONJUNCTION WITH THE UNIVERSITY OF MELBOURNE

Introduction

Here we enter the ancient sanctuary of the godsof Olympia, witness Zeus the wielder of the thun-derbolts, and walk the archaeological site of Olym-pia as it appears today.The works under discussionresult from a recently completed project that aug-mented an exhibition at the Powerhouse Museum,Sydney, entitled 1000 Years of the Olympic Games:treasures of ancient Greece (July 18 November 18,2000). This Exhibition offered an opportunity tosupplement the traditional visitor experience withthe introduction of virtual reality components.These were a 3D digital reconstruction of the ar-chaeological site of Olympia, in 200 BC, the laserscan of the famous statue of Zeus from the archaeo-logical museum in Athens, and the construction ofa large scale Web which delivered the Exhibitioncomponents together with a host of other infor-mation sources and interpretive information.

This paper begins by introducing some of the vi-sions, opportunities and cautionary perspectivesthat are companion to the processes of archaeo-

logical reconstruction using digital tools.Observantto the potential and difficulties inherent in digitalreconstruction, the paper will then discuss somethe curatorial and technical aspects of creating theinformation complex: the reconstruction of Olym-pia; the laser scan of Zeus; and the presentation ofmaterials on the Web.

The paper examines the data acquisition process atthe site of Olympia, and the archaeological and his-torical data considerations that were the basis forcreating the 3D visualisation of the site in 200 BC.The scan and display of the statue of Zeus is alsooutlined. A brief description of the web contentsintroduces the technical, architectural and curato-rial decisions that were used to define the design.This section concludes with a statistical analysis ofthe Web usage since its launch. Reviews of the Webby industry,and education and research awards thatthe works have attracted to date, conclude the dis-cussions.



Kenderdine, Digital Reconstruction at the Home of the Gods

Theory in Digital ArchaeologicalVisualisations

Opportunities

In theory, digital reconstructions represent a para-mount tool of enquiry for archaeologists.As Forte(2000) notes, this can occur when:

..."interactive-computational simulation"becomes the methodical ideal of the scien-tist reason and observation permit the breakdown of an event into its primary elementsand from these it is reconstructed.A tightlink is established between computationalinformation and the "architecture ofthought", because both embody man's de-sire to be the "architect" of the "world". Aportion of the real is no longer representedby a chain of ideas in which all pertinent in-formation is housed; instead it is recon-structed in a way that the observer can im-merse himself, react to it and be reacted to...

...Through the collection of multiple formsof data otherwise lost, that become homo-geneous after calculation, a model of theevent can be constructed.This allows for thestep-by-step visualisation of all phases, pro-viding humanity with extraordinary predic-tive power, since each level is the necessarycondition for the next one.

The techniques of immersion, interaction, and sen-sorial interaction are all based on perceptive mecha-nisms (Gregory, 1998 in Forte, 2000).They repre-sent instruments that permit the user to operatein synthetic space; that is, to be able to understandsynthetic space in all of its richness through theshifting of points of view (Forte, 2000). In archaeo-logical enquiries, the richness of the medium forteaching and theory testing in the form of alternatevisualisations is potentially unlimited.

To situate digital visualisations of archaeological andhistorical sites into museums and to translate themfor the Internet represents a powerful form of in-terpretation. It allows for complex datasets to bedrawn together forming products that are multi-sensory and easily accessible.

159

Colonna (1994) presents his view of the flexiblenature of digitally reconstructed environments (asdiscussed by Barce16).

A virtual world should be, then a model, aset of concepts, laws, tested hypotheses andhypotheses waiting for testing. If in standardtheories, concepts are expressed linguisti-cally or mathematically, in virtual environ-ments, theories are expressedcomputationally, by using images and render-ing effects. Nothing should be wrong or"imaginary" in a virtual reconstruction, butshould follow what we know, be dynamical,and be interactively modifiable.A virtual ex-perience is then a way of studying a geo-metrical model a scientific theory ex-pressed with a geometric language insteadof studying empirical reality.As such it shouldbe related with work on the empirical real-ity (excavation, laboratory analysis).As a re-sult we can act virtually with inaccessibleartefacts, buildings and landscapes throughtheir models (Barceló, 2000).

For the Olympia project, the empirical reality wasbased on numerous excavation reports and histori-cal and scholastic sources discussed below, and inaerial and geological ground survey data.A consid-erable component of the project was the acquisi-tion and analysis of source information, especiallyas a high degree of veracity and verisimilitude wasrequired. The digital reconstruction process at-tempted at all times to create an archaeologicallycorrect interpretation of the research materialsavailable.

Caution in reconstruction

Increasingly, archaeological reconstructions are usedin cinemagraphic-digital formats or 3D interactiveenvironments. These virtual models are criticizedfor their ability to portray scientific speculation as'truth'. In general the more advanced the level oftechnology used in the reconstruction and display,the greater the belief in its authenticity (Emele,2000).This mechanism is also reinforced when re-constructions are displayed in the context of mu-seums with the inherent authority these institutionsconfer to the works.



Virtual archaeology has been accused of being more'an artistic task than an inferential process' (Forte,2000).Uncritical acceptance of the product has ledto a point where 'fundamental questions relating toissues such as what we actually mean by virtual re-ality, and what our expensively assembled modelstruly represent have been left largely unexplored'(Emele, 2000).

As Barceló (2000) notes:

...future advancement of virtual reality tech-niques in archaeology should not be re-stricted to "presentation" techniques, but toexplanatory tools...

...VR techniques [are] not only for descrip-tion, but for expressing all the explanatoryprocess. An explanation can be presentedas a visual model, which is as a virtual dy-namic environment, where the user asksquestions in the same way a scientist use atheory to understand the empirical world(Barceldo, 2000).

The perspectives introduced above were to act asboth visionary statements and cautionary notes tothe visualisation project. Against the potential of adigitally reconstructed archaeological model, theproject was constrained in a number of ways withregard to the final product. Resource restrictions(time and computing power) meant that it was nec-essary to fix an approximate date in the history ofOlympia rather than use the model to examine thechanges at Olympia over time. Interactivity was alsorestricted to force the users to complete a tour ofthe site (with choices along the way) in 20 minutes,rather than allowing free interaction. Juxtapositionof the reconstruction and the archaeological sitewas introduced by the use of panoramas from theOlympia site as it appeared in March 2000.

The paper will now introduce the Exhibition thatgave the opportunity for the creation of the worksand the impetus for the Web.

The Exhibition 1000 years...

In 2000 Sydney hosted the 25th Olympiad and wasalso the recipient of the most significant selectionof antiquities from ancient Greece ever seen in

Australia. Sydney's Powerhouse Museum featured54 ancient objects most of them usually on per-manent display in their museums of origin in Greece,and many being famous icons of Greek art.The ex-hibition, 1000 Years of the Olympic Games: treasuresof ancient Greece (July November 2000) capturedsome of the magic that constituted the essence ofOlympia, and the vibrant ancient Greek society inwhich it thrived.The majority of objects coming fromGreece to the Powerhouse Museum, including sculp-ture, grave markers, votive offerings, ceramic ves-sels and sporting equipment, date from around 700to 200 BC. This slice of time is one of the mostenergetic periods of human endeavour ever re-corded. Subdivided chronologically for our conve-nience into the Archaic (about 660-480 BC), Clas-sical (480-323 BC) and Hellenistic (323-27 BC) Pe-riods, it was during this time that the essentials ofwestern life were defined including philosophy,poetry, drama, architecture, art, and sport(Donnelley, 2000).

Due to a significant sponsorship offer by the IntelCorporation, the Exhibition had two interactivecomponents the digital reconstruction of Olym-pia and the 3D Zeus incorporated into it. Mostimportantly (according to sponsors wishes), bothof these works would be translated for the Internetin a state-of-the-art Web. Data acquisition was amajor task in creating the initial exhibits.

The site of Olympia-data acquisition

The site of Olympia is in fertile countrysidesqueezed by the steep and tree covered MountKronos into the elbow of two rivers the Alpheiosand Kladeos. Archaeology and historical recordsshow that little has changed at this site over thepast few millennia. In the prehistoric period, its ver-dant topography inspired the worship of nature godsand set it on a course for greatness as one of theglories of ancient Greece. This isolated glade wasto spend 1000 years as one of the most importantreligious sanctuaries in Greece, with its OlympicGames a fundamental component of worship to thesupreme deity in the ancient Greek pantheon:Zeus(Donnelley, 2000).

A team comprised of a curator, archaeologist, sur-veyors and photographers traveled from Australiato Greece early this year to capture the data sets

160@Archives & Museum Informatics, 2001 I 65


in high resolution using laser scanning and digitalphotography. Most of the archaeological remains atOlympia are scattered across the site, the result oftwo earthquakes of the 6th Century AD and nu-merous floods.Those foundations that survive dateto different periods,from theArchaic,Classical,Hel-lenistic and Roman times.

The excavations at Olympia were begun in May I 829,by French archaeologists.The initial finds (metopesfrom the opisthodomus and parts of the metopesfrom the pronaos of theTemple of Zeus) were trans-ferred to the Louvre where they are still being ex-hibited.When the Greek government was informedof the looting of artefacts, the excavation wasstopped.

Excavations were started again 45 years later byGerman archaeologists. Research continues todayby the German Institute of Archaeology in Athensand the Ephorate of Antiquities in Olympia. TheMuseum of Olympia associated with the archaeo-logical site houses many of the sculptural and ob-ject material that has been recovered from the site.Supplementary material on the archaeological siteand its associated Museum can be found on the

-1

Hellenic Ministry of Culture's Web (http://www.culture.gr/).

The bibliography on Olympia runs to hundreds ofitems, although those dealing strictly with the ar-chaeology of the site can be counted in the meredozens.The most important sets of documents forthe reconstructions were some of the oldest. It al-most goes without saying that any work on Olym-pia would be nearly impossible without Pausanias'Guide to Greece (2nd AD).

The five text volumes of the earliest German se-ries, Olympia. Die Ergebnisse der vom Deutschen Reich

veranstalteten Ausgrabung (Adler et. al. (1892- I 897);referred to hence as the Adler and Curtius publica-tions) contain incredibly detailed descriptions ofbuildings and objects, complemented by the hun-dreds of plans, sections, drawings and reconstruc-tions in the accompanying folio volumes.Subsequentimportant updates have been published byAshmole&Yalouris (1967); Grunauer (1971) and (1981); Miller(1971); Mallwitz (1972); Herrmann (1972); Koenigs(1984) (Da Costa et al, 2000).

vz 17KIRCHE

ZEUS-TEMPEL

1

4IW

kir10_11.

ANTIKER BAU

0 t4 --ACP 'A

EH rt.

I

117117 MAT.

cccc

0 0 Eb-MALIHR

-

000 a our) Li&

.Fig 1: Detail from an excavation map of the site of Olympio,Adler & Curtius 1896.

161



Archaeological and historicalconsiderations

The project team decided to attempt to reconstructOlympia, as it had been around 200 BC. Of course,it is actually impossible to pick a single year to re-construct an ancient site, given the relative coarse-ness of archaeological chronology. At a site likeOlympia, used without interruption for over a thou-sand years, almost certainly with frequent refurbish-ment of quite old buildings, it is really only possibleto establish the rough date a building was erected.Whether a building was undergoing renovation in aparticular year cannot be stated.The length of timeit took to complete buildings is also another issue.

The choice of 200 BC was therefore made for mainlypractical reasons:the later the date in the site's his-tory, the more buildings there would be, and themore interesting a virtual tour for modern visitorswould be and would more closely correspond tothe extant ruins; it was an Olympic year; it was justbefore the Roman annexation of Greece began.Weacknowledge the date was stretched slightly in or-der to incorporate the entire gymnasium and thekrypte entrance to the stadium.

Individual buildings

The Adler and Curtius publications provided de-tailed ground plans of most buildings. As well, theelevations were usually calculated, and the extantdecorative elements, mainly terracotta simae andakroteria, were associated with each building. Major

Fig 2.Work in progress, rendering thePryaneion. @ Powerhouse Museum.

additions or changes to some buildings were takenfrom the later publications.The drawings were useddirectly by the modellers to recreate each building.An attempt was made to incorporate as much smalldetail as possible given the limits of time and com-puting power (both in rendering and delivering theanimated tour). For example, the Adler and Curtiusvolumes contained the differing capitals and columnsof the Heraion, and these were included in themodel.

Inevitably, there was missing information, and thelimited timeframe meant that not every known de-tail could be incorporated.The application of colour,which is now standard in our concept of ancientGreek architecture and art, is still difficult becauseof the lack of evidence. One particular difficulty inusing the early German work is that the colour platepublished (Volume II, plate CX11) does not corre-spond with the text description of colours or place-ment, even though it is supposed to be the tem-plate for the painting of Doric buildings at Olympia.This is partly a factor of colour printing techniquesin 1896 and the fact that we had the 1966 reprint,and also due to the difficulty of really knowing whatcolours such as "mild whitish cobalt blue" or "astrong but at the same time transparent blue" ac-tually were. However, we followed the Adler andCurtius publications as closely as possible in termsof applying colour to the buildings, and the evidenceof the terracotta roof decorations excavated at thesite was particularly helpful in this regard.

Buildings which must have had pedimental sculp-ture or for which some fragments of sculpture re-main, such as the Metroon, were given pale bluepediments, rather than attempting a poor recon-struction. Doors and lattice screens were based onthose depicted on black and red figure vases.

Probably the least satisfactory outcome, given thetime restrictions on the project, related to the ad-dition of statuary to the Altis area of Olympia. Lit-erally hundreds of dedicatory statues and objectswould have been placed into the sacred area, butwe were able to reconstruct only a handful, mod-elled on excavated statue bases and the descrip-tions in Pausanias.

Our most ambitious reconstruction was the inte-rior of the Temple of Zeus.This was based only on



Fig 3. Metope X from the Temple of Zeus.The Apples of the Hesperides. Archaeological

Museum at Olympia. Photograph PowerhouseMuseum. © Hellenic Ministry of Culture.

Pausanias' descriptions as no evidence remains.lt isthought that some coins and Christian icons re-flect the seated Zeus figure, once one of the won-ders of the ancient world, but they have alreadypast through the filter of later cultural biases.

A further discussion on the basis for reconstruc-tion is available from the project Web (http://www.phm.gov.au /greek/4/pdf/archaeological_basis.pdf).

The digital model

The topography of the site and the locations of keystructures were surveyed by members of the teamso that the present surface (which approximatesthe original ground level of the site) could be usedin the visualisations.This was performed using sur-veying equipment loaned by the National TechnicalUniversity in Athens.A digital elevation model wasprepared from this data, and combined with a de-tailed terrain model of Kronos Hill obtained fromlarge-scale topographic maps. In addition, the sur-rounding landscape was also modelled from smallerscale topographic series maps.As a result, the land-scape of ancient Olympia used in the visualisationsis a very close approximation to the terrain of theperiod.

Fig 4. Colour interpretation Metope X fromthe Temple of Zeus,The Apples of the

Hesperides. Drawing David Loong.© Powerhouse Museum.

Every graphic element was created for the project.Existing digital Greek architectural elements wouldnot have allowed us to maintain a sufficiently highlevel of contextual accuracy. For example, theTempleof Hera (where the Olympic torch is ignited) hasevery column different as it was constructed andreconstructed over a long period of time. Standardcommercial digital library elements did not easilymodify; in all cases it was easier to create new ele-ments.The human figures and statues were mod-elled originally in PoserV4, and modified in 3D Max.

Fig 5. Reconstruction of the Temple of Zeus. ©Powerhouse Museum.

163I 68 © Archives & Museum Informatics, 2001


Fig. 6. South-south-east aerial view of the reconstruction. © Powerhouse Museum.

Display

The delivery of theVR reconstruction posed manychallenges.The intention was to provide a 3D 'ex-perience' of Olympia to an audience of up to 20people at a time using affordable technology.A rear-projection polarised projection system with inex-pensive plastic glasses was developed.Two JVC DLA-CI5 rear projection capable projectors were usedto overlap the left and right channel images onto a3m wide screen in a 5m square room.

Future

The virtual reconstruction of Olympia was intendedfor the general public (with limited interactivity)rather than for use as a teaching tool. However, theunderlying dataset of the model does allow for sci-entific use. Some areas for investigation include

Modelling sun in relation to temples

Modelling the earthquake effect on the Temple ofZeus

Model allows for further research about the useof the sanctuary and liturgical practice

Choices for colour applications invite responsesfrom the academic community.

Statue of Zeus

One of the most significant sculptures in the Na-tionalArchaeological Museum in Athens is the statueof Zeus from Artemision (Artemesium), also con-sidered to be perhaps a statue of Poseidon. Thisbronze sculpture is slightly larger than life size, andwas found in 1926 in the sea off Cape Artemision.It is one of the few surviving examples of Early Clas-sical statuary. The laser scan of the Zeus allowedthis object to travel, in a virtual sense, from Athensto Sydney.

A laser scan

A Modelmaker Laser Scanner mounted on a Faro3D Coordinate MeasurementArm was used to cre-ate high-resolution point clouds of the surface ofthe statue.A purpose-built scaffold was fabricatedin Athens to enable elevation of the scanner to thetop of the statue while ensuring a stable fixture.The entire sculpture was digitised over several day-time sessions in multiple parts.There were consid-erable restrictions on this part of the project,whichincluded no method of re-orienting the scanningsystem datum and a very short working period inthe Museum.

The data resulting from the laser scanning opera-tion consisted of very dense clouds of points rep-



roART.

Fig 7. Laser scanning the statue of Zeus at theNational Archaeological Museum in Athens,

March 2000. Photo University of Melbourne.@ Powerhouse Museum.

resenting the surface of the statue with a coordi-nate accuracy of around 0.2mm. This presentedmany difficulties as the data sets were huge andneeded substantial filtering before they were use-ful.The decimation process reduced the density ofthe data points without reducing the high level ofdetail in the model. The process used a variety ofsoftware packages,finally resulting in 3D Studio Maxmodel of the statue.This contained over 2 millionpolygons.

Presentation

The Zeus model was presented in a separate vir-tual reality booth at the Exhibition. It was displayedusing shuttered glasses technology driven by pro-prietary software running on Intergraph hardware.

A limited amount of interactivity was providedwhere users could rotate the model about its ver-tical axis, and zoom in and out.A higher degree ofinteractivity was not appropriate given the respectaccorded the actual object itself, and therefore, itsfacsimile! (Da Costa et al, 2000).

The aforementioned works Olympia and the Zeuswere both then translated into the products for a

state-of-the-artWeb.

Web site (http://www.phm.gov.au/ancient_greek_olympics)

Despite the great opportunity to produce theseaforementioned interactive Exhibition materials, the

Intel

Fig 8. Screen grab from the homepage of theweb, 1000 years of the Olympics Games:

treasures of ancient Greece, 2000. (hap://www.phm.gov.aulancient_greek_olympics).

Powerhouse Museum.

sponsors of the project (Intel Corporation) and theMuseums' main interest involved the production ofa state-of-the-art Web. This Web was intended toaugment the Exhibition, and to demonstrate theviability of the Internet to supplement and extendmaterials offered locally in the Museum.The abilityto give access to this material to wide audiencesand the fact that it will eventually outlive the Exhi-bition by two years is also the advantage of thismedium.

To meet the sponsors' expectations, this Web wasconstructed to demonstrate the use of state-of-the-art technologies for the delivery of rich cul-tural content. The web was designed with a two-week exclusive period for the Intel Web OutfitterService in the United States,where subscribers havea minimum of a PIII chip and are regularly updatedwith the latest plugins.From the Museum's perspec-tive the opportunity to design a product with a highdegree of sophistication outweighed the concernsover download times and minimum browser speci-fications.The Web was ultimately intended for aca-demic, school and researcher markets.

Architecture, design and programming

The web was scripted as active server packages.Aflash based navigation system was used.There washigh use of 360° panoramas constructed with LivePicture from Reality Studio. Apart from the stan-dard plugins such as AbodeAcrobat Reader,the mainplugin is the Zoom Viewer by MGISoft.The latterhas two main utilities: it allows the high-resolutionimages to be stored on the server and streamed to

165170 ©Archives& Museum Informatics, 2001 BESTCOPYAVAILABLE

NiVItAtiff 114YOUIIA

ROOM 3 GYM NASI UM/PALAESTRA

1 aomtl

Heed at t, Youthabout 480 - 970 BC

<row, = fru, rand , r

I Ect

ROOM 3 GYM NASIUM/PALAESTRAHead eta Youth

611, el it ,1 0 0

frur,110

Use the toolbar to zoom ,n and out.0 re en.< Miro rtv of ...Nom

Fig 9. Screen grab from the virtual Exhibition of the web, 1000 years of the Olympics Games: treasuresof ancient Greece, 2000. (http://www.phm.gov.aulancient_greek_olympics).© Powerhouse Museum.

the user on demand:and it prevents the users fromsaving the materials to their hard drives. This lastfeature is of extreme importance to the GreekHellenic Ministry of Culture and the Museum whowish to protect copyright. As noted, none of theExhibition artefacts (also on display on the Web)had travelled outside Greece before. With a his-tory of contentious international relations overGreek antiquities, it was very important to all thoseinvolved that every measure be taken to preventunwanted use of the digital materials. VRML wasnot considered to be a robust or user-friendly en-vironment for translation of the Olympia recon-struction. However visualisation in 3D is certainlythe way forward for the display of these materialsin the future, and major developments continue inthis field.

The web is constructed in six sections.These are:

I.The Exhibition: a series of zoomable panora-mas of the Exhibition at the Powerhouse Museumaugmented by 3D object movies and high-resolu-don zoomable images of objects, and supplementedwith object information as text.

2.Virtual Olympia: this section contains a richcollection of materials separated into 4 sub-sec-tions, including.

Archaeological and geographic maps of Olympia

A panoramic tour of the digital reconstruction ofOlympia presented as 360° zoomable panoramas(viewable at full screen) with audio files.The op-tion to download individual buildings to comple-ment scholarly research is also available. Essaysavailable as PDF files

29 zoomable 360° panoramas of the archaeologi-cal site at Olympia taken in March 2000 (viewableat full screen). Essays available as PDF files

The statues from the Temple of Zeus shown as a360° panorama of a room at the Museum of Olym-pia, and a series of zoomable object movies.Themetopes are further supplemented by recon-structed line drawings and colour interpretationsusing archaeological and historical records assources for reconstruction.

3.3-D Zeus:a web version of the laser scan of theZeus of Artemision, with the option to view as a3D anaglyph through downloadable anaglyph glasses.

Resources:14 researched PDF articlesEducation programmesGlossary of terms linked to the PDF articlesA nodal genealogy of the godsSupplementary programme of events to comple-ment the Exhibition at the Powerhouse MuseumAll audio files are listed.



BESTCOPYAVAILABLE


"TT-;(71,1k1fRc,-

Fig 10. Screen gab from the Temple of Zeus statues as displayed in the Museum of Olympia, 1000years of the Olympics Games:treasures of ancient Greece web, 2000. (http:Ilwww.phm.gov.aul

ancient_greek_olympics).© Powerhouse Museum.

5. Image archive:because of the extensive visualcomponents in the web, this section offers a quickentry to all the individual panoramas, zoomableimages, object movies and jpegs. Selected buildingsfrom the digital reconstruction of Olympia are avail-able here, for research purposes, to download andprint

6.Timeline:a timeline of events at Olympia over a1000-year-period is organised into three catego-ries:

Events specific to the Games at Olympia (eventsadded after 200 BC not included)

Events relevant to Olympia itself (buildingprogramme until end of Hellenistic period only)

Events, political, around the Mediterranean worldafter 200 BC only events directly relevant to

Greece or Olympia included: politics, arts & sci-ences.

The Web also contains extensive project informa-don, help information, plugins and downloads pages.

Promotion and marketing

The Web was launched just prior to the OlympicGames. It benefited from a high degree ofmarketing through its association with the Intel

172

Web Outfitter Services and was featured on thehomepage for the Intel.com Web for a short timein the United States and Australia.The site wastranslated into three other languages: Korean,Chinese and Japanese. In conjunction withchildren, the Premier of the State for New SouthWales launched it.There was a high degree ofmedia coverage derived from the Web launch andalso from the Exhibition itself.

Analysis

Statistical data has so far been analysed for the four-month period of the Exhibition from the begin-ning of August, 2000 until the end of November,2000.The author acknowledges the restricted na-ture of basic statistics to determine true use of theWeb. However,general statistics are presented hereto demonstrate several observable trends.

The average number of hits per day over this pe-riod totalled 18,500. Other broad general-level sta-tistics include these: the average number of visitorsessions of 52, 582, and the average visitor lengthof 12 minutes.The majority of visitors were fromoutsideAustralia (international at 55% and unknownat 29%).

A more thorough examination of the period showsa predictable variation in the data display.The mostactive day of use was during the Olympic period


I DETAILS

3-0 ZEUS

34 MOS

(<1 3.1) ZEDS A MIXON VERSIONThs mod41 nftki star. of Zeusmob". ;10 olassel. Ciick ,01 thelink bete. to ornload th.m.

0 3-0 CLASSES

El 001,MMIC IESI

°ACK

tmli.floitVIRT IltAL *MEM

.3azEusAlsowicEsb

04.os AnNt

0

it011111,

Fig 1 1. Screen grab from the 3D Zeus, 1000 years of the Olympics Games:treasures of ancient Greeceweb, 2000. (http://www.phm.gov.au/ancient_greek_olympics).@ Powerhouse Museum.

and on September 16,2000 with 92,352 hits.Week-days rather than weekends showed a larger num-ber of hits although it was after-hours that num-bers were marginally higher. This contradicts ourinitial expectations that the high bandwidth requiredto access the site would preclude extensive homeuse.

What is potentially of more interest is the length ofstay by each visitor: 45% of visitors accessed justthe first two pages of the Web.There was a corre-sponding drop off as the number of pages increasesd.However, a relatively high proportion (8% of thevisitors) was staying over 20 minutes to access 43%of the total page views.This last statistic would sug-gest that those visitors who were interested in thesite found rich and rewarding content.At one pointin the 4-month period over 18% of the visitors werestaying more than 20 minutes. In Web lingo this hasbeen referred to as Web "stick-ability" (and hope-fully not a measure of slow content download!).

The Web will be hosted on the Museum server fortwo years.As is so often the case, the overall devel-opment time for the entire project was just 4months including the Exhibition 3D materials.Thisprevented extensive usability testing before launch,and represents a classic scenario in museumprojects. During the upcoming period it will be pos-

168

sible to further analyse the Web use in schools andtertiary institutions. There have been numerouscomplimentary emails from such organisationsabout the quality of the scholastic research, althoughthere have also been requests for a CD ROM ver-sion to allay the heavy bandwidth requirements.

Awards

The overall project has won many awards in theshort time since its launch.The most important ofthese include aVirtual System and Multimedia awardfor the BestVirtual World Heritage reconstruction;the Best of the Best web for the Australian Interac-tive Multimedia Industry Awards; the Best InterfaceDesign at AIMIA; Hitwise Top 10 for Education andResearch 2000; and the Web was a finalist in theBritish Association for Film and Television Interac-tive EntertainmentAwards 2000 (short-listed to oneof three along with the multi-million dollar 'Walk-ing with Dinosaurs'). The judges described it

A fascinating and highly interactive site, thisis elegantly designed and technically impres-sive and brings you closer to OlympianGreece with effortless navigation and appro-priate use of all media, including superb pan-oramic photography and walk through 3Darchitecture. It is a state-of-the-art web with


an excellent interface and very high-qualitycontent.A technical tour de force deployinga broad range of impressive plig-ins to goodeffect (BAFTA awards catalog, 2000).

Conclusion

Any project of this nature is concerned with howto balance 'realism' with 'reality', and the decisionsmade in the creation of digital reconstruction werefurther constrained by the goals of the project, theintended audience, the desired product, the qualityof archaeological information,and technological ca-pabilities. In the construction of the Web, a furthercaveat on design decisions included the use of state-of-the-art technologies, which must demonstrateimpressive content development and display for asophisticated Web user demographic. This ques-tioned the Museum's role as a provider of 'accessfor all', but at the same allowed for the Museum topush limits in Internet design and technology usingrich cultural content.

The Web and its contents are being re-purposedfor a DVD ROM which will be distributed free ofcharge to all schools in the State of New SouthWales,Australia, and in limited numbers for othermarkets.The 3D version of the digital reconstruc-tion of Olympia has been sold to three interna-tional Olympic museums, and is to be used in uni-versity teaching in some form.

The project is an example of a digital reconstruc-tion project that has been repurposed for a highlysuccessful Web. It also demonstrates that rigorousscholarship is a worthwhile investment and cancommand a significant audience. The repurposingof the underlying content has enabled the works tobe developed for different platforms and deliverymechanisms. The challenge lies in providing userswith the ability to alter input data to test their ownhypotheses.The development of VRML worlds onthe Internet (ones that are robust in implementa-tion) will hold much value for museum Web devel-opments.

References

1000 years of the Olympic Games: treasures of ancientGreece, web (2000). Last accessed 15 January 2001http:/Iwww.phm.gov.au/ancient.greek_olympics

F, Curtius, E., Dörpfeld,W, Graef, P., Partsch, J. &Weil, R. (1897). Olympia 1:Topographie undGeschichte, 1897/1966,Verlag Adolf M. Hakkert,Amsterdam.

Adler, R., Borrmann, R., Därpfeld,W, Graeber, F. &Graef, P. (1892). Olympia 11: Die Baudenkmdler vonOlympia, 1892/1966,Verlag Adolf M. Hakkert,Amsterdam.

Ashmole B.,Yalouris N. & Frantz, A. (1967). Olympia:TheSculptures of the Temple of Zeus. Phaidon Press,London.

Barcel&A. (2000).Visualizing what might be: anintroduction to virtual reality techniques inarchaeology. Virtual reality in Archaeology. BAR S843.CD ROM, Archaeopress. Oxford.

Dittenberger,W. & Purgold, K. (1896). Olympia V: Dielnschriften von Olympia, 1896/1966,Verlag AdolfM. Hakkert,Amsterdam.

Colon na, J.F. (1994). Images du Virtue!. Addison-Wesley,France.As discussed by Barcel&A. (2000).'Visualizing what might be: an introduction tovirtual reality techniques in archaeology'. Virtualreality in Archaeology. BAR S843. pp 29, CD ROM,Archaeopress. Oxford.

Da Costa, K., Kenderdine, S., Ogleby, C. & Ristevski, J.(2000).The digital reconstruction of Olympia, 3DZeus and website. 1000 years of the OlympicGames: treasures from ancient Greece web. Lastaccessed 15 January 2001 (http://203.10.106.180/greek/4/pdf/Grecian2000weba.pdf)

Donnelly, R (2000). Comments from the curator. 1000years of the Olympic Games: treasures from ancientGreece. Powerhouse Museum, Sydney.

Eleme, M. (2000).Archaeological simulation betweenlinear media and a virtual museum. EuropeanInstitute of Cinema Karlsruhe web. Last accessed31 August 2000 (http://eikk.hfg-karlsruhe.de/Eng/simulation.htm) No longer available.



Forte, M. (2000).About virtual archaeology: disorder,interaction and virtuality. Virtual reality in Archaeol-ogy. BAR S843. CD ROM.Archaeopress. Oxford.

Furtwangler,A. ( I 897). Olympia IV: Die Bronzen und dieObrigen kleineren Funde von Olympia, 1897/1966,Verlag Adolf M. Hakkert,Amsterdam.

Greek Ministry of Culture web. Hellenic Culture. Lastaccessed 15 January 2001 (http://www.culture.gr/)

Grunauer, R (1971). Der Zeustempel in OlympiaNeue Aspekte, Bonner Jahrbuch, 171, pp. 114-131.

Grunauer, R (1981). Zur Ostansicht des Zeus Tempels.X. Olympische Bericht, pp. 256-301.

Herrmann, H.V. (1972). Olympia: Heiligtum undWettkampfttatte, Hirmer Verlag, Munich.

Koenigs W (1984). Olympische Forschungen XIV: DieEchohalle,Verlag Walter de Gruyter & Co/DA1,Berlin.

Mallwitz,A. (1972). Olympia und seine Bauten, Prestel-Verlag, Munich.

Miller, S. (1971).The Prytaneion at Olympia, Ath. Mitt(AM), 86, pp. 79-107.

Pausanias. (2"° C AD). Guide to Greece vol 2: SouthernGreece (trans. R Levi). Penguin Classics, London,1979.

Treu, G. (1897). Olympia III:Die Bildwerke von Olympia inStein und Ton, 1897/1966,Verlag Adolf M. Hakkert,Amsterdam.

[Sections of this paper have been included in the projectinformation available on t he web(http://www.phm,gov.au/greek/4/pdf/Grecian2000weba.pdf) and(http://wvmphm.gov.au /greek/4/pdf/archaeological_basis.pdf)].With regard to these documents the author would liketo acknowledge co-authors Kate Da Costa, Cliff Oglebyand John Ristevski].

170

Acknowledgements

This project was made possible through the generousprovision of the Hellenic Ministry of Culture, Athens(http://www.culture.gE).A number of people in the Min-istry worked hard on our behalf to make the data collec-tion process possible. Sponsorship to complete the workwas generously provided by the Intel Corporation (http://www.intel.com.au/). MG1 Software Ltd (http://www.mgisoft.com) kindly supported the license for theZoom Server software.

Thanks to co-developers Cliff Ogleby and John Ristevskifrom the Geomatics Department at Melbourne Univer-sity for their work on the digital reconstruction of Olym-pia and the Zeus; to Vu Nguyen and Asher Graham ofOsmosis for their excellent rendering and animation skills;to Kate Da Costa, archaeological advisor on the project,who was an invaluable and seriously hardworking talent,without whom we would have retired to Disney....

The Web was conceived by the Powerhouse Museum.The photographic work at the archaeological site wasundertaken by Peter Murphy, Sydney. It is his genius alsothat enabled much of the data that is displayed on theWeb.The overlays for the metopes were draw by a me-ticulous David Loong.

The sonorous tones of David Rutledge were recordedfor the tour and used for a number of audio files on theWeb site.

Thanks also go to Lefki Pavlidis for her assistance withthe data collection and laser scanning in Greece; and toSebastain Chan and Jessie for the installation tweeks thatneeded to be undertaken for the virtual Olympia model.

Our time in Athens was also eased by the welcome ofthe Australian Ambassador in Athens, Ross Burns, andconsul Deborah George. We are also indebted to JanCasson-Medhurst at the Australian Archaeological Insti-tute at Athens.Thanks to Judith Swaddling at the BritishMuseum for her time in reference to the scale model ofOlympia housed at the Museum.

© Archives & Museum Informatics, 2001 I 75

Engaging Museums, Content Specialists, Educators, andInformation Specialists:A Model and Examples

C. Olivia Frost, University of Michigan, USA

AbstractThis paper describes a model for the use of the Web to engage communities in better understanding and appreciationof cultures through partnership among a knowledge organization. K-I2 teachers and learners, museums, and regionaland virtual communities. Museums and other content providers offer artefacts for object-based learning. Contentspecialists contribute specialized content expertise and evaluate resources. Education specialists develop, evaluate,and use materials for instruction. Information specialists identify, evaluate, organize, and promote the use of web-based and other information resources. They also develop tools to capture and display content, to engage theteachers and learners in dialogue, and to reach out and extend the content and resources to both local and virtualcommunities. Communication tools enable members of the local as well as virtual communities to provide reflectionsand engage in dialogue, as well as to contribute their own content. Two recent projects developed from the modelare described, in addition to four previous projects developed by the project team, that exemplify earlier applicationsof the model. Lessons learned from the project work are identified, and the paper discusses ways in which this typeof project work reflects new roles for the professionals participating in the projects, as well as changes in the missionof their organizations and professions.

Introduction

Successful partnerships among different institutionsor disciplines are an attractive way of providingcomplementary expertise in addressing needs ofcommon concern and tasks that require a range ofknowledge areas. As professions evolve in theirmissions and their roles expand and intersect withthose of other disciplines, the potential for effec-tive synergistic partnerships increases. Profession-als in museums, libraries, the performing arts, andK-12 education have all undergone profound missonchanges which have resulted in greater emphasison outreach and new connections with membersof other professional communities.At the Univer-sity of Michigan, some interesting and fruitful part-nerships have engaged members of the arts com-munity, museum and library specialists,faculty schol-ars, K-I2 educators, and students from the Schoolof Information. From these has emerged a partner-ship model utilizing the information and technol-ogy skills of librarians and information profession-als, the content knowledge of curators and schol-ars, and the expertise of K-I2 teachers and otheryouth educators.The partnerships developed havebroadened the reach of arts and cultural heritagecontent by taking advantage of the unique poten-tial of information technology to disseminate con-tent, knowledge and ideas to a wide array of audi-ences, to provide context to increase the under-standing of cultural content and to provide a meansof interactive and cross-cultural exchange of ideasand new creative work.

This paper describes partnerships in two recentprojects which build on the model, as well as les-sons learned in these and related projects whichexemplify earlier applications of the model. Finally,some observations are offered on the ways in whichthis type of project work offers the opportunity forprofessionals to assume some of the new roles whichare emerging as a result of the evolving missions oftheir organizations.

The Projects

The two projects which illustrate the partnershipmodel described in this paper are ones in which afaculty member and graduate students at the Uni-versity of Michigan School of Information play orga-nizing roles as part of the CHICO project (CulturalHeritage Initiative for Community Outreach (http://www.si.umich.edu/CHICO/).CHICO has as its mis-sion to broaden the reach of cultural heritage re-sources through the innovative use of informationtechnology and through partnerships with muse-ums, schools, libraries, content specialists, and com-munity arts programs. CHICO collaborated withcampus partners to provide a digital showcase forevents and exhibits in the arts.The resources wereaugmented with contextual enrichment provided bycontent specialists.The site was targeted to generalaudiences, with a focus on K-I2 users. Delivery in-cluded both Web-based resources as well as public



BESTCOPYAVAILABLE

Frost, Engaging Museums and Outside Specialists

events and activities in the classroom with the part-nership of K-I2 educators.

Royal Shakespeare Company Project (RSC)http://www.slumich.edulchico/rsc/

A landmark performance event at the University ofMichigan is the appearance in March 2001 of TheRoyal Shakespeare Company from London for athree-week residency on campus and in the sur-rounding region.The Company's performance of thecomplete cycle of Shakespeare's eight historical plays

and the residency of the actors, stage crew, andeducation department engage University scholarsand students, regional community groups , K- I 2schools, theaters, and local museums and libraries.Educational and outreach activities developed forthe event help K- I 2 teachers and learners learnnot only about the plays themselves, but also aboutthe relevance the themes in the plays have in theireveryday lives. School of Information graduate stu-dents in the CHICO project have a role in develop-ing web resources to help increase appreciation andunderstanding of the plays.

Food/Animals in Ancient Cultures Projecthttp://www.slumich.ed u/CH I CO/AncientFood/ andhttp://www.si.umich.edu/CHICO/Animals/

A University of Michigan professor who has donepioneering work in providing a multidisciplinaryapproach to the study of classics and archaeologyworked with CHICO students to develop a webresource to help K-I2 students understand the roleof food in ancient culture through examination ofartefacts in a local museum. SI graduate studentshad a role in developing resources and tools to builda virtual and local community of scholars, K- I 2teachers and learners, museum experts, and otherinterested people in the community with the goalof understanding how people lived in ancient cul-tures and how these cultures have relevance intoday's times.

Collaboration roles

With a collaboration among museums, content spe-cialists, K-12 and outreach educators, and informa-tion specialists, there are roles for partners provid-ing service to the end-user, as well as roles for end-users engaging with the content. Individual part-ners can play different roles according to their de-

gree of expertise, level of commitment, and abilityto participate in terms of time and resources. Forus, the role distribution was as follows:

Content provision

creation of content (curators, scholars, informa-tion specialists)

resource discovery of information on related top-ics (information specialists)

interpretation of content (curators, scholars, in-formation specialists)

Content organization and distribution

selection and organization of content (informa-tion specialists)

mounting and display of content in digital form(information specialists)

provision of contextual information and back-ground for learning purposes (information spe-cialists, curators, scholars)

site design and implementation (information spe-cialists)

Organization of project activities

liaison (information specialists)

identification of relevant content specialists (in-formation specialists)

recruitment of content specialists and partners(information specialists)

Activity creation

providing web-based activities to engage in andconnect with the content (information special-ists)

providing on-site activities which link to the physi-cal and virtual exhibits (museums, performing artsagency)

coordinating activities among the various part-ners (information specialists)

Audience

For the end-users and audiences, roles include:

engagement with peer learners

creation of artefacts and content which build uponthe collection

.172I 78 © Archives & Museum Informatics, 2001


feedback on the usability and usefulness of thesite

Engagement with Content Specialists

In creating web-based materials for both projects,the team's goals were to create interest in and un-derstanding of the topic; augment the performanceor museum visit experience; provide a means forfurther study of the topic; provide mechanisms toengage in dialogue; and provide a means for partici-patory experiences.

The broad themes for structuring the sites includedRelevance and Connections; Context and HistoryBackground; and Behind the Scenes (stagecraft, ormounting a museum exhibit). In both projects,CHICO team members were able to take advan-tage of local content specialists,as well as resourcesand events, all which contributee to content of theweb-sites and built upon the broad themes.

Content specialists in the RoyalShakespeare Company project

The visit of the Royal Shakespeare Company, orga-nized by the University Musical Society, provided aready-made infrastructure, including a collocationof content specialists and a wide array of plannededucational outreach activities, including public lec-tures, museum exhibits, seminars, university classsessions open to the public, interviews, workshops,and in-school visits. The audiences for the eventsinclude the general public, university students, K-I2teachers and learners, and theater groups through-out the area of Southeast Michigan and centralCanada.

In designing the site to establish relevance and con-nection, the CHICO students envisioned a resourcecontaining information, enriched by images, maps,and links,to make clear the relevance of these playsto a contemporary audience. The site would alsocontain material to encourage people to make theirown connections to the plays.An example is show-casing adaptations of Shakespeare's plays that havebeen set in modern times yet maintain the integrityof the text while framing the play in a contempo-rary setting. The site would also provide contextand history background for the plays themselves,aswell as provide information, images and links to elu-cidate the history in the plays; for example, a his-torical timeline and comparisons with instances of

civil strife in the United States. Behind the scenesinformation would provide background on the artof the play in general,and specific information aboutthe art of staging these plays in particular.

The CHICO student team members were inter-ested in including content about the political cli-mate of the time when the plays were written andshowing how that related to the histories portrayedin plays to be staged in Ann Arbor by the RSC.Theteam first attempted to do their own research andcompile materials from generalist sources to bepresented from a generalist's perspective. Uponencouragement, they looked for expert help andwere successful in recruiting assistance from spe-cialists to write key pieces.A widely-renowned En-glish professor involved in the RSC residency pro-gram will be giving a series of public lectures andagreed to contribute content on the play's histori-cal and literary significance.This professor had justspent a week in England sitting in on rehearsals ofthe plays and talking with the plays' director.A doc-toral student in theater, who coordinates theShakespeare reading group on campus, also agreedto contribute. The content specialists were alsoasked to write short essays about Shakespeare's life,the history plays, and the relevance of the historyplays for a contemporary audience.

The Royal Shakespeare Company submitted imagesof costume designs and other materials; includedwas a file of photographs depicting rehearsal shotsand images of historical engravings. The studentswere also successful in garnering permission to in-clude ten essays requisitioned by the UniversityMusical Society for the event. In conjunction withthe project, the Curator of Special Collections atthe Graduate Library prepared an exhibit on cos-tumes used in Shakespearean plays, to be on displaybefore and during the time that the RSC is in resi-dence.The exhibit contains costumes, prints, pho-tographs, rare books and drawings illustrating theclothing crafted for performances of Shakespeare'shistorical dramas from the mid-18th century to thepresent.With the assistance of the curator, CHICOteam members selected and photographed materi-als from the exhibit to include on the website.TheCurator also gave the students guidance on how todo research for other background content on theproject. In addition to providing their own originalcontent, the content specialists reviewed contentdeveloped by the students and evaluated parts of



the website to make sure the historical content ofthe website is accurate.

Content specialists in the Ancient Food/Animals project

Through a university committee assignment, theCHICO director made the acquaintance of a fac-uky colleague in archaeology and classical studies,and the two exchanged information about theirwork. The classics professor was teaching an un-dergraduate class in Classical Civilizations called"Food in the Ancient World: Subsistence and Sym-bor The students were planning to use themesfrom the class to create an exhibit at the archaeo-logical museum on campus.The undergraduate stu-dents selected artefacts and photographs from themuseum, created display labels, and helped do re-search for a website that would showcase the newexhibit.

A partnership was quickly begun,and CHICO teammembers were recruited to work with the profes-sor and her class to help them with the design andcreation of a website to be targeted to K-I2 teach-ers and learners. The professor continues to getinquiries from visitors to an earlier site on the topiccreated for a similar class four years ago, and stillreceives frequent e-mail from K- I 2 students andteachers asking for more information: "Are morereading materials available?" "Are there recipes?""And did they really eat peacock?" The professorresponds when she has time, but is not always ableto keep up with the demand for further informa-tion. She felt the interest was clearly there for anenhancement of the existing foods web site as wellas the creation of a new site on animals in the an-cient world.The new site would continue the themeof Food in the Ancient World but would focus onanimals: in hunting, as food, as symbol and myth, aspets, and in relationships with humans,with K-I2 asthe primary audience.

For the professor, time was an issue, since she wasinvolved in supervising the museum study for herundergraduate students and developing signage forthe exhibit.She felt it would be helpful if team mem-bers could work independently or in tandem withthe undergraduates. At the same time, the teammembers wanted to work closely with the profes-sor and her students in creating the content mate-rial. They believed it would be difficult, and even

counterproductive, to create the content for thewebsite in isolation from the structure and designthat the site would have. Therefore, it was agreedthat the team members would visit the class andspend time observing the professor and her stu-dents

For the Context and History part of the exhibit,CHICO team members gained insight into the back-ground of the content from the professor and wereprovided with readings. In turn, CHICO gave pre-sentations to the professor's class about signage andlabels for the exhibit artefacts, and also shared in-formation from the Education Curator at the UMArt Museum about signage practices at that mu-seum.A local radio station provided digital versionsof an interview with the professor conducted atthe opening of the Animals exhibit of the interview(in .WAV and RealAudio format) to be mounted ontheWeb site.

CHICO team members added new features to theexistingAncient Foods site and updated the revisedsite to include both a regular HTML version and aversion in Shockwave format, with a link providedfor viewers to download Shockwave if needed. AWeb site was created to accompany the Animalsexhibit that opened at the end of the term andwould run for several more months.The Museumhad just acquired a touch-screen kiosk for the 'ingallery' activities, and the team members created arelated resource for the kiosk, preparing a Direc-tor movie to use at the Museum's kiosk based onthe revised Foods site.A day was planned for fami-lies to visit the museum, and this would be an op-portunity to advertise the project.

For the Behind the Scenes component, team mem-bers photographed the undergraduate students andsome of the exhibit installation and exhibit-buildingprocess. Students also met with the Associate Cu-rator of Slides and Photos at the Museum and weregiven background on the animal exhibit.This will beincluded in a part of the site devoted to "Buildingan Exhibit" and will provide a behind-the-scenes lookat exhibition design and creation.

To incorporate a Relevance and Connections di-mension for the site, CHICO worked closely witha sixth grade teacher and her class, as describedbelow.

174



Engagement with K- 1 2 Audiences

Ancient Foods/Animals Project

One of the CHICO team members was enrolled ina class on digital resources for K-I2, a school mediamethods course. She was interested in "test driv-ing" ideas learned in class with K- I 2 students andteachers, and as part of the course, the student wasrequired to teach some aspect of information skillsto K-12 students.Through her class instructor, theCHICO student was connected with a local middleschool sixth grade teacher who had expressed in-terest in participating in the Ancient Foods project.

The sixth grade teacher had begun by asking herstudents if anyone was interested in "old things, re-ally old things." From the initial response, twelvestudents were identified, drawn from several classes.The team member and teacher agreed upon threedifferent project parts which could be exploredusing Eisenberg's model of the Big Six skills(Eisenberg & Berkowitz, 1990). The CHICO teammember helped the students find sites with infor-mation about animals in the ancient world andworked with the sixth graders as fellow research-ers.Together they searched the web for appropri-ate sites from the new Web site that could be usedas resources for the project. The learners partici-pated as collaborators on researching the site, andthe CHICO team member worked with them ontask definition, information seeking strategies, useof information, appropriate citing of sources, orga-nizing the information and putting it together, andevaluation of the information found.

An important activity was creating stories and ma-terials from the information gained in the researchprocess. Once the students had done preliminaryWeb research about the ancient works (specifi-cally on themes the UM undergraduate class wasexploring for the project), they explored ways thatinformation could be used and interpreted by hav-ing the students create stories and draw picturesbased on their findings. Some of these stories andpictures could then be included in the final Websitecreated for the Ancient Foods projects. The sixthgraders are creating clay models of ancient coinscontaining animal images, and these will be includedon the site as an example of student work.The firststep involved pencil sketches to be used as guidesfor the clay models of animals.The team member

photographed the art work with a digital camerafor inclusion on the site.The sixth graders have alsofound a number of interesting sites related to ani-mals in the ancient world.They may be included in aresource section for teachers.

The sixth graders will also be providing feedbackon the animal site as it develops. A key role for theK-I2 learners is evaluating the Website under con-struction for usability. Learners "test" the newWebsite while under construction to determinewhether it meets their information needs. For theSchool of Information student, this provides a goodopportunity to apply usability assessment skillslearned in coursework.

CHICO also arranged a field trip to the Museumfor the teacher and students from the local middleschool.The group was accompanied by a team mem-ber who took photographs for later posting on theAnimals Website. The sixth grade teacher helpedthe team member organize the project activitiesand made arrangements for a computer lab. Theconnection with the middle school, the resourcesfor museum education directed to youth, will beuseful to the museum as it focuses on becomingmore oriented to youth audiences.

Royal Shakespeare Company project

Engagement with K-I2 in the RSC project was builton a wide array of activities organized around theevent by the University Musical Society and its Edu-cation program. The RSC project afforded a num-ber of opportunities for engagement with schoolsand youth-centered activities.These included

youth performances

workshops with K-6 and K-12 teachers

activities taking place at K- I 2 schools

RSC visits with local schools

workshops for museum curators

Working with these groups,CHICO is gaining feed-back on the development of its site, as well as anopportunity to promote the site's existence and toencourage audience participation by submitting es-says and artwork to be included on the site.

CHICO students are developing a word game fea-ture which will allow K-12 learners to interact withthe text of Shakespeare's plays and get insight into



its meaning. A map feature will allow learners tolook at a map of Great Britain and see where thevarious characters and events in the plays are situ-ated.The interactive map of cities helps visitors tothe site explore the theater scene.A participatoryfeature of the site will encourage audiences to sub-mit their own content, reflecting on their own ex-periences with the plays, and creating new material

essays, drawings inspired from the plays. Con-tent will be reviewed, edited if necessary,and postedon the site.

Partnership on a Base of Mutual SelfInterest

A key factor contributing to the success of gener-ating and sustaining commitment to the partner-ship is that there be a high degree of mutual self-interest underpinning the participation of each part-ner. We have found it critical that there be a "win"for each participating party.

Museums

For the participating museums or other contentproviders, the benefits include

Rreaching out to a wider audience through digitaldisplay of artefacts and increased visibility andaccess for the collection

providing a marketing vehicle which sparks inter-est in the museum's content, promotes themuseum's mission and has the potential to build anew user base

providing contextualization which enhances ap-preciation of objects, and provides interpretationfor their understanding

providing enhancements to viewing objects, supple-menting but not replacing the physical museumvisit.

Information Specialists

For students preparing for careers in the informa-tion professions and their educators, the benefitsinclude

real-world experiences which bring togethermultiple dimensions of information service

an opportunity to act in a service role

an opportunity to apply technical skills to

humanistic endeavours

the opportunity to apply a range of informationskills: selecting, acquiring, organizing, presenting,and evaluating information

contact with end-users in museum and K-1 2 yen-ues

project-based learning from an interdisciplinaryperspective

K- I 2 communities

For K- I 2 educators and learners, the benefits in-clude:

gaining access to a broad array of objects and in-formation in distributed form and thus expandingthe base of learning resources

the opportunity for learners to take what theyhave seen and have read about an artistic experi-ence and engage in dialog with others

the opportunity to create their own artistic andinformation content, building on what they haveseen and heard.

the chance to apply critical thinking skills

the chance to extend an artistic experience be-yond the initial performance or museum visit

Content specialists

For content specialists, the benefits include

engaging with a broader audience in discussion oftheir topic

an opportunity to provide outreach and service

a chance to think about their topic from a differ-ent perspective and share that perspective withothers outside the academy or immediate schol-arly or specialist community.

Changing Roles

The benefits of partnership will most likely appealto those who are exemplars of the changing bound-aries of their profession. Each type of institutionrepresented by the partners has undergone somedegree of change in its mission, and a key part ofthe value of the partnership can lie in its ability tooffer opportunity to test some of the new rolesemerging in each of the professions represented.

176



Information professions

Education for information professions has beenundergoing profound changes recently, and at theheart of the changes are the changing roles withinthe information professions themselves. Librarians,for example, are now recognized increasingly as fa-cilitators of learning and information discovery, asproviders of distributed content to users in remotevenues,and as practitioners whose skills are of valuein a variety of information venues. In addition, thedelivery of instruction has changed to provide moreproject-based, experiential learning. More explicitattention is being paid to long-standing fundamen-tals of library and information service, with greateremphasis on the end-user. As a result, schools edu-cating library and other information professionalsmust now provide educational experiences whichintegrate classroom learning, technology, and prac-tice, deal with real organizational settings, are cli-ent-centered, and have a strong service philosophy.

While technology has recently been a driving forcein the profession, there is recognition that digitaltechnology is at once an enabling tool which makesdifferent kinds of learning possible, and an object ofstudy itself, with impacts reaching into almost ev-ery branch of human experience. Students of theinformation professions gain not only from assist-ing many different types of clients in the use of in-formation and information technology, but also fromobserving the impacts of information and informa-tion technology on environments such as museumsand K-12 classes (Frost, 1998).

While technical challenges are daunting, more chal-lenging still are the human challenges brought aboutby information technology. Students need to be-come acquainted with issues such as technologyacceptance, intellectual property, and informationaccess.The project-based experiences working withmuseum and K- I 2 clients can provide excellentexamples of these issues.

In addition, projects with an arts and technologyconnection are effective in attracting students whohave learned technical skills and want to apply themto interesting problems.Added to this is the oppor-tunity to provide a real service with a human di-mension. These projects have attracted studentsacross the spectrum of our curriculum, from archi-vists, librarians,and information policy specialists to

2

human-factors and usability experts and e-com-merce specialists.

When invited to participate in the RoyalShakespeare project, one student exclaimed,

I'm so eager to work on this project! In ad-dition to being a chronic Anglophile, I wouldfeel honored to contribute to a website thatwill provide such a valuable service to thecommunity. What makes my skill set espe-cially applicable to this project is diversity.Usually one finds programming proficiencyin one person, creative/graphic ability in an-other person, and writing skills in a third.My background, however, includes all of theabove.

Museums

Museum professionals also find themselves at a timeof change, with changes in their professional mis-sions, an expansion of their target audience, andnew options for delivery of their content. Manymuseum staff have not grown up with computers,but now find that technology has the potential tofundamentally change the way in which museumcontent is delivered.At the present time, web surf-ers can visit over five thousand online museum sites(Davis, 2000), and for many, the digital experiencemay become the primary way a visitor comes intocontact with a museum exhibit. At the same time,the digital experience differs in many fundamentalways from the experience of the original, and thisdifference can vary with adult viewers (Taylor,2000)and younger viewers (Thompson, 1982).

While it is obvious that the images seen by viewerson their computer screens will vary in significantways from those seen by viewers in their visits tomuseums, perhaps less apparent is the way in whichdifferent digital representations can also vary, de-pending on method of image compression, screenresolution,and the like.Visitors scanning a screenfulof thumbnail images or pursuing a series ofhyperlinked images will browse the collection dif-ferently from those who are visiting the physicalcollection in a sequential way (Frost, 2000).And ofcourse the experience of walking through a mu-seum exhibit provides a whole array of sensory ef-fects of smell, sound, touch and institutional aurawhich can at best be found in a computer experi-ence in only a limited and experimental way



(Greenman, 2000).Added to this mix of questionsis the social dimension of museum viewing and itsrelated impacts on individual and group experiencesboth within and outside the museum (Chadwick,1992, 1999, Falk & Dierking, I 992).With the poten-tial for interactive communication,will museums beable to engage audiences in new ways? What is therole for computers within museums? What are thecurators' roles in interpreting museum content, andhow are these roles now being shared with educa-tors? (Roberts, 1997)

Similar to "talk-back" areas in a museum, where visi-tors can post comments and questions about anexhibit, the Web offers opportunity, for examplethrough a Web-board, for visitors to respond towhat they have seen and to have questions answeredby museum staff. How does this new form of inter-action affect the viewer's experience and relation-ship with the museum?

All of these new and interesting questions will re-quire that museums have direct experience in of-fering their collections in digital form, in providingcomplementary and interconnected experiences,and in giving viewers an opportunity to interact withboth physical and online forms of the exhibit.Thiswill help museums address questions such as therole of physical and virtual objects, and whether thevirtual medium will serve to promote or deter vis-its to the physical collection. How can the virtualexhibit promote the afterlife of an exhibit, and howcan it help users become more informed beforevisiting the museum?

At the same time that the technology revolutionhas been changing museum practice, there has alsobeen a change among many museums in their coremission.Whereas the mission used to be to solelyto collect, preserve and exhibit works of art, thatmission has now expanded to include education andoutreach to new audiences, especially for youth andpopulations which have been historically under-served. Museum professionals' roles have expandedfrom curators and directors to include educators,marketing officers, and technology specialists. Edu-cators may be more closely involved in the planningof an exhibition. New vehicles exist for displayingor enhancing the exhibits to include video, audioguides, games, printed guides, CD ROMs, andWebs ites.

Content specialists

For the scholar, faculty member or other contentspecialist, there is the opportunity to expand thereach for the topic of research, to provide oppor-tunity for engagement with new audiences, and toprovide an outreach service to the community. Itgives an opportunity to bring together the threekey missions of teaching, research, and service. Itgives the scholar a chance for interdisciplinary en-gagement with content specialists in related areas.It also allows the faculty member, along with thestudents, to explore new means of knowledge workusing digital technologies.While this type of involve-ment is not typical for faculty members, developingon-line resources gives interested faculty an oppor-tunity to expand their roles as universities begin toplace greater emphasis on interdisciplinarity, out-reach, and the integration of service into teachingand research.

K-1 2

Several decades ago, Dewey pointed out the im-portance of experience in helping give meaning toabstract concepts and saw a special place for muse-ums and objects in education (Dewey, 1938), ideaswhich still remain a force in contemporary K- I 2museum education activities (Hein, I 998).With theappearance ofWeb-based museums and the abilityof access to primary resources and backgroundmaterials, these concepts take on new significance.In addition, new opportunities are afforded for learn-ers to interact with content and with other inter-ested learners, and to create and make sense oftheir own content.

Previous Projects Exemplifying theModel

The Ancient Foods/Animals and the RoyalShakespeare Projects build upon a base of pilot ac-tivities which were examples of similar collabora-tions with museums, K-I2, information specialists,and content specialists.

CHICO team members worked with museum cu-rators to develop a Website called Mummies ofAncient Egypt.This resource is aimed at K- I 2 stu-dents studying ancient Egyptian culture and is in-tended to complement visits to local archaeologi-cal collections. (http://www.si.umich.edu/CHICO/

1 73



mummy/). While we did not directly involve K- I 2communities in the development of the projects,the site has generated years of correspondence fromK-I2 teachers and learners offering feedback.

"Harlem I900-1940:An African-American Commu-nity" is a collaboration between CHICO and theSchomburg Center for Research in Black Cultureat the NewYork Public Library.The Schomburg askedCHICO to provide an online virtual exhibit of oneof its most popular collections,a photographic port-folio depicting the Harlem Renaissance.At the coreof the exhibit is an online presentation of theSchomburg's compelling portfolio,"Harlem 1900-I 940," featuring more than 30 archival photographs.A section for educators, with lesson plans and dis-cussion guides, is aimed at K- I 2 teachers and learn-ers. (http://www.si.umich.edu/CHICO/Harlem/) Asimilar partnership with the Schomburg resulted inthe creation of a site called The African Presence inthe Americas (http://www.si.umich.edu/CH1CO/Schomburg/).

In partnership with the Smithsonian's NationalMuseum of the American Indian (NMAI), CHICOteam members helped transform a site-specific ex-hibit into an interactive, multimedia resource avail-able to global audiences. CHICO staff developedthe exhibit based on content developed by AlaskanNative American Yup'ik Elders, with an NMAI ex-hibit curator (http://www.si.umich.edu/CHICO/Yupik). Building on this resource, a faculty memberand graduate student in the University of MichiganSchool of Education developed an interactive edu-cational resource with content provided by Yup'ikstudents and educators.Yup'ik elders also workedwith CHICO staff to transform the online exhibitinto a CD ROM featuring additional oral historiesand audio resources.

CHICO and the UM Museum ofArt (UMMA) joinedforces in a collaborative project with local middle-school students, art and technology teachers, themuseum director of UMMA, and the outreach co-ordinator of CHICO to collectively enhance thelearning process through an experimental electivecourse. In the Stylistic Journey Project, CHICOdeveloped aWebboard which was used extensivelyto facilitate online discussions among teachers, cu-rators and students (http:llwww.si.umich.edu/CH1CO/Emersona.

1,79

In a CHICO partnership with the University ofMichigan Museum of Art and a local school, a set offifth graders from a local school were engaged indeveloping an exhibit guide which would reflecttheir perspective but also have its content reviewedand validated by UMMA curators. Together withthe UMMA museum professionals, the students,their art teacher and the school's technology spe-cialist, CHICO helped create "Monet at Vetheuil",an online, interactive exhibit guide.Youngsters wereable to research specific artistic resources and paint-ing techniques in consultation with curatorial ex-perts, and to create their own works of art basedon their experiences and learning. A bulletin boardand online quiz facilitated communication amongstudents, educators and the exhibit curator, andwelcomed responses from a global audience (http://www.slumich.edu/CHICO/monet).

Lessons LearnedIn the course of implementing and evaluating theprojects, a number of recurring themes and issueshave surfaced to assist us in future development ofsimilar sites.

Intellectual Property issues

Since copyright takes on new dimensions in the digi-tal world, many of the assumptions to which wehad become accustomed are changed or no longervalid; many old assumptions are challenged. Oneexample is the ownership of items in the collectionof a museum or archive and our ability to use themin collections. For example, if a museum educatorwants to create a site for teachers in which sheuses images from her local museum, and if thatmuseum "owns" all the material in its collections,and with it all donated items, does that mean thatthe museum is free to make images of these itemsuniversally available through the Internet? Althougha museum or other repository may "own" a photo-graph, painting or statues, and may have been giventhe right by the donor to display that object in thecollection, that does not necessarily entitle the in-stitution to post images of the item on the Web.Instead, such "ownership" conditions needed to bere-negotiated, and securing permission to use digi-tal objects may involve going beyond their original"owner:'



In addition, we found it critical that agreements beworked out in advance and in writing indicating theownership and disposition of the intellectual prop-erty of the content provider, as well as how theproject team may use or distribute the content inthe course of the project work.

Rules of disengagement

The CHICO projects have as one of their main goalsthe provision of educational experiences for infor-mation students.While the development ofNiVebsitesis a key means of providing these experiences, theproject is not serving primarily as a production unit,and decisions must be made in a way which bal-ances the educational needs of the students withthe needs of the project partners. Staff change eachyear, with new students entering, graduating, ormoving on to other projects. Staff skill levels vary.Accordingly, decisions must be made as to when tomove a site to the CHICO archives (and notify view-ers that the site is no longer being actively main-tained), when to continue maintenance and upgrad-ing, and when to remove a site outright. Ideally, asite would be handed off to the participating projectpartner for maintenance, but this is not always fea-sible. If a site is to be continued, how can it be keptactive and fresh while new projects are developed?This can be a key problem since student participa-tion is voluntary. Students are more likely to be at-tracted to the development of a new site in whichthey have a stake, and for which they can take creditand can cite in their job search portfolio.

Coordinating schedules and goals

University course projects need to be coordinatedwithin the framework of the academic term or se-mester, but in real life, project activity does not al-ways fit neatly within this timeframe. Project ex-pectations and deliverables need to be managedwithin these constraints, without penalizing studentsfor difficulties beyond their control. Difficulties incoordinating meetings are to be expected, and thiscan cause a lag in project work, yet the studentsexpect to be actively involved in a project in thecourse of a term. If the project experiences a seri-ous delay, the students are encouraged to try to fillin this time with project activities within their con-trol (e.g., developing background information, locat-ing information resources, or preparing draft pro-posals of work for reaction from the project part-

ners). If the student teams experience a major set-back, the project goals need to be adjusted, and inthe unlikely event that the problem is not correct-able, the students are encouraged to focus insteadon an evaluation and assessment, with recommen-dations based on lessons learned.

Matching skill levels with project needs

The projects are designed to allow student partici-pation at various skill levels. Projects start off bydoing a skills inventory to determine what skills stu-dents have and need. Students are then pointed toresources, on campus and elsewhere, that can as-sist them in gaining new skills. Students are pairedwith experienced team members, and novices ofone year become seasoned trainers in their sec-ond year.Where possible, projects are matched withskill level and interest. SI students generally have avariety of backgrounds, with both content and tech-nical expertise reflected in the team membershipat any given time.

ConclusionThis paper has described a model for communitypartnerships involving information specialists intraining, museums and other content providers,content specialists, and K-12 teachers and learners.The focus has been on two recent activities involv-ing partnerships with museums, libraries, and a per-forming arts group.The experiences have provideda valuable opportunity for information specialists,museum curators, and K-I2 educators to gain in-sight into new roles in their changing professions.

AcknowledgementsThe project work undertaken in this project wasaccomplished by the following team members:Yifan Ji,Katrice Lewis,Adina Lipsitz, John Northup, JanetSzcsesny, and Zhiping Zheng. Special thanks are due toteam member Zilia Estrada, who had a particularlyinstrumental role in the development of both projectsand in reporting on the group activities. UM ProfessorSusan Alcock and her undergraduate class provided thecontent for the Ancient Foods and Animals sites.Materials from the Royal Shakespeare Society and theUniversity Musical Society provided key content to theRSC site. Support for the projects was provided in partby the W. K. Kellogg Foundation.



References

Chadwick, J. ( I 999).A survey of characteristics andpatterns of behavior in visitors to a museum website. In D. Bearman & J.Trant (Eds.) Museums andthe Web I 999: Selected Papers from an InternationalConference. Pittsburgh,Archives & MuseumInformatics. http://www.archimuse.com/mw99/papers/chadwick/chadwick.html

Davis, D. (2000).The virtual museum, imperfect butpromising, New York Times, Sept. 24, 2000

Dewey, J, ( I 938). Experience and Education. New York:Macmillan.

Duckworth, E. Museum Visitors and the Developmentof Understanding. Journal of Museum Education 15(I):1990, pp.4-6.

Eisenberg, M.B., & Berkowitz, R.E. (1990). InformationProblem-Solving:The Big Six Skills Approach toLibrary and Information Skills Instruction.Norwood, NJ:Ablex.

Eisenberg, M.B. and Berkowitz, R.E. The Big6 Skills:Information Problem-Solving Approach...hap:IIwww.big6.com/

Falk, J.H. & Dierking, LD. (1998). Understanding free-choice learning:A review of the research and itsapplication to museum web sites. In D. Bearmanand J Trant (Eds.) Museums and the Web 98Proceedings (CD-ROM).Archives and MuseumInformatics, 1998.

Frost, C. O. et al. (2000). Browse and search patternsin a digital image database. Information Retrieval, 1,287-313.

Frost, C. 0. (1999). Cultural heritage outreach andmuseum/school partnerships: Initiatives at theSchool of Information, University of Michigan. In D.Bearman & J.Trant (Eds.) Museums and theWeb1999: Selected Papers from an InternationalConference._Pittsburgh,Archives & MuseumInformatics. 223-229. http://www.archimuse.com/mw99/papers/frostlfrost.html

Frost, C. 0. ( 1998) "Library Studies at the School ofInformation", Library Hi Tech,Vol. 16, No. 2, 91-102.

Frost, C. O. When the Object is Digital: Properties ofDigital Surrogate Objects and Implications forLearning . In S. Paris (Ed) Multiple Perspectives onChildren's Object-Centered Learning; Erlbaum Press(In Press).

Greenman, C. (2000). Museum goers get a virtualhands-on experience, New York Times, September14, 2000.

Hapgood, S.E. & Palinscar,A.S. (2001). Fostering anInvestigatory stance: Using text to mediate inquirywith museum objects. In S. Paris (Ed) MultiplePerspectives on Children's Object-Centered Learning.Erlbaum Press (In Press).

Hein, G. (1998). Learning in the museum. London:Routledge.

Hindle, B. (1978). How much is a piece of the truecross worth?, in I. M. G. Quimby, (Ed.). MaterialCulture and the Study ofAmerican Life, New York,WW Norton, 5-20.

Keim,A. (2000). Discovering dinosaurs on 2 CD-ROM's.NewYork Times, September 1 4, 2000.

Roberts, Lisa C. From knowledge to narrative: Educatorsand the changing museum.Washington and London:Smithsonian Institution Press, 1997.

Taylor, B.L (2000).The effect of surrogation on viewerresponse to expressional qualities in works of art.Unpublished preliminary results from a doctoraldissertation. University of Michigan, School ofInformation.

Thompson, C. (I 982).Why do you need to see theoriginal painting anyway? _Visual Resources, 2, 21-36.

18.1© Archives & Museum Informatics, 2001 187

Unseen Discussions: [email protected]

Anthony Huberman P.S.I Contemporary Art Center, USA

Abstract

For the recent exhibition Greater New York: New Art in New York Now, the Education Department at RS.IContemporary Art Center organized a unique email-based discussion. The museum set up an e-mail address formost participating artists using the free Hotmail.com service. These addresses were displayed on RS. I 's website,where each artist had his/her own Web-page, and on sheets available in the museum lobby; and most importantly, theaddresses were printed directly on to the wall labels in the galleries.This established a platform where public andartist could discuss the work and exchange ideas. The forum was easy, free, international, immediate, private, andalmost intimate.The museum relinquished its monopoly over the interpretation of the art by opening up unmediatedchannels of communication between artist and public.

P.S. I Contemporary Art Center

RS.1 Contemporary Art Center is a large museumlocated in the Long Island City section of Queens,NewYork.Since 1971 under the direction of Found-ing Director Alanna Heiss, RS.1 has showcased thework of thousands of local and international emerg-ing artists and has been a defining force in NewYork's cultural life. Housed in a 100-year old schoolbuilding, the museum organizes critically acclaimedsolo and group exhibitions year-round and a sum-mer outdoor live DJ Series.Since 1997, RS.1's reno-vated building has been the largest contemporaryart center in the world.The museum combines aworld-class exhibition program, a prestigious Na-tional and International Artist Studio ResidencyProgram, and a broad spectrum of education andpublic programs that serve its many audiences. Inearly 2000, RS. I became an affiliate ofThe Museumof Modern Art, New York, and looks forward toexploring the many possible collaborations withMoMA over the next years.

The education program

Since 1985, RS. 1 's Education Department has of-fered its various audiences opportunities to get in-volved with the museum, the exhibitions, and theartists. Continuing RS. 1 's commitment towardsworking closely with artists, the education programreflects a particularly active effort to bring a widepublic in contact with the emerging artist commu-n4Youth and family programs include art classesand art-making workshops with artists included inRS. I exhibitions.School programs include tours andart-making workshops with artists,collaborative on-line project development with artists, and a Teen

182

Curators Series.Adult programs include panels anddiscussions bringing artists, writers, and curators incontact with the public, a Writers Series showcas-ing texts by emerging writers on pieces in the mu-seum galleries, brochures and print materials withartist statements and interviews, and interactiveWeb projects.

Greater NewYork

In February 2000, RS.1 and MoMA presented theresult of their first curatorial collaboration.The ex-hibition Greater New York:New Art in NewYork Nowbegan as an open-call for submissions, and severalthousands portfolio submissions and hundreds ofstudio visits later, became a massive building-wideshow for which 30 P.S.I and MoMA curatorsbrought together the work of over 140 NewYork-based artists who have emerged as vital, creativevoices in their fields over the past five years. Theexhibition turned out to be one of the most suc-cessful shows in the history of RS.1, drawing thou-sands of visitors and critical acclaim. Hits to RS. 1 'swebsite, http://www.psl.org, tripled over the courseof the exhibition, as the museum hosted its mostambitious on-line exhibition to date, now on viewat http://www.psl.org/cut/tours.html. The manycomponents that made up the exhibition contrib-uted to its unprecedented success.

The E-Mail Project

One of the components initiated by the museum'sEducation Department was dubbed the E-MailProjectA month before the opening of the exhibi-

Museums and theWeb 2001:Proceedings


BEST COPY AVAILABLE

Huberman, Unseen Discussions: artist

tion, museum staff set up an e-mail address for eachartist using the free Hotmail.com service.To betterdistinguish the newly created e-mail addresses, thefollowing format was used:

[email protected].

After receiving a long explanation of what theproject required and the responsibilities that ac-companied it, over 80 of the 146 artists included inthe Greater NewYork exhibition agreed to join.Allartists were shown how to use the service and howto access their private account.As an important stepto ensure the success of the project, the artists werestrongly warned that their participation requiredthem to check the account several times a weekand to respond to all relevant messages.Those notable to make that commitment were excluded fromthe project.

Each artist was made aware that the dialogue thatoccurred through the e-mails would not be screenedby the museum and would be seen by the artistalone. Different from a "chat room" or a "bulletinboard," the messages would not be posted on themuseum website but would remain in the privateHotmail inboxes of the recipient artist. Since noinstitutional control was possible, all were warnedof the possibility for "spam" or other unwanted orunwelcome messages. Furthermore, since there wasno precedent for the project, the quantity and qualityof the messages remained completely unknown andimpossible to predict. The museum was eager toshowcase selected messages on the P.S.1 GreaterNewYorkWebsite, and artists were encouraged toforward particularly interesting messages to themuseum Web site staff.

The e-mail addresses were then displayed in threeplaces:

on RS. I 's Website, where each artist had his/herown personal Web page

on sheets available in the museum lobby for visi-tors to take away

on the wall labels of the artists' works in the gal-leries

hotmail.com

Through the Website, a local and international au-dience was able to contact the artists directly bysending them direct and private e-mail messages. Inthe Greater New York site, each artist had his orher personal Web page, complete with an image, anexhibition history, a selected bibliography, an artiststatement, and an e-mail address.When and if theywere fonmarded to museum staff, excerpts from e-mail correspondence were also posted in the artist'spage, adding new insight to what was an invaluableresource for information about significant up-and-coming artists.

Visitors to the museum could take home with theman "address book" of e-mail addresses for the art-ists.One computer was placed in the museum lobby,but was meant for browsing the Web site, not forsending e-mails. Since e-mail correspondence re-quires users to log into a personal home or officeaccount, visitors could not use the computer on-site at RS.1 to correspond with the artists. Instead,museum-goers could take home a free pamphletwith the complete lists of existing e-mail addressesfor Greater New York artists and compose a mes-sage at a more comfortable time and place.

Most importantly, visitors in the galleries, when con-sulting the wall label for the usual name and titleinformation, were confronted with the possibilityof communicating directly with the artists in theexhibition. By simply adding a line to the standardmuseum wall label,the presence of the E-Mail Projectwas carried from the digital world of the Internetto the physical world of the galleries themselves, anexercise which remains unusual for many webprojects.

Jeremy Blake

Angel Dust, 1999

Digital projection

Courtesy Feigen Contemporary, NY

[email protected]

Fig. l:A sample wall label from the exhibition,including an email address

183



Results

Several factors played a part in evaluating the re-sults of the project. First of all, since there exists nopre-existing model with which to compare it, a com-prehensive understanding of its successes and short-comings is difficult. In addition, the very nature ofthe project, and indeed one of its most significantand characteristic traits, was that the messages werekept in private e-mail accounts, inaccessible to mu-seum staff. Any assessment was possible only byasking the participants to discuss what they had re-ceived.

Once the exhibition closed, Education staff did in-deed interview many of the artists in order to getan idea of the results. An important question wasthat of identifying the different types of users. Feed-back from the artists revealed that e-mail messagescould be divided into 5 roughly defined categories,here listed in no particular order:

curators and gallery owners

collectors

writers participating in the Greater New YorkWriting Project

the "unidentified" general public

teachers

An obvious but admittedly unanticipated groupof e-mail corresponders were local and internationalcurators and gallerists. Partly due to its sheer size,to the involvement ofThe Museum of Modern Art,to an aggressive publicity campaign, and naturally, tothe quality of the works, the exhibition received asignificant amount of press in national and interna-tional papers and art magazines. Many curators, fromNewYork galleries to prestigious institutions aroundthe world, visited the exhibition and the exhibitionWeb site. Many of them took advantage of the E-Mail Project to contact artists and request addi-tional information about their work, images, CVs,or studio visits. Artists engaged in continuous dia-logue with curators and became involved in exhibi-tions in other spaces, from Houston to Berlin, Lon-don and Tokyo. Important to keep in mind is thatmany of the artists included in Greater New York

were not yet represented by a commercial gallery,and this direct and facilitated link to curators provedvery promising and helpful.A particularly potent ex-ample is that of a young and unrepresented artistreceiving a personal message of interest from theDirector of the Brooklyn Museum of Art, ArnoldLehman. Of course, artists already working withcommercial galleries referred the interested cura-tor to the appropriate person at the gallery to con-tinue the correspondence. E-mails from curators,in the case of some artists, made up more than halfof all messages received during the course of theshow.This impressive tally was certainly a welcomeby-product of the E-Mail Project, especially from theartists' point of view. Nevertheless, the project wasconceptualized by the P.S. I Education team moreas a tool for outreach and for encouraging an openunmediated exchange of ideas and interpretationsbetween a general public and the artists rather thanas a way to further careers.This particular outcomeof the project was not among the intended orsought-after goals, yet it stands undeniably as aprominent one.

Another frequent user of the e-mail addresses wasthe art collector. Once again, this audience is onewhich the initiative did not aim for, but was never-theless significant.As with curators, artists were onlytoo pleased to receive messages from collectors.One young artist, who had shown her work verylittle before her inclusion in Greater New Yorkshe is now frequently featured in exhibitions in manyNewYork galleries was quite unprepared for theresponse from collectors, and sold all of her pieceswithin the first few weeks. She then spent the nextmany weeks telling collectors from across the coun-try that all her work had been sold and that shewas working on new pieces (as fast as she could,presumably). An issue that becomes immediatelyapparent is that of artists selling their work with-out going through the gallery that represents them.Of course, it is difficult to convince artists to di-vulge details concerning the sales of their artworks,but the opportunity to by-pass the gallery is onewhich surely crossed the mind of several.

The Greater NewYorkWriters Project was anotherinitiative of the Education Department and whosefull description deserves a paper in itself.To accom-pany the open-call for artist submissions, the mu-seum organized an open-call for writer submissions.

© Archives & Museum Informatics, 2001 I 9 I


Given the difficulty of articulating a cohesive themeand a concise curatorial essay for a show so het-erogeneous in content, the exhibition catalogue wasmade up of texts selected from the group of emerg-ing writers from theWriters Project.Approximately100 texts were selected by a screening committeeat PS.I and MoMA, giving the exhibition a broadscope of interpretation and insight.Texts varied fromthe art-historical to the creative, as theoretical es-says appeared alongside more creative short sto-ries or poems.The e-mail addresses proved to bean invaluable resource for many writers.Artists re-ceived messages filled with questions about theirwork, and later on, drafts of essays-in-progress. Ea-ger that the text be included in the catalogue, art-ists responded to the questions generously andparticipated in what could be called an e-mail inter-view. Following is the beginning of one such re-sponse:

Dear Brian, thanks for writing. I've had somecontact through the PS.I e-mail address.Actually, more than I thought and a few help-ful things for the future. I will see what I canfax to you. I will copy my bio and a review ofone of my CDs. The Wire article you canfind through my label JDK. If you do asearch on Radiants lab and JDK, you shouldbe able to find the site.You can find somearticles reprinted there.Attached to this e-mail is the bio, etc. If you can, also check outmy small piece at Postmasters on W. 19th.Thanks a lot.

Given the artist's accessibility through the E-MailProject, several writers who would not have doneso otherwise decided to contribute a text. Indeed,many artists and writers, over e-mail of course, ar-ranged to meet in person to further discuss ideasfor the essay. Having the Greater New York e-mailaddresses as resources for private exchanges thatguaranteed a personal,tailored response promptedmany young writers to move beyond the often para-lyzing intimidation that is so common among peopleuninitiated in, and unfamiliar with, the New York artworld.

The last two segments of the E-Mail Project audi-ence were those that the project most aimed toreach. Most artists confirmed that about a third ofthe messages they received were from a more gen-

hotmaiLcom

eral public, unaffiliated with the professional artworld.Art students, local residents, tourists, and thesimply curious found the opportunity to communi-cate directly with the artist to be an unusual exer-cise in sharing thoughts, impressions, opinions, andsuggestions. For example,for a piece involving a fullyfunctional sauna in the museum's courtyard wherevisitors were invited to disrobe and join in, responsesvaried from the simply helpful:

There is a good article on nudity in Austin atwww.austinchronicle.com that you might beinterested in.

to the more profuse:

1 8 5

the next morning I woke up thinking aboutmy experience and how it was that you man-aged to coax total public nudity out of me insuch a short time. what was at play in yourpiece that made it so easy for me? and formy girlfriend as well for that matter? Andmore importantly, do those factorsaproximate [sic] the culture of public bath-ing you described in Finland?You see, to us,the fact that the curtain is slightly parted,the changing room door does not actualy[sic] connect with the curtain, the ability tosee people outside from behing [sic] thedoor, the clear plastic of the ventilation slotetc... all those elements created a kind ofkinkyness. they all played on the notion ofvoyeurism, it was all a kind of game wherethe bather plays with how much or how littlehe is willing to show, every time he moves,he has to consiously [sic] decide to bedescrete [sic] (prudeish?) or not. Every timehe notices an opportunity for exposure heis confronted with either exitement [sic] orfear or a realization of shame about that fear.It is therefore key to the experience that itbe a repetition of the same series of deci-sions. with each cycle, the bather may be-come more daring (or less so). that daringevolves very quickly over those forty fiveminutes because the bather has to makethose kinds of decisions so often. In any case,I would really appreciate it if you could writeme back and tell me a little bit about whatrole you think sexuality plays is your piece.



People seemed to be unusually frank and almostvulnerable in their responses:

Your video piece was very endearing.Well,that makes it sound almost cuddly. It wasalso kind of frightening.There was some sug-gestion of violence, to me. Maybe I associatecircuses and carnivals with mayhem.

Do you consider this work to be a productof your Colombian heritage, or is it strictlyan illustration/embodiment of the five physi-cal dimensions?

Straight-forward "fan letters" were also very com-mon. Instead of feeling the need to ask questions,people enjoyed being able to simply express theirenthusiasm for the piece.Artists read these letterswith great appreciation. Repeatedly, artists men-tioned that these types of letters are a missing com-ponent in the art world: rarely does an artist re-ceive a letter of praise from an unknown exhibitionvisitor. Most of the time, feedback exists only in criti-cal reviews in the press or through friends. Indeed,many artists saved and treasured these short "fane-mails." Following is one of the favorite notes re-ceived by one of the artists. It is written by a youngstudent who visited the museum on a class trip:

To Artist,

I recently visited RS. I with my school, I goto The Ursuline School in New Rochelle. Ichoose your piece as my favorite. I thoughtit was very creative and calming. I enjoyedthe way the cotton candy actually movedwith the music. I also liked the split screenwhich made it look like it was a mirror im-age. The music was soothing and watchingthe cotton candy was amusing. I think youridea was very original and creative. I didn'tunderstand the music so i was happy thatthere was a translation. Please write back.Thank you.

In addition to their students, many teachers usedthe e-mail address to get in touch with artists inorder to better prepare lectures or workshops fortheir classes. Several artists sent teachers ideas forworkshop projects or provided them with a more

elaborate explanation of their work. After a classvisit to the museum, a teacher gave her students anassignment to use the e-mails to contact an artistof their choice and ask him or her a specific set ofquestions. Some artists were bombarded with 20messages from 8-year old children and patientlyanswered their questions. One curious and coura-geous artist, taking advantage of the anonymity ofe-mail, actually attended the lecture of a collegeprofessor who had been in touch with her abouther work, simply to hear what the lesson woulddiscuss.

Used to more ordinary and "safer" responses toartworks that appear in Bulletin Boards or are heardat panel discussions and public forums, the RS.IEducation Department was pleased to see an ex-change of ideas that seemed less inhibited than inmany other systems and strategies for gatheringaudience feedback.

A reporter for the online zine FEED,Claire Barliant,independently interviewed several participating art-ists for an article. Noting the absence ofWeb art inGreater New York, the reporter named the E-MailProject as the museum's stab at including theInternet in the exhibition. Her conversations withartists suggested a mixed result, some artists re-ceiving a significant amount of correspondence, butothers voicing disappointment in the turnout.Thearticle noted a particularly intriguing occurrence,"prov[ing] the elasticity of the e-mail project"(Barliant, 2000), where an artist not included in theexhibition nevertheless created a [email protected] account for himself.

Implications for Museum Outreach

The Greater New York E-Mail Project is an intelli-gent and effective strategy for engaging an exhibi-tion audience. Museums across the world have triedmany different methods and systems to foster arelationship with their visitors, ranging from docenttours, art classes for children, and museum-schoolpartnerships to scholarly lectures and film screen-ings, not to mention merchandise and brand names.Rarer, however, are attempts to put the public, orrather, the publics, in direct contact with the artists.This is a commitment that P.S. I has always main-tained.



The strongest aspect of the E-Mail Project is whatit stands for.The FEED magazine article mentionedabove determined that since artists did not receivean overwhelming amount of insightful and intellec-tually provoking messages, the project as a wholefailed in its purpose as an effective forum for dia-logue. On the contrary, the project succeeded evenbefore anyone put fingers to keyboard. Whetherthey were consulting the wall label, visiting thewebsite, or taking the e-mail address book home,viewers of Greater NewYork were confronted withan exhibition interested less in convincing its publicof a specific curatorial viewpoint or appreciation ofart, than in providing a stimulus for others to con-sider and interpret. By adding e-mail addresses intothe exhibition, Greater NewYork curators openedup important channels of personal and private in-terpretation and evaluation.

Curators, and the museums that legitimize them,have traditionally been the authors of"correct" and"incorrect" understanding of art and gatekeepersof good and bad taste. Exhibitions have been ve-hicles and frameworks through which curators andtheir museums tell their viewers how to think aboutart. Little or no room is left for the visitor to de-velop, much less articulate, his or her own meaningfor the works. In recent years,an increased intereston the part of museums and curators in seekingthe participation of the audience has spawned non-linear exhibition design and layout, education andoutreach programs, including bulletin boards (bothphysical and digital), interactive website projects, and"meet the artist" events, among other many ex-amples. More and more, especially in contemporaryart museums, the visitor's input is sought after andlegitimized. This trend is fueled by the belief thatleaving space for public participation makes themuseum less of an intimidating fortress for the eliteand more of an appealing and accessible center forshared ideas.

The E-Mail Project functions with similar goals andpost-modern theoretical roots but brings new en-ergy to an overly-saturated world of museum out-reach initiatives. The project's simplicity is deceiv-ing the mere appearance of the e-mail address onthe wall label, on the website, or in the lobby hasthe symbolic effect of encouraging discussion anddebate. Viewer opinions and impressions are sud-denly acknowledged and legitimized before any such

hotmail.com

opinion or impression has even been written down.The negotiations around the content of the exhibi-tion are removed from the hands of the curatoralone and extended to include anyone interestedin joining.The project provides visitors with an openchannel for the pursuit and development of theirown personal interpretations and associations. In-deed, the personal e-mail addresses were commu-nication tools as much as they were symbols for ademocratization of the understanding of art. In ef-fect, the simple act of displaying the artist's avail-ability to discuss the work, unsupervised by mu-seum staff in the shape of an e-mail addresspoints to the willingness on the part of the cura-tors to relinquish their monopoly over a "correct"interpretation of the artworks.

Setting this project apart from other systems ofaudience outreach is its invisibility. E-mail correspon-dence provides both sender and recipient with aprivate, immediate and almost intimate medium ofcommunication. Being part of over 140 included inthe exhibition, many artists welcomed the intimacythe e-mails brought to their experience as partici-pants in Greater New York. In a show of its scale,one-on-one communication and channels for feed-back are difficult to establish, and the exchange ofprivate e-mails provided a way for artists and audi-ence to maintain a relationship with the show. Asalready noted, privacy helped audience membersfeel more comfortable about communicating withthe artists.As is often the case in public forums andpanel discussions, visitors can find it intimidating tovoice their opinions, questions, or personal asso-ciations and interpretations, a difficulty easily over-come by the unmediated nature of e-mail messages.Especially important was the assurance that theircomment would be heard and responded to. In themore public bulletin boards or feedback notebdokscommon in many museums, viewer contributionsare rarely, if ever,answered or acknowledged. In thiscase, the corresponder had the luxury of a specificlistener and could look forward to a reaction to anopinion or an answer to a question.This aspect ofthe project heavily contributed to the volume ofmessages sent and helped the museum position it-self as the host for a tailored and careful platformfor meaningful exchange.As many pedagogues havestated, education works best if it occurs throughdialogue rather than monologue. Indeed, the E-MailProject transformed the museum experience from

1 71 94 © Archives & Museum Informatics, 2001


monologue to dialogue: from quietly absorbing acuratorial stance (in public) to actively negotiatingthe importance and meaning of the works (in pri-vate). It seems as if museums can use the Internet,or more specifically, e-mail, as an appropriate "toolfor the task." Interestingly, museum education canbe left as a private experience.

Other significant traits of the project are its cost-efficiency, its international reach, and its role in theactual gallery display of the exhibition. First, the E-Mail Project was completely free of cost.The mu-seum could set up e-mail accounts on the freeHotmail.com service, after which the project wasleft in the hands of the public and of the artists.During the private exchange of messages, museumstaff members were not involved in screening, edit-ing, or maintaining the correspondence and couldlet it run its course. Second, by including the ad-dresses on the museum Web site, the project wasinternational in scope. Initiated by a comprehen-sive site which included images, bios, analytical andcreative essays, and streaming sound and video,web-surfers from around the world sent artists theirimpressions and questions. Many artists noted thatthe wide reach of the website led to an excitingvariety of opinions and remarks.While some art-ists received invitations to participate in exhibitionsabroad, others heard from people with surprisingreactions, sometimes written in a foreign language.This characteristic also suggests that one could setup a similar project for an exhibition in any loca-tion. Being in New York certainly helped keep at-tendance figures high, but an e-mail based dialoguecould be an effective and successful component toexhibitions in any city in the world.

As briefly mentioned above, a final aspect rich inimplications is the role played by the project in thegalleries themselves. Including the addresses on thewall labels on one hand disrupted, or rather, rede-fined a traditional relationship between artwork andits viewer,and secondly,fully integrated the Internetas a central component in the exhibition. As oneartist emphasized, the insertion of the e-mail ad-dress on the wall label adds new life to a long-stand-ing standard in the techniques of museum display.Adocument which usually contains nothing more thanfactual information suddenly gains a certain open-endedness that can powerfully alter the viewer'sexperience with the work of art itself. Furthermore,

the introduction of the universally understood"@"sign into the gallery space highlights the role playedby the Internet in the physical experience of theworks. No longer a marginal subsidiary to the physi-cal display of the show in the form of an online-only project, the Internet is woven into the fabricof the exhibition itself. In addition to the e-mailaddresses, selected texts from theWriters Project,all submitted to the screening committee over e-mail, were on view in the galleries. Interestingly, inher article for FEED magazine, Claire Barliant, not-ing that Greater NewYork did not include any art-ists who use the Web as their chosen medium,names the E-Mail Project as the Net.art segment ofthe exhibition (Barliant, 2000).

Limitations

Far from flawless, the project gained from its firstincarnation several insights as to its limitations andpossible improvements. A first limitation is that itmay only be possible with relatively emerging art-ists. Since its success depends on the reliability ofthe artists to check their e-mail accounts and torespond to any mail, it is difficult to expect thatworld-renowned artists would be willing to investthe time and energy necessary. Immediately follow-ing Greater NewYork, RS. I presented Around 1984:A Look at Art in the Eighties, with artists such asPeter Halley, Jenny Holzer, Julian Schnabel, SophieCalle, and Anish Kapoor, among others.With verybusy schedules and near-celebrity status, these art-ists would surely not have agreed to maintain an e-mail address. Artists in the beginning of their ca-reer, on the other hand, are more likely to take onthe responsibility of corresponding with the generalpublic over e-mail on an on-going basis.

188

Another problem lay in the quality and quantity ofthe e-mail messages themselves. For example, manyartists complained of "spam" mail.Artists receiveddozens of messages from various mailing lists. Someartists took advantage of the network of e-mailaddresses to publicize shows of which they werepart, and many exhibition announcements were sentout.This problem is one that plagues all e-mail us-ers and seems to be an unfortunate by-product,and one difficult to avoid, of the technology. Also,some artists received significantly more mail thanothers.With a show of its scale, it was impossibleto guarantee that each artist would get an equal



number of messages. Perhaps a smaller exhibitionwould lead to a more equal distribution of corre-spondence.Additional publicity might also have beenhelpful. Had the E-Mail Project been announcedmore formally on the exhibition print materials(press release, brochure, postcards), or even better,had it profited from its own press release or an-nouncement card, it might have attracted more us-ers. Again, the existence of the project was visibleon the wall labels, on the Web site, and as a take-away pamphlet in the lobby, and not in any mailedor otherwise distributed materials.

One last suggested improvement is to stress theimportance to artists of forwarding excerpts ofparticularly interesting messages to the museumstaff, for Web posting. With artists acting asgatekeepers, the project could have profited from alarger amount of excerpted correspondences be-ing made public. Not only would these short textshighlight the dynamic interaction going on behindthe closed doors of the Hotmail addresses, but theyalso would act as a bulletin board moderated bythe artists, and surely prompt more people to re-spond and react to the postings by sending mes-sages of their own.

Conclusion

Greater New York and its use of the Internet ande-mail was an exciting and unique experience forall involved: museum staff, artists, and audiences.Emphasizing a space for an open, free, and privateinterpretation of the works on view, yielding unin-

hotmaiLcom

tended but beneficial results such as exhibition andsale opportunities for artists, reaching audiences ofall ages and nationalities,and bringing new media tothe forefront of museum outreach techniques, theE-Mail Project combined the complexities of themuseum experience with the enthusiasms of themuseum audience to create a one-of-a-kind syn-thesis of art and communication. It is surely a strat-egy that deserves the attention and considerationof education, curatorial, and new media museumprofessionals.

References

Greater New York Artists referenced:

Adriana ArenasMatthew BuckinghamBeth CampbellJordan CrandallHope GinsburgGareth JamesJulian LaverdierePia LindmanJennifer and Kevin McCoyMick O'SheaPaul PfeifferJohn PilsonJavier TellezAnton VidokleStephen Vitiello

Barliant, C. (2000).The Challenge of Exhibiting aSupernatural Instrument.Consulted January 10th,2001. http://www.feedmag.com/essay/es3431offhtml.

s


A Curatorial Resource for Upstart Media Bliss

Beryl Graham and Sarah Cook, University of Sunderland, UK

Abstract

Using resources gathered for the CRUMB website (http://www.newmedia.sunderland.ac.uk/crumb0.this paper presentssome key information and debates regarding the curating of new media art, both on the web and in conventionalgallery spaces. Including quotes from Sarah Diamond, Steve Dietz, Peter Weibel, Benjamin Weil and Kathy RaeHuffman, the paper discusses issues of how artists are paid, how new media art is archived, how audiences mightrespond to new media art, and the aesthetics concerning the presentation of new media genres. Practical examplesfrom the authors' experience of curating new media art are also given.

Museums consider having web presence as obligatory as membership, informa-

tion, e.commerce, etc. and then eventually an art space.The art space is not quitean afterthought, but it is not the prime focus. (Weil, Sins of Change 2000)

... the Susan Collins artwork Audio Zone is spread around the exhibition space.The audience must wear infra-red headphones, which at certain points receiveseductive voices urging you to 'touch' and 'stroke' the triggered video projections of

nipples, lips and keyboard buttons.The desk staff who issue the headphones quickly

noticed a very common misconception in the audience, and now carefully explainto each person that this is NOT 'a guide to the exhibition.' (Graham, 1994)

The problem with curating new media art is thatthe fascinating range of challenges is matched onlyby the dearth of data and material available to helpcurators.Comparatively, the field of museum inter-pretative and educational new media is wellresourced and debated. On seeing a piece of newtechnology in a gallery or museum, a member ofthe public is justifiably likely to assume that it issome kind of interpretative aid rather than an art-work in itself.

General new media theory seems to be in plentifulsupply; so much so, that the few accomplished newmedia artworks which manage to get produced tendto emerge staggering under the weight of eageracademics.When it comes to the practical issuesof presenting new media to the public, however,even Peter Weibel of ZKM admits that it is "diffi-cult not to make an error because there is not muchinformation ..." (Weibel, 2000). There have beensome conferences and some skirmishing on dis-cussion lists concerning new media art curating inparticular, but nothing like the regular, archived, pub-lished events devoted to the web as interpreta-tion.

It is this relative absence which the Curatorial Re-source for Upstart Media Bliss attempts to address,starting with the collection and presentation of somepithy information and opinion relevant to over-worked new media curators. CRUMB forages be-neath the tables of education, museum evaluationand media theory for tasty morsels, as well as goingstraight to the horses' mouths of curators and art-ists for useful experience and hindsightThe websiteincludes bibliographies, links, interviews with lead-ing curators, and some useful nitty-gritty concern-ing contracts, etc. The CRUMB website, and thispaper, draw from a wider research interest in newmedia curating at the University of Sunderland, andparticularly from the experience of the curators.Working both in new media and mixed visual artscontexts, between us we have experience in work-ing with budgets from 000 to £200,000, and in do-ing research in places from Banff and Minneapolisto Bangalore and Ljubljana. Having worked with in-stitutions from artists-run centres to local author-ity galleries, we have sympathy with various devils,and can commiserate with Barbara London whenshe says thatit's tough to get museums to change,to keep moving in new directions. In early days of



Graham and Cook, CRUMB: A Curatorial Resource for Upstart Media Bliss

video we didn't have access parcel post is howthey got around:' (Sins of Change 2000).

It is very much early days for any discussion of newmedia art curatorship. At times the debate seemsto move very fast, and at others it seems fossilized,with huge disparities in awareness and expectations.In Britain in 1997, the announcement of the TurnerPrize nominees unleashed a surprising flood of popu-lar press puzzlement that video could be consid-ered to be art. On the other hand, in India, a health-ily hybrid approach means that artists like ShilpaGupta (http://members.tripod.com/shilpagupta/) canmove between sculptural installation and Internet-based art fairly comfortably.The field of 'new', new-ish or upstart media is rather difficult to discusswhen neither the terminology nor the genres arefixed.The conference Computing Culture:Defining NewMedia Genres (1998) suggested that artefacts couldbe considered within the genres of Database, Inter-face, Spatialization and Navigation. Festivals and in-stitutions do seem to be starting to divide theircalls for work into rough categories of Net.Art,Single Screen, Performative, and some variety of in-teractive or non-interactive Installation (Public Artand 2D Digital Images being occasional orbitingcompanions).Whilst net.art has netted the major-ity of debate and controversy, CRUMB also coversphysical new media works in conventional galleryspaces.

This paper uses some resources from the CRUMBwebsite to consider some current debates forcurating new media art, falling under four headings:Artists,Archives,Audiences, and Aesthetics.

Artists

His idea was clear:a museum has to follow what

artists are doing. Art history has to follow artNot the opposite.Too much today the museumwants to prescribe what art is. (Weibel ,2000,referring to Alfred H. Barr of MOMA)

Curatorial decisions are made through a needto justify hardware and software investments.Artists are a test case. (Vuk Cosic, netart pio-neer, Sins of Change 2000)

Art and science institutions have sometimes invitedartists in to play with their equipment, only to find

them fundamentally challenging their whole valuesystems:

[new media art] practice challenges the no-tion of authorship, has to do with collectiveauthorship; non western ideas of discourseis something the museum has always hadtrouble with. And what has happened on thenet is a brain of a social collectivity, that al-lows discursive practice ... How do you sup-port and preserve a critical practice that isinclusive ... how can you do that when it isdifficult to pin down authorship? (Sara Dia-mond, Banff Centre for the Arts, Sins of Change

2000)

Artists, whether individually or collectively, have notonly presented museums with major conceptualheadaches, but also been implicated in major shiftsin how educational and commercial research insti-tutions think about what they do.As Lynn Hershmansays:"digital artists have to adopt interdisciplinaryways of researching" (Sins of Change 2000), not leastin order to get access to equipment. In the USA,forexample, the Xerox PARC experience (Harris, 1999)openly explores the creative conflicts between artand science research, whilst in the UK the develop-ment of 'art-practice-led' PhDs has been involvedwith a fundamental questioning of'what is research?'(Malins and Gray, 1999). It is the artists who makeand push the new forms of media art, even if theseforms are risky, challenging or an unfinished pro-cess:"... only the net allows us to make these worksin progress. If you are a curator it is a crime not touse it." (Olia Lialina, net.art pioneer, Sins of Change2000). Certain generous and brave curators havebenefited from the spirit of this openness by mak-ing their research trips public (such as BarbaraLondon's Japan journal http://media.moma.org/dot.jp/) or by sharing their process and knowledge(very notably, Steve Dietz's publications at http://www.yproductions.com/).

Concerning audience, archiving and economics, art-ists tend to ask the most difficult questions, such asEster Robinson's queries:"Who is paying for whatyou are doing? ... Does it live in a place that no oneshould care about it? ...Who has ownership? ...How does it give the audience sustenance?" (Sins ofChange, 2000). Of obvious importance to artists isthe whole question of how they can make a living.

1.91198 ©Archives & Museum Informatics, 2001


Olia Lialina has stated that"My contribution to thisdiscussion was the first net art gallery,Teleportica.It was to show that net artists are not the cheapestartists on the market." (Sins of Change, 2000). Shouldmuseums be paying`per screening' like video or cin-ema,? Should they be funding co-productions likefilms? How do web artists fit into the ExhibitionPayment Right (UK) or CARFAC schedules(Canada)? Media artists can't survive

... unless museums start to pay artists fairamounts for linking to their work.., amountsthat are comparable with what any otherartist showing in a gallery space in the mu-seum would expect to receive for a wholebody of work.., one net site is usually notthe equivalent to one painting ... more like amajor body of work [such as] a large in-stallation, a substantial video, a series of sculp-

tures. (Rackham, 2000)

Other artists may see the more popular debateconcerning payment and intellectual property as partof the commercialization of the Internet rather thanas an art debate per se (but these may, of course,be artists with other sources of income):

I'm always struck by how the mainstreampress is obsessed with the question of howInternet artists are ever going to make a buck... I recently plodded through stacks of re-views from the '60s of Fluxus artists smash-ing violins and pissing in buckets. At the timeplenty of journalists asked Why is this art?'or 'Is it theater?' or 'Is it any good?', but Inever saw a single article that raised thequestion of how these artists were going tomake a living. (lppolito, 2000)

The struggle for artists' access to money, resourcesand presentation facilities will no doubt continue,but as artists are the ones who are making the de-bate as well as the work, they may be "a test case"in a positive as well as a negative sense testing theinstitutions in unpredictable and productive ways.

Archives

How can we address the preservation of ideasas opposed to objects? (Weil, 2000)

Lev Manovich:I understand your position regard-

ing museums, art institutions, preserving,archiving databasing but it's so different fromthe Futurists who said,'shoot the painters, burnthe museum:Here we are the avant-garde

and we want to keep all the stuff (...) Maybewe should be looking towards the future.

Sara Diamond: It's different when a canon isbeing created, as opposed to a movement (Sins

of Change 2000)

It seems that every media art festival of the pastyear had a panel to address the question of archiving

often as art (for example, the panel at ISEA2000moderated by Karen O'Rourke with Patricia d'Isola,Christophe Le Francois, Eduardo Kac, GeorgesLegrady, and Lev Manovich), but also as a curatorialstrategy.Yet despite all the discussions there remainsan ambivalence toward the notion of archivingwe don't know what to keep yet, but we need to beable to refer to it.

We either let it decay online and that's fine orwe preserve the memory and find a form torepresent it in a manner that makes sense.Who

in this room has seen the Spiral Jetty? But we all

know about it, at least we have a sense of what

the intent was. Going backwards to understandhow this work was produced is a good lesson.(Weil, 2000).

... art critics have suddenly found pioneeringfigures in video installation, all from the early90s.There is in fact a half-century of pioneers,and now we have to get the word out, not justthe work out. ... the genesis of this conferencewas to address the amnesia of recent art critics.

You must get up and think about them [theearly media artists], teach them, write aboutthem, make people watch them. (Bruce Jenkins,

film and video curator at Harvard University,Sins of Change 2000).

Furthermore, we're aware that technologically, we'reusing media which have, in Bruce Sterling's words,"the life-span of a hamster" (1995). Showing thework in order to keep it alive, in order to make itpart of a canon, is in fact destroying the work. In thecase of film and video, the work is deteriorating aswe watch and learn about it.

192@Archives & Museum Informatics, 2001 1 99


So, if we can't count on being able to keep the ob-solescent technology, we need to be able to keepthe intent, the words.Yet the question of archivingeven the discourse which surrounds new mediapractice (which we anticipate will be useful to fu-ture curators) is itself debated.For every book pub-lished by MIT press, for example, there are a thou-sand e-mails on listservs across the globe that getread and deleted. Jennifer Crowe is constantly re-vising the guidelines for submissions of projects andinformation to the Rhizome database (http://www.rhizome.org). Then there's the question ofwhether the discourse exists in the first place:

... people in art worlds didn't know how tolook at my work, or treat it So I wrote myown reviews (under the pseudonym of) Pru-dence Juris. The reviews would talk aboutand argue about the work. Then I would takethose articles and show them to the galler-ies to develop my own credibility. You haveto create the language yourself to promote,historicize your work. Just doing the workisn't enough. You have to create preserva-tion on your own. (Lynn Hershman, artist,Sins of Change, 2000).

E-mail has made even the museum tradition of fil-ing correspondence with artists more complicated(few people in the museum world are consistentlydiligent enough to print and file a letter about themaking or exhibiting of even a painting, let alonethe details of a web commission).

The Walker and other museums have been ques-tioning the feasibility of offering open source, sharedserver space to artists. Oliver Grau at the Univer-sity of Humboldt in Berlin (http:// www.arthisthu-berlin.de) is developing a database of virtual art thatdocuments not only the work of art but also all itsdifferent versions, each time it was presented,whatthe publicity/criticism of its exhibition was, even whothe technicians who worked on it were.This is anacademic (and highly theoretical) endeavour, butonce online could be a model for the preservationof ideas as well as of the documentation of inher-ently transitory work.

This, after all, is the mandate of the ZKM. But onthat front there is also the curatorial dilemma ofthe authority of authoring through the practice of

archiving. How many future new media curators willsee the exhibition net_condition produced by theZKM as the official history of net.art? Why aren'tHeath Bunting and Irationalorg in the show?

They [ZKAI] had a show on Internet art I'd

like this to go on the record very, very late in

the game. They came very late to discoverInternet art.They trailed on the coattails of other

curators.They picked up and accumulated thechoices of other curators. They accumulatedthem in their show. (Huffman, 2000)

Sarah Cook:So then what is the future of a newmedia curator?

Peter Weibel: to protect media art against thetakeover of the historical art world. Seriously.It's not an easy job.

Sarah Cook: But the historical art world isfounded in part on museums where they collectart That is why you have new media institutions

like the one we're in now, ZKM, which also col-lects, so how is that protecting media art fromthe art world if museums the world over arecollecting new media art?

Weibel:By two things. First by emphasizing pro-

duction, of contemporary, risky, you ng artists, and

then by preserving the work which is discounted

and marginalized by the art institutions. (Weibel,

2000).

In the field of new media art, the canons are begin-ning to appear amongst the hand -to-hand combatconcerning archiving. Dust may continue to obscurethe view for some time, but if we archive docu-ments and records of the early exhibitions as wellas recording the work itself, the history of newmedia may perhaps be usefully retained. By utilisingthe possibilities of new media,the archives may evenbe creative tools in their own right.

Audience

Who wants to walk around a gallery if it's full oftoffee-nosed elitists? (lames Bloom of Wiredmagazine at the First UK Internet Art Festival,quoted in Cavendish, I 995, 23)



If interactive art simply mirrors the game its

themes and values it becomes symptomaticof uncritical postmodernism where there is nodifference between entertainment and art,where consumerism reigns. And when, loadeddown as amusement, it knocks on the museum

door, it insists on altering how and why muse-ums function, further institutionalizing art asconsumer fun. (Cornwell, 1933, 12)

In the early nineties, the utopian excitement aboutnew and different audiences contrasted sharply withfears that the allure of 'hands-on technology fun'might be deeply implicated in the Disneyfication' ofmuseums. Some time later, the debate is perhapsless polarized, but still rather contradictory.ShankarBarua (2000) summarised a situation that has inter-national resonance: "In India the audience for artgalleries is a thin elite.The audience for new mediaart is also an elite, but a different one."

Is the audience for net.art a new audience or justthe regular art audience logging on when they shouldbe doing something else? How does the audiencefind the site? How does the number of hits to awebsite relate to actual use, benefit and pleasure?Is it more about participation than audience any-way? And how about hacktivise net.art, which maybe deliberately anonymous and covert in its baitingof multinationals, rather than seeking an audience?For all the internationalist rhetoric, how does theInternet cope with the practicalities of cross-cul-tural art communication? For all of these questions,certain exhibitions and artefacts provide particularexamples, but there is much less information avail-able from curators and artists -about behind-the-scenes data, illuminating failures, or audience feed-back.Those who design museum interpretative siteshave been gathering information for some time onthose who use the sites, how and when.Artists mayhave very different parameters for judging the 'suc-cess' of their sites, but some data on who visit, andwhether they stay for two seconds or two hours,might be useful, even if only to discover how bestto shock and appall.

As for new media art installations displayed in con-ventional gallery and museum spaces, there are some

scattered items of information concerning audiencedemographics. For example, for one exhibition ofinteractive installations in a regional British art mu-

194

seum, the show, when compared to other contem-porary and historical art exhibitions in the sameyear, showed little difference in gender numbers,but a significantly higher proportion of visitors fromthe "under-20" age bracket (Graham, 1997, 102).

Gathering demographic information is one thing,but judging the subtleties of audience interaction isanother.The entry of new media art into museumswas very much spearheaded by 'the romance ofinteractivity'.

The word interactive sounds like it will alleviate thealienation of modern life by generating a dynamicalliance between artists and their audiences, joiningthem together in a splendid waltz that lets viewersbecome equal partners with artists in creating art(Wooster, 1991, 294). Since then, the romance hasbeen tempered by some more critical views onexactly how interactive artefacts are:"For a multi-media program, a human audience is just a randomnumber generator." (Cubitt, 1999) The knowledgeof those designing educational exhibits has also be-come pertinent:"To interact is to act reciprocally,to act on each other.... not merely a machine thatthe visitor operates ...'Non-interactive mechanisms'perhaps sums them up adequately?' (Miles, 1988, 95).Although artists are understandably wary of 'audi-ence evaluation', they may be interested in, for ex-ample, the results of Stevenson's (1994) researchinto hands-on science exhibits which indicates thatthere are significant impacts on the long-termmemories and understanding of the audience, ratherthan merely a case of'running around pushing things'.

Some of the commentary on the audience's use ofinteractive artworks in galleries has come from crit-ics:"... you really need an hour alone with the thing,which is impossible under the circumstances of ev-eryday museum attendance?' (Coleman, 1994, 14).Some has come from artists themselves, such asDavid Rokeby's detailed observations on"command"gestures versus "tentative questioning gestures" inhis interactive art systems (1995, I 48).Weinbren(1995) and Feingold (1995, 401) have also madeuseful observations on 'control' and 'mastery' in theirown artworks. Research into interactive artworksin galleries (Graham; 1997, 1999) expected inter-esting gender differences, but uncovered instead asurprising tendency for groups to want to use art-works together, even when this meant squashing



uncomfortably into spaces designed for an individual.This led to a particular interest in artworks whichencourage interaction between audience members,rather than solely between artwork and audience.

There is still a great deal of uncertainty amongstcurators (and the audience) as to what kind of ex-perience is being offered by new media art: Quick-fire game pleasure? Information pleasure? Sculpturalpleasure? Sit-on-a-hard-chair-and-watch-a-videopleasure? (Graham, 200 1). As with the Audio Zoneartwork using infra-red headphones, this confusionmay be used creatively, but the incredibly diverserange of expectations about new media art meansthat it's more than usually important to try to getsome kind of feedback. Matthew Gansallo (2000),when interviewed about the Mongrel commissionfor the Tate (http://www.tate.org.uk/webart/), re-vealed a wide range of reactions:

... we got a lot of responses back: 'who did it?'and 'how scaryr, and 'is this a site for knowl-edge?'and 'what is it?' and 'anyone can put thistypically Tate shock tactics crap (up]r And wegot responses that were 'gosh this is interesting,'

and 'I'm glad they've done this,' and 'it's goodthe Tate is large enough to say what you want.'

Whoever the target audience for new media artmay be, and whatever the feedback, Robert Atkinshas some pithy parting advice:

The Top Five strategies for overcoming sloth inyour artistic/curatorial practice in relation to theissue of having your audience gain access to the

work:

5. Don't think technological barriers will fallthere are ever-increasing barriers of technology

business ...

4. Focus your energy on technology that is be-ginning to arise, something not too far away ...

3. Don't lose sight of producers and of qualityfor audiences; the more you watch the less youknow.

2. Do think hybrid, not just hardware or soft-ware. Digital culture can be simultaneously many

things at once;producers and audiences can both

185

be content creators ...The art-world mix of pro-ducers [can create] another form of knowledge

from media into media art. (i.e. Muntadas'Archive Project and mediachanneLorg)

I. Don't censor yourself, don't wimp out, don'tthink anyone knows more than you do ... Find

niche opportunities ..."

(Robert Atkins, art historian, writer, media con-tent producer, Sins of Change 2000)

Aesthetics

User experience is what art does best.To change

the interface is to dramatically change the work.

(Lee Manovich, artist and theorist, Sins of Change

2000)

It's only under huge pressure that a visual artscurator would agree to hang a video projector,and only if it is agreed the projector will projectan image on the wall and take us back to paint-

ing. Only under threat of torture will a visualarts curator put a computer in the galleries.(PhilippeVerge,WalkerArt Center , Sins of Change

2000)

The aesthetics of new media art is easily ignored infavour of its function. Moreover, the way a piece ofnew media art looks and holds up within the realmof aesthetics is usually masked behind a whack ofmedia and communication theory about networks,about spectacle, about invisible architectures.Aes-thetics, as a philosophy, is an old-fashioned one, andcertainly not one taught in most media schools. Itis, however, a philosophy taught to curators. Hegeltaught that to each age there is an art form, frompainting to sculpture to the architecture of thetemple itself; from this we learned not to see art asseparate from the age in which it was produced.New media's very nature interactivity being oneaspect has demanded of curators different crite-ria for the aesthetic evaluation of the works.WouldHegel have placed net.art as the highest art formfor our age?

In his article "The Death of Computer Art", LevManovich (1996) made a distinction between twoworlds where art is made and shown, but where



the social definitions of art differ enormously.Thefirst he called "Duchamp-land" and defined it as:"gal-leries, major museums, prestigious art journals"; thesecond he called "Turing-land" and defined its pa-rameters as "ISEA,Ars Electronica, SIGGRAPH artshows, etc." Manovich's point was that the twoworlds would not converge, that we should notexpect what is being made in Turing-land to beshown in Duchamp-land. The reason for this, hesurmised, was that the typical object admitted toDuchamp-land prioritizes content and fits within adiscourse of irony, self-referentiality and other thingsgenerally postmodern, but that the objects beingmade in Turing-land are not ironic because they areoriented towards ever-new technology rather thancontent, they are simple in ideas,and they take theirtechnology very seriously.

This distinction is still valid. Manovich establishedcriteria for the type of art accepted in each land, butwhat he didn't mention then, and what has becomeparamount now in 2001, are the criteria for art tobe shown in each land namely, the questions ofexhibition practices and the field of curatorship. Forart to be shown in Turing-land, it need only apply.Allthat is needed is some money, a truck, a power-bar,a plug, a network connection plus a space in a fair ina convention center, or maybe none of that but ftpand a server. For art to be shown in Duchamp-landit has to have the interest of a curator, and withthat comes the backing of an institution, a board ofdirectors, a funding body, an intellectual mass. Un-like the art made in Turing-land, art in Duchamp-land goes through some process of curatorial le-gitimization before it is shown.Curators make regu-lar visits to Turing-land in order to find things tobring to Duchamp-land (that's how Lev Manovichended up in the Walker Art Center one of thepowerhouses of Duchamp-land).So in that respect,it might have very little to do with what the work isactually about.While the two worlds may not haveconverged, the way curators move between themhas had an effect on the work being produced ineach. Where do you think curators have learnedthe communication theory behind which they canhide the lackluster aesthetics of much of new me-dia art?What is found readymade inTuring-land canalways be signed and exhibited in Duchamp-land.

And has this curatorial "research" had an effect onthe way the work is presented in each?You betcha.

Curators beware:`changing interface' can change thework. One only has to think of the different experi-ence of visiting a convention center from visiting awhite cube to realize that the interface is the inter-action.The same is the case on the web seeing artin the context of a commercial product-based siteand seeing it in the context of a cultural content-based site can drastically affect how that work isinterpreted.

The best recent example of this was Vivian Selbo'sdesign for the exhibition Art Entertainment Network(http://aen.walkerart.org). Creating a customizableinterface for the works on view was the artwork initself.There was a place to shop, a place to read, aplace to listen, a place to chat ... all things we ex-pect from our web-experience, whether commer-cial or cultural.These types of interfaces are clearlyneeded as curators have recognized that Duchamp-land-type media art simply doesn't fly in any placewhere its irony and content-based form is subsumedby the pressures of the bottom line.

Though who says aesthetic culture isn't commer-cial anyway? Obviously it always has been. Never-theless, the curating of media art into Duchamp-land has caused a number of Turing-land artists toreconsider how they want their work to evolve inthe world. Why should it be self-referential andpostmodern just to get the money to be shown inDuchamp-land when now they can get a lot moremoney by masking the irony and selling their skillsto the market (which is neither Duchamp nor Tur-ing aware) instead?As such, a number of web-basedartists have turned the lion's share of their atten-tion away from simply making works of net.art tostarting dot.coms:Vuk Cosic and the Slovenian start-up Literal (http://www.literal.si), the boys from etoyselling shares as art (http://www.etoy.com),hans_extrem and his self-professed very expensiveconsulting company, Ubermorgen (http://www.ubermorgen.com).

This has created a third problem and brings us backto the beginning of the vicious circle. How do cura-tors present this type of overtly commercial andyet slyly Duchampian practice when museums struc-tures have traditionally dictated that the art shownis not explicitly commercial, and when at the sametime the commercial galleries won't touch newmedia art with a ten-foot pole because they can't

1.96


see its inherent ephemerality as an investment?Should they even exhibit it? Until Turing-land andDuchamp-land (and now perhaps Monopoly-land)get themselves sorted out, the answer seems to bethat curators tend to go back to their philosophiesof aesthetics and mask the work behind moretheory: beauty with seduction always sells.

In Conclusion

The reason curators expose themselves to the con-fusion and challenge (and carping) of newmedia artmight be that the whole mess is undeniably fasci-nating and offers opportunities for trying a boldexperiment,and getting it right (or more right thananyone so far). CRUMB aims to provide some in-formation which might help avoid the alternative ofgetting it spectacularly wrong (in a way that some-one else has previously got it wrong). We aim toextend the life-span of the media art hamster byfilling our cheeks with crumbs dropped by curatorsand others, and all thanks should go to those whoshare the knowledge (Steve Dietz and Sara Diamondparticularly), and to all those who agreed to be in-terviewed about awkward subjects.

The Museums and the Web conference is one ofthe few occasions on which educators and archi-vists get to meet artists and art curators and toshare information from their varied experience.Wehope that people will contribute to the CRUMBsite and discussion list, and add to the public bodyof knowledge concerning new media curating.

References

Barua, Shankar. (2000). Unpublished interview by BerylGraham, 4 December, New Delhi.

Cavendish, Dominic. (1995)."The virtual state of theart." The Independent, 27 March. 23.

Coleman, A.D. ( I 994)."Letter from New York." PhotoMetro (Jan). 28-29.

Computing Culture: Defining New Media Genres (1998).San Diego: University of California San Diego.Available from URL: http://jupiter.ucsd.edu/culture/main.html

Cornwell, Regina. (1993)."Interactive art and the videogame: Separating the siblings:' Camerawork (SanFrancisco), 20 (1). 10-13.

Cubitt, Sean. (1999). Presentation at Matrix(art researchconference). July. London: Central St. MartinsSchool of Art.

Feingold, Ken. (1995)."Ou: Interactivity as divination asvending machine:' Leonardo, 28 (5). 399-402.

Gansallo, Matthew. (2000). Unpublished interview bySarah Cook. December. Forthcoming editedtranscript on URL http://www.newmedia.sunderland.ac.uk/crumb/

Graham, Beryl. (1994). Unpublished notes fromobservational case study of the exhibition V-topia,July-Sept 1994 at The Tramway, Glasgow.

Graham, Beryl. (1997).A study of audience relationshipswith interactive computer-based visual artworks ingallery settings, through observation, art practice, andcuration. Unpub. Ph.D. thesis, University ofSunderland.

Graham, Beryl. (1999)."A study of audience relation-ships with interactive computer-based visualartworks." Leonardo, 32 (4).326-328.

Graham, Beryl. (forthcoming 2001)."Fun + pleasure:Games and interactive computer-based art inmainstream galleries." In: Erkki Huhtamo (ed.)Mariosofla:Elektronisten pelien kulttuuri (Monolog:Electronic game culture). Helsinki: Gaudeamus.

Harris, Craig (ed.). ( I 999). Art and Innovation:The XeroxParc Artist-In-Residence Program. Cambridge andLondon: MIT Press.

Huffman, Kathy Rae. (2000). Unpublished interview, inconversation with Julie Lazar. Forthcoming editedtranscript on URL: <hap://www.newmedia.sunderland.ac.uk/crumb/>

Ippolito, Jon. (2000). EAT digest. [Online] Digest 8 March2000.Available from URL: <http://www.walkerartorg/salons/ead>.

Malins, Julian and Gray, Carole. (1999)."The digitalthesis: Recent developments in practice-basedPhD research in art and design:' Digital Creativity,10 (I). 18-28.

197

Manovich, Lev. (1996). The Death of Computer Art.[Online].Available from URL: <http://www-apparitions.ucsd.edu/manovich/text/death.html>OR <http://www.thenetnet.com/schmeb/schmeb12.html>.

Miles, R.S. (1988). The Design of Educational Exhibits. 2ndedition. London: Unwin Hyman.

Net_condition. (1999). Curated by Peter Weibel, ZKM,MECAD, and NTT: Sept. 23 Nov. 19, 1999.[Online] Available from URL:_http://www.mecad.org/net_condition

Rackham, Melinda. (2000). EAT digest. [Online] Digest 8March 2000.Available from URL: <http://www.walkerart.org/salonskat/>.

Rokeby, David. (1995)."Transforming mirrors: Subjectiv-ity and control in interactive medie In: SimonPenny (ed.) Critical issues in electronic media. NewYork SUNY.

Sins of Change: Media Arts in Transition Again (2000).Conference at the Walker Art Center, Minneapolis,April 2000.Transcriptions of panel discussions bySarah Cook

Sterling, Bruce. (1995)."Dead media:' Presentation atInternational Symposium of Electronic Arts. Montreal:1SEA 95.

Stevenson, John. ( I 994)."Getting to grips." MuseumsJournal (UK) May. 30-32.

Weibel, Peter. (2000). Unpublished interview by SarahCook, September. Forthcoming edited transcripton URL: http://www.newmedia.sunderland.ac.uk/crumb/

Weinbren, Grahame String. (1995),"Mastery: Computergames, intuitive interfaces, and interactivemultimedie Leonardo, 28 (5). 403-408.

Wooster,Ann-Sargent. (1991)."Reach out and touchsomeone:The romance of interactivity:' In Halland Fifer (eds.) lluminating video. New York:Aperture. 275-303.

.08@Archives & Museum Informatics, 2001 205

Maintainingthe Magic ...

199

The Virtual Aura - Is There Space for Enchantment in aTechnological World?

Susan Hazan,The Israel Museum, Jerusalem, Israel

Abstract

As Walter Benjamin described in his famous essay,"The Work of Art in the Age of Mechanical Reproduction", the roleof art in society and the notion that art has become modified through mechanical reproduction has engaged not onlyartists, but also curators and the museum public. Benjamin embraced the severing of the quasi-mystical 'aura' from theoriginal as a potentially liberating phenomenon, both for the reproduction of works of art and for the art of film,thereby making works of art widely available, introducing new forms of perception in film and photography andallowing art to move from private to public, from the elite to the masses. While the loss of the aura for Benjaminrepresented new possibilities, what was forfeited in this process were the 'aura' and the authority of the objectcontaining within it the values of cultural heritage and tradition.This paper evaluates the different ways that museumsare responding to life on the Net.and will look to three models of museum Web-sites: the documentation of traditionalcollections through online databases, the virtual museum with no concrete counterpart to resonate the online experienceand the proliferation of Web based contemporary art. This attempt to map out the different ways that museumsformulate their identity on the Net will address the notion of the lost aura or perhaps the emergence of new culturalphenomena, the virtual aura.

There are as many reasons to seek out an onlinemuseum as there are to visit the real museum, andmuseum Web sites are addressing a wide range ofthese experiences.While some Web sites demandthat we sit upright in our chairs for a traditional,pedagogical kind of experience, at other times, weare less focused and enjoy a more passive, relaxedkind of interaction with the screen. In much thesame way as we may watch a film or a televisionprogram in a state of distraction, we act as observer,but an absent-minded one.As we sit in front of themonitor and connect to the dancing images in frontof us, we allow electronic images and sounds toradiate from the screen and seep into our innerworld. In a post-modern, wired society, the onlinemuseum brings the virtual image from remote lo-cations into the foreground.This paper will investi-gate different models of online museum and thecultural processes involved when objects in theconcrete museum are not only represented by sur-rogates, but also have been liberated or dispelledfrom their rhizome and distributed in electronicpackets to remote locations in a networked world.

This paper evaluates the different ways museumsare responding to life on the Net, and will explorethree models of museum Web sites.The first modelprovides a metaphor for the traditional museumwhich provides access to digital collections throughonline databases such as the Thinker ImageBase,(http://www.thinker.org/) of the FineArts Museumsof San Francisco (the de Young Museum and the

Legion of Honor), USA. In the second model, thevirtual museum metaphor is explored in MUVA,El Pais Virtual Museum of Art (http://www.diarioelpais.com/muva2/).The El Pais site ex-emplifies a virtual construction which maintains atenuous base in reality. For the third model I haveselected two further web sites, both net art projects:Uncomfortable Proximity (http://www.tate.org.uk) at the Tate Modern Web site, and Geist,(http://www.mediascot.org/geist). Where the 'real'Tate provides floor plans, the Tate Mongrel project,created by Harwood, takes us under the floor ofthe Tate.The Geist Project, by Colin Andrews of theNew Media Group, Scotland, presents an eerieWebsite involving ghosts.

This paper will attempt to map out the differentways museums formulate their identities on the Netand will address the notion of the lost aura or per-haps the emergence of a new cultural phenomenon,the virtual aura.

The Age of Mechanical Reproduction

Walter Benjamin, writing against the backdrop ofthe Nazi era in 1936, in his famous essay,"TheWorkof Art in the Age of Mechanical Reproduction", de-scribed the role of art in society and the way inwhich art had become modified through mechani-cal reproduction. Benjamin embraced the severingof the quasi-mystical `aura' from the original as apotentially liberating phenomenon. Making works

200 BEST COPY AVAILABLEMuseums and the Web 2001:Proceedings


Hazan,The Virtual Aura

of art widely available opened new forms of per-ception in film and photography, and the accessibil-ity of art could move from private to public, fromthe elite to the masses.While at the same time ques-tioning the need for authenticity, Benjamin wel-comed the close-ups and slow motion of the mov-ing image in that they opened up new art valuesthat were no longer so dependant on cult values orritual.Thus Benjamin's work was seminal in bringinginto focus the notion of art as politic.This insight,according to Benjamin, meant that

for the first time in world history, mechani-cal reproduction emancipates the work ofart from its parasitical dependence on ritual.(Benjamin: 1935, 1992).

What had been forfeited in this process, were the'aura' and the authority of the object, scarred, yetalso embellished with the patina of time and pris-matic with the marks of human endeavour. It wasthe aura that contained within it the values of cul-tural heritage and tradition. Even though for Ben-jamin, the loss of the aura meant the loss of theoriginal, the transformation or liberation of the artobject to the ordinary represented a gain. For Ben-jamin, what had then replaced the original at thattime was the illusion of the moving image, and theduplication of the photograph. For post-modernsociety, it has become the digital image.While Ben-jamin celebrated the magical aura that had beenforfeited as a liberating phenomenon, one cannot helpbut speculate whether there is still a need for aspace of wonder or enchantment in a technologi-cal world. Perhaps society still craves such a space,now more than ever, and seeks it in extraordinaryplaces, such as in the museum. If so, then can thislost aura be compensated for or reconstituted inany way in a virtual environment in a networkedsociety?

Click to Go!

The speed with which we are able to access re-mote museums and pull them up side by side onthe screen is alarmingly immediate.We do this atthe click of a mouse, and in a nano-instant of time.Scott Lash and John Limy, (1994), argue that in thewake of organized capitalism, the flows of objects(goods, capital, money, communications, commodi-ties) as well as subjects (labor, immigrants, tourists)

;.,) 01

are accelerating in ever widening trajectories. Ac-cording to Lash and Urry, this has disturbing impli-cations for society. As social relations are`distanciated', it 'compresses' time and space, and isleading to an emptying out of both subjects andobjects.This accelerated mobility causes objects tobecome disposable and to decline in significance,while social relationships are emptied of meaning(Llash and Urry, I 994).With one click we can ac-cess an entire museum.We are able in one glimpseto visualize in miniature all of its multi-functionsrepresented on its home page. While there is nodoubt that this provides efficient and meaningfulinformation for remote visitors regarding potentialvisits, should we query what kind of experience visi-tors gain when they view the digital collections rep-resented as tiny, two dimensional, electronic reflec-tions of the original works? Should we concludethat this medium is orientated exclusively towardsinformation and interpretation, or could there bemore to it?

Anthony Giddens, (1994), argues that in post mod-ern society there has been a separation of time andspace, and their recombination in forms which per-mit precise time-spacing "zoning" of social life, butwhich has consequently led to the disembedding ofsocial systems (Giddens, 1990). Giddens points tothis severing of time as the cause of place becomingincreasingly phantasmagoric in that locales are thor-oughly penetrated by and shaped in terms of socialinfluences that are, in fact, quite distant from them(Giddens, 1990). Giddens emphasizes that

what structures the locale is not simply thatwhich is present on the scene; the "visibleform" of the locale but what is present isalso the relations concealed in the distance,and it is the remote relations that is deter-mining the nature of the local (Giddens,1990)

For example, the city with numbered houses basedon boulevards and a grid structure, in which thehigh street has its Benettons,McDonalds,and Nexts,is more abstract, more emptied out than the pre-modern city of winding streets and numberlesshouses.The remotely accessed database of museumcollections seems to offer a similarly disembeddedexperience.The vast distance covered in mere nano-seconds acts to dissolve the concrete-ness of the



real museum in front of our eyes, resulting in anemptying out of the traditional visitor/museum ex-perience and the disembedding of cultural systemsreflected in the museum.

Living In-Vitro

In post-modern society, our participation in thepublic sphere, understanding of current events, en-tertainment, and life long education have come todepend more and more on mediated resourcesrather than first hand,getting our boots dirty experi-ence. The second hand, virtual narratives, eitherthrough television or digital interaction, cause usnot merely to reflect on these experiences but toactively construct our daily lives through them.Wecannot be physically present at every national cel-ebration, and we do not want to be present in awar-zone.We are content to let the camera be oureye and the anchorman our mouthpiece. Wheremuch of our life is lived through mediated ratherthan first hand experience, much of our daily inter-action is becoming more vitreous than visceral. Overthe last 40 years, most of the world's populationshave spent countless hours watching the world invivo, in vitro on screens in their living rooms, bed-rooms and classrooms. We are content to watchlive sporting events from the comfort of an arm-chair; we receive the daily fix of news on the allot-ted time slot; and we are faithful voyeurs of otherpeople's lives, some real, some not, playing out onweekly dramas on the screen. Marc Auge remindsus of

the false familiarity the small screen estab-lishes between the viewers and the actorsof big-scale history, whose profiles becomeas well known to us as those of soap-operaheroes and international artistic or sportingstars (Auge, 1995).

On the screen, the remote players become minia-turized in our own personal microcosm,and as thesetiny, yet familiar images flicker into our internal vis-tas, they penetrate our lives just as potently as otherdaily interactions and are often just as persuasive intheir messages.How often have we put down a bookonly to be rudely awakened to the reality that weare sitting in our favorite armchair? How is it thatwe are shocked by the brutality of the house lightsturning on as the last scene of a film fades away

from our inner eye? Margaret Morse, (1998), de-scribes this state of distraction as the fiction eftbct;that is, the partial loss of touch with the here andnow, a state of distraction experienced on the free-way, in shopping malls, and in television viewing(Morse, 1998). Can digital images and messages bejust as compelling as the magic of literary spaceand cinematic experience? Can the tiny electronicstage, set for one person and for one person only,be a convincing space of enchantment?

Seeking the Sacred Space

According to Michel Foucault, (1964) we live ourlives within many kinds of intersecting social rela-tions that overlap yet remain discrete. He arguesthat while social space has been moving towardsde-sacralization since the time of Galileo, this pro-cess has been mainly theoretical.Where once therewas a hierarchy of sacred and profane spaces, asFoucault describes, "all laden with qualities andhaunted by fantasy", in practical terms, both in pri-vate and public space, relations are still being con-trolled by an unspoken sacralization (Foucault,1964). In order to fulfill this desire for the sacred,contemporary society seeks to define spaces, sepa-rate from mundane, everyday living. Foucault de-scribes these spaces as utopias, as spaces havingno real place, as fundamentally and essentially un-real where they act as an analogy with the real spaceof society. However, according to Foucault, everycivilization creates real places, actual places, thatserve to stage experiences, and consequently setsthem aside for extraordinary action. DespiteBenjamin's celebration of the mechanical reproduc-tion that caused the separation of art from cult val-ues and ritual, we cannot deny that there is still aneed for the sacred in post-modern society. Theliminal spaces that Foucault calls heterotopias,while based in objective reality, act as the mirrorthat reflects.While this reflected space may be con-crete, in that it exists in a real location, its socialfunction at the same time serves to provide societywith an abstract locale to act out experiences at atime of crisis; a locale for adolescents, menstruatingwomen, women in labor, old people and so on (Fou-cault, I 964).A derivation of the heterotopian space,according to Foucault, is the heterochronia of timethat accumulates indefinitely - for example, muse-ums and libraries.



Museums and libraries are heterotopias inwhich time never ceases to pile up and perchon its own summit, whereas in the seven-teenth century, and up to the end of theseventeenth century still, museums were theexpression of an individual choice. By con-trast, the idea of accumulating every-thing,the idea of constituting a sort of generalarchive, the desire to contain all times, allages, all forms, all tastes in one place, theidea of constituting a place of all times thatis itself outside time and protected from itserosion, the project of thus organizing a kindof perpetual and indefinite accumulation oftime in a place that will nor move - well, infact, all of this belongs to our modernity(Foucault, 1964).

If we follow Foucault's argument that the museumis already an exceptional space, set aside by societyfor extraordinary activity, when we view this alreadyfabricated construction remotely through the glassof the television or computer monitor, we surelywill encounter a further dislocation. In order toseparate technological determination from aestheticexperience, it would be useful to look to AlfredGell's approach on what it is that makes an art-work compelling in the first place.

Enchanting Technology

In his essay, TheTechnology of Enchantment and theEnchantment of Technology' (1992), Gell attemptedto evaluate this mystical yet nebulous trait from asocio-anthropological perspective. He looked to thetechnological processes of evolution of the art-work to ascertain what it was that was inspiringenchantment in the beholder. Gell determined anessential property imbued within the art object thatserved as a component of social systems. In hissearch for an overarching anthropological theoryto explain not only Western art, but what he termsnon-Western art, he implied that art objects can-not be evaluated when dislocated from their socialsystems, but need to be seen in light of their circu-lation in society. In this circulatory format, art isthought of as agency, invoking a series of responsesor interpretations perceived as indexes of relationsto the object,the artist,and the recipient (Ge11,1998,in the foreword by Thomas, ix). Gell does not denythe aesthetic component of art, but considers art

0

as a technical system producing social consequencesin much the same way as a religion or belief systemweaves the social fabric of a community.This he callsthe technology of enchantment and proposes that art-systems be considered networks in society where.neither the individual creator, nor the agent, northe enchanting art object itself, can be separated fromthe collective process. This over-arching technicalsystem, contributing to the dynamic of culture, em-bodies a mystical process that"casts a spell over usso that we see the real world in an enchanted form".

The technology of an art object differs from otherobjects valued by society, such as beautiful horses,beautiful people, (Gell: 1992, 43) in that it is manmade, emerging from the alchemy of the occult tech-nician, and reflects what Gell defines as the 'enchant-ment of tech nology.'The magic, or the aura of the artobject, is brought into being through the techno-logical process itself. He illustrated the universalattributes of technology common to all categoriesof man made art objects, including classical oil paint-ings, sculptures by Picasso and Matisse, as well asthe canoe prow-board of the Trobriande Islands ofPapua New Guinea (Gell: 1992, 44). Gell empha-sizes the technological process whereby all of theseartefacts came into being and suggests that these'outcomes of the technical process' have all beenproduced by magical means. It is through this magicthat the object evokes a profound sense of awe inthe beholder, a potent visceral response that Gellidentifies as the 'halo-effect' and Benjamin wouldconnote the aura. But how does the artist/sculp-tor/carver produce this magical 'halo-effect'? Gellreminds us in discussing J.F. Petos' Old Time LetterRack, 1894, that

People have great difficulty in working outhow colored pigments (substances withwhich everybody is broadly familiar) can beapplied to a surface so as to become an ap-parently different set of substances, namely,the ones which enter into the compositionof letters, ribbons, drawing pins, stamps, bitsof string, and so on (Gell, 1992, 49).

Gell acknowledges that this mystical process doesoccur, but questions if it is simply a matter of a fas-cination with colored pigments.As Gell goes on toexplain, it is in this process, the way in which theobject has come-into-being, which eludes us, and

2 I 2 © Archives & Museum Informatics, 2001


forces us to interpret it as magical (Gell: 1992, 49).Gell compares other technological processes ashaving more or less prestige and cites the lowly but-ton-pressing of the photographer as having 'no pres-tige' unless there is perceived to be virtuosity in thephotographic process and sophistically producedimage. Only then would the photograph reflect askill that exists beyond the beholders' capabilities.Gell identifies 'concept' art within his 'interpretive'theory of art, as being more attuned to the realitiesof the present-day art world which, according toGell, has long abandoned the making of 'beautiful'-looking pictures and sculptures in favor of 'concept'art. Gell points to Damien Hirst's dead shark, in atank of formaldehyde,"The Impossibility of Deathin the Mind of Someone Living", 1992, as,

an object that could not be called appealing,nor a work of any excellence in terms ofcraftsmanship. But Hirst's shark is a highlyintelligible gesture in terms of contempo-rary art-making and thoroughly grounded inthe post-Duchampian tradition of conceptart, and, as such, is capable of being evalu-ated as good art, bad art, middling art, butdefinitely art of some kind (Gell, 1996).

Fredric Jameson, (1992), describes conceptual artas "perceptual paradoxes that we cannot think orunravel by way of conscious abstractions and whichbring us up short against the visual occasions".Jameson compares conceptual art to some of theclassic texts of deconstruction and anchors con-ceptual art to the spatial relationship of the galleryspace. The art form creates a material pretext tomake a mental circuit from the institution, into itsnetwork of trustees, their affiliations with multiplecorporations, and finally the global system of latecapitalism proper (Jameson, 1991).

In trying to explain the source of this magical po-tency, Gell describes how the Trobiande carversspiritually prepare themselves in order to overcometechnical obstacles in the carving process.Only thencan this magic be transposed through channels ofexchange into the perfect 'performance' of the ca-noe-board, allowing it to slip effortlessly throughthe water.Carvers undergo magical procedures whichopen up the channels of their minds, so that boththe carved designs and the canoe will flow smoothly.This was accomplished, according to Gell, by means

of the metaphoric use of water and other liquids,especially blood and bespelled betel-juice (Gell.1992). Gell compares this moment to a musician inour culture getting technically prepared to give aperfect performance of an already existing compo-sition, such as the 'Moonlight' Sonata (Gell, 1992).Gell acquiesces to Duchamp's accomplishments, inspite of the banality of the urinal, and recognizesthe alchemy of the artist, in that he had the capacityto transform, which Gell describes as

The essential alchemy of art, which is to make whatis not out of what is, and to make what is out ofwhat is not (Gell, 1992).

Museumification

The visitor in a museum is responding to culturalprocesses that are reflected through the trajecto-ries of the aggregated and contextualized objects.These sometimes precious, sometimes mundaneobjects are modified in the exhibition context wherethey go through a process of museumification, ex-tracted from distant locations and placed on a spot-lighted pedestal, or isolated in a glass cage. Theybecome re-conceptualized and re-contextualizedand serve to petrify cultural values in much the sameway as the theatre projects the human conditionthrough metaphor and allegory.While both the the-atre and the heterochronical spaces of a museumare both artificial and temporary projections, thestructure of the exhibition relies on 'real', culturallyrobust objects. Could this wonder resonate in thedigital image of the new media? Could there evenbe such a thing as a virtual aura?

While the exhibition is a discursive space of a me-diated message or sets of messages, visitors tradi-tionally expect to encounter the 'rear object. If visi-tors come to the museum and find, say, videos, whywould they even need to come into the museum inthe first place when they could enjoy the workequally well from the comfort of home on privatetelevision sets or video players? In a media satu-rated society we are bombarded with a surfeit ofimages that have been mechanically reproduced allaround us: advertising on the street, art posters inpublic institutions and art catalogues in school andhome. If the museum were to be relied on as alocation of culturally robust objects, the exhibitingof videos, electronic interactives, and virtual reality

204@Archives & Museum Informatics, 2001 2 I 3


installations would seem to defeat the purpose ofthe art museum experience.

The history of photography has long left behind thenotion of the photograph as historical document,and through aesthetic appreciation it has come tobe a theoretical object, no longer perceived merelyas a stand-alone simulacrum, eventually attaining astatus of its own. This ontological evolution tookalmost a century and we now recognize the capac-ity of the photographic image to stir emotions andevoke wonder. Roland Barthes (Barthes, 1981, 2000)affirmed that photographs do radiate a certain kindoraura:The aura of the lost in me and of lost memo-ries act in much the same way that Proust's textualreminiscences of the Madeleine pastry and the po-tency of its smell served to evoke buried memory.Barthes distinguishes the"punctum" as that accidentof photographic detail that pricked him, bruised himand was so poignant to him that it evoked an al-most transcendental experience, conjuring up poi-gnant lost memories of his mother. The historicalprocess of the photographic image emerged fromwindow to artefact,much as video has migrated fromdocumentation tool to art form. Perhaps we needto maintain aesthetic distance from theWorldWideWeb to distinguish the parameters of this still newmedium, in order to identify new spaces of enchant-ment amongst the cacophony of information.

Culturally robust objects serve as referents to cul-tural discourses and historical processes, and weare all aware of the many ways they evoke wonderin us. For the majority of society, without the capi-tal to surround themselves with the original, themuseum, the zoo, and the botanical garden offer apublic space to languish in the authentic. However,with leisure time a limited asset, we depend on themediated experiences and the surrogate to fill inthe gaps. For many years, our reliance on the signi-fier rather than the signified, the Baudrillardsimulacra, a duplicate without an original, has beenconstantly debated. It has even caused us to takeactions in surprising ways that are not based in ob-jective reality.The alarm over the Y2K bug that in-creased as the new millennium drew closer was anexample of the fear of the signifier, a mere numeri-cal representation, rather than of a concrete his-torical eventAccording to Kevin Robins,

The idea that we are already living in a simu-lation culture has now almost become a cli-che. We have actually come to feel rathercomfortable with our new condition ofderealisation (Robins, 1996).

In order to ascertain if the potency of Gell's en-chantment, the Barthes"punctum" or the resonanceof Benjamin's lost aura can be found online, I willturn to the online projects and investigate how theyhave made their screen debuts.

Ghosts, Magic and Enchantment

TheThinker ImageBase, from the collections of TheFine Arts Museums of San Francisco (the deYoung Museum and the Legion of Honor),USA,rep-resents the traditional museum metaphor, includ-ing activities, visitor information, membership, edu-cation activities and online shopping at the museumstore. At the same time, the site offers the visitoror surfer access to the digital holdings of the realmuseums.While the deYoung Museum is closed tothe public until Spring, 2005, the databases provideauthoritative background material both on the ex-hibitions on display as well as the collections be-hind the scenes.According to theWeb site, the col-lections belong to the public, and because the mu-seum is able to show less than 5% of the collec-tions in the galleries at any given time, they feel aspecial responsibility to make them accessible inother ways.The ImageBase is a fully keyword search-able database containing 110,000 images from thecollections and is promoted as an expression of themuseum's mission to provide meaningful public ac-cess to the collections.lt behaves more like a re-source and less like a repository.The database of-fers a compelling educational experience, and re-calls Andre Malraux's message of universality in his"Museum Without Walls." Just as Malraux predictedand applauded the globalization process that wasyet to evolve, both he and Walter Benjamin wouldundoubtedly have celebrated the unrestricted dis-tribution of art resources that are now freely be-stowed upon remote visitors by such museums overthe Internet.While the educational value of such asite is indisputable, in that it effectively replaces thetraditional learning tool of slides or exhibition cata-logue, I would question the notion of 'meaningful



access:WhileWeb authors invest considerable timeand energy in making images speedily accessiblethrough limiting to low resolution and cropping, itis precisely this immediacy of access that makes theprocess so alarmingly effortless.The speed factor,the 'click to go' phenomenon, may actually act as adisservice to the collections and act as the antith-esis of the enchantment of technology, that in factcauses the disembedding of cultural systems.

A museum that does not exist in objective realityand is exclusively constructed electronically on theWorld Wide Web is the MUVA, El Pais VirtualMuseum of Art.This museum is a virtual fabrica-tion, and maintains only a tenuous connection toreality. MUVA utilizes a 3D technique,Web2mil, toconjure up a magic environmentAlicia Haber, theDirector of the museum, welcomes visitors to themuseum, which specializes in contemporaryUraguayan and Latin American art, and hosts ex-tensive collections of paintings by leading Uraguayanartists. Four architects, Jaime Lores, Raul Nazur,Daniel Colominas and Marcelo Mezzottoni, werecommissioned to prepare the plans for the build-ing, on Avenida 18 de Julio, the main artery ofUruguay's capital, Montevideo.They created a finearts museum consisting of galleries for permanentand temporary exhibitions, as well as spaces for in-formal shows, sculpture garden, restoration work-shops, and administrative service areas.The build-ing has five main floors where galleries are open tothe public, twenty-four hours a day... virtually thatis! Some sixteen graphic and Web designers, pro-grammers, photographers and system managersmodeled textures of the walls, stairways, windows,sidewalks, roofs and elevator, pixel by pixel, to pro-vide a sense of 'reality' for the visitors. Intuitive navi-gation tools, allowing for fluid exploration aroundthe galleries and collections, were studiously hungand discretely lit.Through embedded 'hot-spots' inthe paintings, click-able links refer to in-depth stud-ies of the artists' work, biographies and further in-formation on the thematic presentation of the ex-hibition.

To construct the same museum in concrete, steeland glass would have cost over 100 million dollars,a prohibitive sum for the Uruguayan reality. Due tothe efforts of this highly motivated and imaginative

team, Uruguay's artists can now show their workscollectively, substituting their own virtual museumfor that impossible museum.This echoes Gell's com-ment that the essential alchemy of art is to makewhat is not out of what is, and to make what is outof what is not. In this case one is not describing anart object, but an entire museum. But we might alsobe reminded of Lash, Urry and Giddens' dubiousimplications for society and discern that the virtualmetaphor of a museum might be a reflection of theemptying out of subject and object. Even so, whilewe do recognize a substantial loss, we might alsoside with Benjamin that, in this loss, there is also awelcome gain.The liberation of the original objectand its distribution over the Internet opens up, forthe first time, the availability of Uruguayan art forremote visitors and the opportunity for these art-ists to reach a broader audience.

"Uncomfortable Proximity" is the project cre-ated by Harwood, a member of the Mongrel col-lective with critical texts by Mathew Fuller and com-missioned by the Tate.While the 'real'Tate providesfloor plans, the Mongrel version takes us to what isbeneath the floorboards of the Tate.The texts andimages introduce us to the precarious foundationsof the Tate galleries, the Millbank penitentiary, thefilth of the Thames and the hidden history of theslave trade.The accrual of wealth through the slavetrade had implications for generations of Britisharistocracy that inevitably translated into the cur-rency of art, some of which found its way into theTate collections.The Web site may be accessed viathe main Tate site, where it kind of sneaks up onyou with what appears to be a clone of the specificpage you intended to visit on theTate web-site.Theextra windows need no invitation.They unscrupu-lously appear on your browser in the backgroundas you click your way through the site, and take youinto the underbelly of this institution of Britain'snational heritage and the decaying matter of the20th Century.This challenge to the very institutionthat is partner to the project and acts as host tothe scathing message is remarkable in itself. How-ever, what this project does is to limelight the veryfoundations of the Tate cathedral, the circulatorysystem of art in society, and specifically the sacred-ness of the British art system. This recalls Gell'sobservation that art objects cannot be evaluated



when dislocated from their social systems but needto be seen in light of their circulation in society,where art, as art agency, the object, the artist, andthe recipient, invokes a series of responses or in-terpretations.The use of this media is resourceful,perniciously using the electronic stage to challengeall that is embedded in the mythological nature ofthe circulatory art system, and symbiotically locatedinside of the very fabric of its embodiment, the of-ficial Web site.

The Geist Project, by Colin Andrews, of the NewMedia Group, Scotland, is an eerie Web site involv-ing ghosts, and was exhibited at The Pier Art Cen-tre, Stromness, in November, 2000. Four remotesites across Scottish, trad itionally 'haunted locations'are networked and act as nodes,gathering data suchas changes in temperature and fluctuations in elec-tromagnetic radiation.The information is then re-layed via electronic networks to the library at thePier Arts Centre, Stromness, Orkney where it isused to 'feed' an audio installation.The audio is de-rived from traces of 'voices' recorded at each loca-tion.The work is experienced as a four channel audioinstallation, with each of the four channels repre-senting one of the four remote 'haunted' locations.The audio is derived from traces of`voices' extractedfrom recordings made at each location at an earliertime.According to the New Media Scotland Com-mission, Geist is not about the existence or other-wise of ghosts, but rather about ghost or spectralityas metaphor. It attempts to explore our contempo-rary condition of omnipresent absence - presencethrough the use of haunted locations, recordedsound, and network technology.This spine-chillingproject reminds us that contemporary communi-cations technologies belong to unseen places - theyconnect us instantaneously across vast distances yetmake our words, impulses and feelings pass throughan uninhabited and invisible domain.

Geist is about being and not being here andthere simultaneously. It is about communi-cation through the exchange of electricalenergy, about recording and playback, aboutreturns and repetition. It is the domain ofspecters and spirits, of slippages in time andspace and communications across bound-aries.

(From the New Media Scotland Commissionelectronic promotion).

The World Wide Web offers many kinds of spaces,and while the traditional museum flows through andinto many kinds of cultural discourses, the art andthe artistic processes provide the vehicles of en-chantment for society.Artists are beginning to real-ize the potential of the interconnectivity of themedium and respond to its resonance. Geist is aproject that clearly articulates these phenomena.The enchanting technological processes of net-worked art, firmly embedded in circulatory, socialsystems, propel sacred spaces from remote loca-tions across vast distances into our foreground atthe blink of an eye.We recognize the inextricablelink between art and culture, yet we are awed bywhat we do not fully understand.When we are ableto step back and maintain critical distance from theseexperiences and appreciate the craftsmanship of thenew medium, we might even be able to discern thepatina of human endeavour and regain in some waythe lost aura; the virtual aura.

The next time you sit in front of your computerand crank up the pixels, don't be surprised if youfind a poltergeist hanging out there inside your harddisk, dancing across the fiber optic connections, orlurking somewhere out there online, waiting to findyou!

References

Auge, M. (1995), non-places,Verso, London, New York

Barthes, R. (2000), Camera Lucinda: Reflections onPhotography,Vintage, UK

Benjamin,W. (1992), Illuminations, The Work ofArt in theAge of Mechanical Reproduction, Fontana Press, UK

Different Spaces, from a lecture presented to theArchitectural Studies Circle, March 14, 1967, firstpublished in 1984.

Foucault, M. Aesthetics, (1994) Essential Works ofFoucault 1994-1984,Vol. 2

Gell,A. ( I 992),Anthropology,Art and Aesthetics, ed.Coote, J. and Shelton, The technology of enchant-ment and the enchantment of technology, OxfordUniversity Press

e, 0



Gell,A. ( I 996),Vogers Net, Traps as Artworks andArtworks as Traps,

Gell,A. (1998), Art and Agency, Oxford University Press,UK

Giddens,A. (1990) Consequences of Modernity, PolityPress, UK

Jameson, E ( I 992), Postmodernism, or Cultural Logic ofLate Capitalism,Verso, 1992

Llash, S. and Urry. J. ( I 994). Economies of Signs and Space,London: Sage

Malraux,A., ( I 967), Museum Without Wa lls,The Voices ofSilence, Doubleday & Company, Inc, Garden City,New York

Morse, M. (1998), Virtualities,Television, Media Art, andCyberculture, Indiana University Press

Robins, (1996) K.S, Into the Image, Routledge, UK


About theAuthors

209

About the Authors

Carrie Adams is currently working as the WebSite Designer for the Seattle Art Museum. Threeyears prior to working for the Museum, Carrieworked at Hansen Design Company,an award win-ing graphic and web Design Company located inSeattle,Washington. Her work has appeared in suchperiodicals as Graphics, Communication Arts, andDavid Carter Publications.Carrie received her BFAin Graphic Design from Cornish College of theArts.She continues to bring web design, illustration andorganizational skills to the various web projects atthe Seattle Art Museum.

Adam Alonzo has been a World Wide Web de-veloper and trainer within the academic environ-ment since 1995. His areas of concentration includedistance learning and accessible Web design. Cur-rently employed in the Office of Disability Servicesat Wright State University in Dayton, Ohio, he is aCoordinator of the university's Accessible Arts pro-gram, which seeks to provide arts opportunitiesfor people with disabilities.Adam holds a BachelorofArts degree in Communications and a Master ofScience in Education.

Susan Anable has worked in the field of educa-tion for 15 years and is currently employed as theAssociate Director of Education atThe Dayton ArtInstitute, where her primary responsibilities are do-cent education and school and teacher services.TheEducation Department of the museum is commit-ted to making the visual arts accessible to peoplewith disabilities and strives to provide programmingwhich is inclusive. Susan holds a Bachelor of FineArts degree and a Master of Education.

Renee Arora was a Summer Intern at IBM T.J.Watson Research Center. Her research interestsinclude human-computer interaction and wirelesscomputing. She received an MS in Management In-formation Systems at The University of Maryland,Baltimore County in 2000.

Thimoty Barbieri has a M.Sc. in Computer Engi-neering, and is pursuing his PhD at the Politecnicodi Milano, Italy. He is active in various fields of Com-puter Science, mainly 3D Graphics collaborationover the Internet, and applications for Mobile De-vices.

I. 0

Ana Carolina Bertoletti received a B.Sc. Com-puter Science in 1996 from II/PUCRS. She receiveda M.Sc. Computer Science in 1998 from CPGCC/UFRGS. She is currently the Coordinator forMultimidia and Internet Division at MCT/PUCRS.

Henri Betaille is Associate Professor at InstitutUniveritaire de Technology Montpellier-Beziers,Computer Science department.Works on HumanComputer Interaction and Hypermedia at LIRMM.Proeceedings Chair at IHM99.Works on the OpalesProject.

Christian Breiteneder is a full professor for In-teractive Systems with the Institute for SoftwareTechnology at the Vienna University of Technology.Before joining the institute he was associate profes-sor at the University ofVienna. Breiteneder earneda diploma degree (Dipl.-)ng.) in computer sciencefrom the Johannes Kepler University in Linz in 1978and a PhD in computer science from theVienna Uni-versity of Technology in 1991. Research interestsinclude media processing systems, content-based vi-sual information retrieval, teleconferencing systemsand museum information systems.

Thomas A. Cofino is a Senior Manager, ResearchStaff Member, e-Business Applications andTechnolo-gies in the Applications and Solutions Technologiesdepartment, IBM T. J. Watson Research Center, Mr.Cofino is responsible for leading Research's workin technology and application development in sev-eral areas: e-commerce intelligence;.human-centricweb application design; and customer relationshipmanagement (CRM).

Traci Cole joined the Seattle Art Museum in 1998as a software engineer and web programmer. Shehas been working in the computer industry sincethe early 1980s, and with the Internet andWeb since1995.

Sarah Cook is a Sunderland University doctoralresearcher in New Media Curation (a collaborationwith the Baltic Centre for Contemporary Art). Sheshe worked at the Walker Art Center in Minneapo-lis, and organised the EAT on-line forum there. Shehas been a curator-in-residence at the Banff Centrefor theArts in Banff,Alberta, Canada and has workedat the National Gallery of Canada in Ottawa.

Museums and the Web 200I:ProceedingsArchives & Museum Informatics, 2001, p. 221

About the Authors . . .

Antonio Carlos da Rocha Costa is AssociateProfessor, Escola de Informática, UCPel. He holdsdegrees in Electronic Engineering, 1977, EE/UFRGS,M.Sc. Computer Science, 1980, CPGCC/UFRGS,anda Ph.D. Computer Science, 1993, CPGCC/UFRGS.

Steven De Bryne is a student in computer sci-ence at heVrije Univeriteit Brussel.

Christina De Paolo is Promotions and Web SiteManager at the Seattle Art Museum. She began herarts administration career working for a small non-profit organization, coordinating grant programs thatsupport the work of Washington State artists. Atthe Seattle Art Museum she works with a cross-departmental team to expand the Museums Website and initiate new media projects.

Susan Edwards is currently continuing work withthe curators at SAM as a Research Associate and isalso a PhD candidate in the History of Art at theUniversity of Michigan. For two years Susan workedas the Curatorial Associate involved with the Grow-ing UpWith Art project at the Seattle Art Museum.She worked with curators and museum educators,and Seattle Public School teachers during this grantperiod to bring 6th graders from two Seattle PublicMiddle Schools into the museum to curate theirown exhibitions. Susan helped students select ob-jects from SAM's permanent collection, developedlesson plans with museum educators, and traveledwith SAM educators to the schools to teach theselessons on how to curate an exhibition.The lessonson the web site are the same lessons that theseSeattle students learned in the process of creatingtheir own exhibition at the Seattle Art Museum. It isthis experience that the "My Art Gallery" web siteattempts to keep alive by bringing it to the web.She continues to bring web design, illustration andorganizational skills to the various web projects atthe Seattle Art Museum.

Gilles Falquet received the PhD degree in com-puter science from the University of Geneva, Swit-zerland, in 1989. He is now an assistant professorat the Department of Information Systems of theUniversity of Geneva. His research interests includeinformation system interfaces, hypertexts, hypertextviews on databases, and knowledge base interfaces.

C. Olivia Frost is Associate Dean for ProfessionalPrograms and Professor at the University of Michi-gan School of Information. She has been a facultymember at the University of Michigan since 1977.She has a B.A. in German Literature from HowardUniversity, an M.LS.from the University of Oregon,an M.A. in German Languages and Literatures, and aPh.D. in Library Science from the University of Chi-cago.

Beryl Graham is a writer, curator and artist witha special interest in interactive art both high-techand low-tech. She has curated the major exhibitionof international interactive artwork,Serious Games,for the Laing and Barbican art galleries, and has alsoworked with the Exploratorium, San Francisco. HerPh.D. concerned audience relationships with inter-active art in gallery settings, and she has writtenwidely on the subject for books and periodicals in-cluding Leonardo, Convergence, and PublicArt Jour-nal.

Jacques Guyot received the PhD degree in com-puter science from the University of Geneva, Swit-zerland, in 1987. He is now an information systemdesigner and manager at Rolex watches S.A. andlecturer at the University of Geneva. His researchinterests include information system interfaces,hypertext views on databases, electronic documentmanagement, and indexing techniques.

Ilse Harms is assistant professor at the Depart-ment of Information Science, University of Saarland.She holds the degree of PD Dr. habil. She teachescourses in e-learning, media research and usabilityengineering. Her research interests are media ef-fects and mass communication research and distanceeducation online.

Susan Hazan is the Curator of New Media Edu-cation Unit, Head of the Internet Office at The Is-rael Museum, Jerusalem and Doctoral Student, Me-dia and Communications, Goldsmiths, UniversityLondon currently researching the virtual in themuseum. She has created museum applications andteaches hands-on multimedia workshops for art-ists, teachers, children and museum staff. Her Mas-ters is in Media and Communications, Goldsmiths,University London, UK.

4

i 1



Martin Hitz is a full professor for Interactive Sys-tems with the Institute for Computer Science -Sys-tems at the University of Klagenfurt Before joiningthe institute he was associate professor at the Uni-versity of Vienna. Hitz earned a Dipl. Eng. in com-puter science in 1982 and a PhD in computer sci-ence in 1989 from the Vienna University of Tech-nology, respectively. Research interests include mul-timedia systems, usability patterns and web-basedinformation systems.

Anthony Huberman is the Education and PublicPrograms Coordinator and current Acting Educa-tion Director at RS.1 Contemporary Art Center, inNew York. He received a B.A. in sociology fromGeorgetown University. He has worked at the TheWashington Project for the Arts/Corcoran, theMuseum of Modern Art, NY, Exit Art, NY, and DiaCenter for the Arts, NY. In February 1999, he be-gan his position at RS.1 Contemporary Art Center(a MoMA Affiliate).

Clare-Marie Karat received her Ph.D in SocialPsychology from the University of Colorado andhas worked for the last fifteen years in private in-dustry and government as a user interface designerin the development of a variety of customer solu-tions. Dr. Karat has published papers and contrib-uted to several books on the topic of human fac-tors cost-benefit analysis, usability engineering meth-ods, universal access, personalization, and speechrecognition.

John Karat is a Research Staff Member at the IBMT.J.Watson Research Center. His research work inhuman-computer interaction (HCI) has coveredtopics from input techniques, information visualiza-tion and the design process.John is theACM SIGCHIAdjunct Chair for International Information and amember of the SIGCHI Extended Executive Com-mittee. He is also the ACM representative to and aVice-Chair of the IFIP Technical Committee onHuman-Computer Interaction (IFIPTC 13).John wasco-chair of the recent ACM Conference on De-signing Interactive Systems (DIS 2000).

Sarah Kenderdine is traditionally a maritime ar-chaeologist and museum curator.She has researchedand excavated shipwrecks throughoutAustralia andthe Indian Ocean region and has written a numberof books on shipwrecks. She has also completed aMaster of Arts (design research thesis) on virtualmuseums.

Robert Meersman is full professor atVUB Com-puter Science Department since end of 1994, Di-rector of its new Laboratory for System Technol-ogy and Applications Research (STARLab). He wasa past Chairman of IFIP WG 2.6 (Database), cur-rent Chairman of IFIP Technical Committee TCI2(Artificial Intelligence). He was responsible atTilburgUniv. for the ESPRIT projects SPRITE on documentmanagement, ISCORE 1 and II (BRA-s), involved inNoE-s on Machine Learning and Logic Programming,and for the RIDL* project that led to the industrialspin-off of a new generation of innovative CASEtools.

Mircia Cristina Moraes was born in Porto Alegre,RS, Brazil and received a BS and MS in ComputerScience from Pontificia Universidade Católica do RioGrande do Sul (PUCRS), respectively in 1997 and1999. Currently she is PhD student in ComputerScience at Universidade Federal do Rio Grande doSul. She is also professor in the Informatic Facultyat PUCRS. Her current research focuses on Artifi-cial Intelligence, Software Agents, Human-ComputerInteraction and Distance Learning.

Slavko Milekic holds a medical degree (BelgradeSchool of Medicine) as well as Ph.D. in Experimen-tal Psychology (University of Connecticut). His re-search interests include digital design, the psychol-ogy of human/computer interaction and building ofWWW-based tools for knowledge transfer. Dr.Milekic is frequently leading workshops and con-sulting on the educational uses of technology andhuman/computer interface design. He currentlyholds the position of an associate professor of Cog-nitive Science & Digital Design at the University ofthe Arts.

Jocelyne Nanard is a Professeur University ofMontpellier 11 & LIRMM Montpellier. Since 1978, herresearch activity mainly concerns document engi-neering, Human Computer Interaction, and moreprecisely hypertext and hypermedia since 1988. Hercurrent positions include Vice head of KnowledgeRepresentation Department LIRMM 92-95. Headof networking dept at Institut Universitaire de Tech-nology Beziers 94-96, co-Chair Human ComputerInteraction research group PRC CHM.

Marc Nanard is Professeur des Universités, CNAMParis & LIRMM Montpellier. Since I 978, his researchactivity mainly concerns document engineering,


About the Authors . . .

Human Computer Interaction, and more preciselyhypertext and hypermedia since 1988. His positionsinclude Head of Human Computer Interactiongroup at LIRMM-1990-now, member of EEAI dept.Council 92-94; member of LIRMM Council 95- 97,Vice head of Knowledge Representation Depart-ment 95-98, Head of CS Department at CRACNAM Montpellier, 90- .

Luka Nerima teaches computer sciences at theFaculty of Arts, University of Geneva, Switzerland.He received his B.A. degree from the ComputerScience Department at the University of Geneva in1981. Since then, he has been active in softwareengineering and database research domains. He isresponsible for the design and the maintenance ofa lexical database in an ongoing research project onNatural Language Precessing. He is also involved ina research project targeting the publishing of data-bases on the Web.

Paolo Paolini is the manager of HOC laboratory.He is full professor at Politecnico di Milano (Com-puter Graphic), Lecturer at the Faculty of Engineer-ing and Industrial Design at Politecnico di Milanoand Lecturer at the School of Communication Sci-ences at the University of Italian Switzerland (USI)-Lugano. He has more than 70 published papers onthe following subjects: relational Data Bases, DataModeling,Abstract DataTypes and Data Bases,Viewsfor Data Bases,Automatic generation of documents,Hypermedia and WWN/V modeling and design, Multi-media implementation, Hypermedia evaluation,Au-tomatic generation of Hypermedia (WWW), col-laborative access toWWW, cultural applications ofnew technologies, use of advanced technologies foreducation and training.

Seongbin Park is a postdoctoral researcher atthe Information System Interfaces group of the Uni-versity of Geneva in Switzerland. He graduatedfrom Korea University in Seoul, Korea with BS de-gree in computer science and received both MSand PhD degrees in computer science from theUniversity of Southern California. His research in-terests include adaptive hypermedia, informationsystem interfaces,Web-based system design, inter-active museums, markup languages, and formal lan-guage theory.

John Perkins operates Mus*Info, an international

consulting company specializing in information andtechnology management for museums, archives, andcultural heritage organizations. He is Executive Di-rector of the Consortium for the Computer Inter-change of Museum Information (CIMI). Previouslyhe was Director, of Client Operations for a soft-ware firm dedicated to museum collections infor-mation management systems, and prior to that di-rected the international, Conservation InformationNetwork, with 450 subscribers in 23 countries forthe Getty Conservation Institute (GC1) in Los An-geles, California.

Claudio Pinhanez recently got his Ph.D.from theMIT Media Laboratory, working on the design andconstruction of physically interactive environmentsand in computer theater. He also researched newparadigms for computational representation of hu-man action and scripting of interactive stories, us-ing Allen's interval algebra as the underlying repre-sentational structure. He is currently working at IBMTJ Watson Research Center on web-based culturaland entertainment experiences and on physical in-terfaces to information.

Hubert Platzer is a researcher with the Institutefor SoftwareTechnology at theVienna University ofTechnology.Hubert Platzer has a more than 10 yearsexperience as consultant, software developer andproject leader in various industrial projects. Heearned a Dipl. Eng. in computer science from theVienna University of Technology in 1983. His inter-ests include web programming and web informa-tion systems.

Mark Podlaseck is interested in the aggresive useof pixels. At IBM's Watson Research Center, he isworking with Philip Glass on a one page catalog ofthe composer's work, soon to be online.

Doug Riecken is a Research Staff Member at theIBM T.J.Watson Research Center.

Werner Schweibenz is a Ph.D. student with theDepartment of Information Science, University ofSaarland, where he teaches usability engineering. InJuly 1999 he participated in the International Sum-mer Workshop Exploring a Communication Modelfor Web Design at the University of Washington,Seattle. He spent the academic year 1996/97 at theSchool of Library and Informational Science at the

ci j224 © Archives & Museum Informatics, 2001


University of Missouri-Columbia on a Fulbrightgrant In the Summer of 1997 he interned with theSmithsonian Institution.

Jim Spadaccini is the founder of Ideum(www.ideum.com),a multimedia design and produc-tion company which specializes in educational andmuseum Web sites. Before that he was Director ofInteractive Media at the Exploratorium in San Fran-cisco. He was the major contributor to, lead de-signer, and manager of the three-timeWebbyAwardwinning Exploratorium Web site (Best Science Site1997-99). In 1995, Jim worked with Apple Com-puters, where he authored a book, two CD-ROMS,and an educational Web site. Along with develop-ing educational Web resources, Jim has been teach-ing Internet and Multimedia classes at San Fran-cisco State since 1994.

Leonard Steinbach joinedThe Cleveland Museumof Art as its first Chief Technology Officer in fall,1999. Prior to this position he was Chief Technol-ogy Officer for the Solomon R. Guggenheim Mu-seum (1996-99) where he was responsible for ex-hibition as well as general business technology. Forseven prior to that he was Associate Director ofComputing at a major health sciences center inNewYork

Peter Stuer has been working on diverse projectsin the fields of Artificial Intelligence and InternetInformation Systems.

Kevin Sumption has worked for nearly ten yearsas both a science and social history curator. He iscurrently the National Project Manager of Austra-lian Museums On-Line (AMOL) and at the sametime Curator of InformationTechnology at the Pow-erhouse Museum. Kevin's original passion and in-deed qualifications, are in the areas of design his-tory and theory.As such he also lectures in DesignHistory, Design Theory and Research Methodologyat the University ofTechnology, Sydney.

johnVergo is the manager of the User ExperienceResearch Group at the IBM T.J.Watson ResearchCenter in Hawthorne New York. His research in-terests include human-computer interaction, user-centered design methods, multimodal user inter-faces, e-commerce user experiences, speech rec-ognition, natural language understanding. scientificvisualization, 3D graphics and software developmentmethods. He has a BS in Mathematics and Psychol-ogy for the University atAlbany, and an MS in Com-puter Science form Polytechnic University.

214


Archives Ce MuseumInformatics

publishes current reports from experts wordwide onissues critical to cultural heritage in the information age

Archives & Museum InformaticsReaderEdited by David Bearman, Jennifer Trant, andCostis DallasISBN: 1-885626-21-5

Approx. 250 pages$40.00

A collection of outstanding papers from both theICHIM and Museums and the Web conferencesprovides an excellent background into museumweb development and hypermedia. The Readerprovides a historical perspective of the develop-ment of hypermedia in museums, surveying thepast 10 years, and exploring subjects ranging fromweb site design, to virtual museum tours, andintellectual property rights issues.

IrnH$Ii"

Museums and the Web2000:Selected papers from aninternational conferenceEdited by David Bearman andJennifer TrantISBN: 1-885626-20-7208 pages and CD-ROM(requires Web browser)

$50.00

This print volume includes the best of the paperspresented at Museums and theWeb 2000 onApril 16-19th, 2000. The CD-ROM includes allpapers submitted, abstracts of all presentations andbiographical information for all presenters.Topicscovered include Virtual Museums, Evaluation,Design Issues, and much more.

215

Managing Web SitesBy Stephen SmithISBN: 1-885626-22-3Approx. 120 pages$35.00

Is your website taking over? Get practical adviceon how to regain control and plan futuredevelopments in this workbook prepared by anexperienced web site administrator.

Museums and the Web 99:Selected papers from aninternational conference

Edited by David Bearman and Jennifer Trant

ISBN 1-885626-17-7 (available March 1999)245 pages, and CD-ROM (requires web browser)$50.00

Museums and the Web 1998:Proceedings (CD-ROM)Edited by David Bearman and Jennifer TrantCD-ROM (requires web browser. Netscape 3.0 or equiva-lent) 1997. $25.00

Includes beyond interface:netartand art on the net" curated by Steve Dietz.

Museums and the Web, 1997:Selected PapersEdited by David Bearman and Jennifer TrantISBN 1-885626-13-4 (1997), 373 pp., $30.00

International Cultural Heritage Informatics Meeting(ICHIM)

International Cultural HeritageInformatics Meeting, Milan, Italy,September 3-7, 2001: Selected Papersedited by David Bearman, Franca Garzotto,Paolo Paolini and Jennifer Trant.

Watch www.archimuse.com for details!

Cultural Heritage Informatics 1999:Selected papers from ichim99, theInternational Cultural Heritage InformaticsMeetingEdited by David Bearman and Jennifer TrantISBN 1-885626-18-5 (1999) 255 pp., $50.00

Papers from the Fifth International Cultural Her-itage Informatics Meeting, ichim99 in WashingtonDC, have been edited for this print publication.Sections reflect the major themes of the conference:interactivity, converging technologies, userinvolvement, and new models for museummultimedia.

Museum Interactive Multimedia:Cultural Heritage Systems Designand InterfacesEdited by David Bearman and Jennifer TrantISBN 1-885626-14-2 (1997), 233 pp., $30.00

A decade of progress in interactive multimedia inmuseums forms the basis for papers on systemsdesign and user interface from the Fourth Interna-tional Conference on Hyper-media andInteractivity in Museums (le Louvre, Paris 1997).These papers focus on design systems developmentand evaluation methodologies), interfaces (visitoraware systems and inter-actives providing geo-graphical and chronological views of data), and casestudies of museum ew multimedia ranging fromcollection catalogs to 3D environments.

216

Multimedia Computing andMuseumsEdited by David BearmanISBN 1-885626-11-8 (1995) 388 pp., $20.00

Volume 1 of selected essays from the ThirdInternational Conference on Hypermedia &Interactivity in Museums (ICHIM 95 / MCN 95)on the technological, cultural and intellectualissues raised by the use of multimedia technolo-gies to represent cultural heritage. Papers profilethe iimpact of technologies on museum applica-tions and audiences, and on the relationship ofmuseums to society.

Hands on: Hypermedia andInteractivity in MuseumsEdited by David BearmanISBN 1-885626-12-6 (1995) 293 pp., $20.00

Volume 2 of selected papers from the Third In-ternational Conference on Hypermedia &Interactivity in Museums (ICHIM 95 / MCN 95)reflecting the evolution of delivery mechanismsfor interactive multimedia, the new social andinstitutional arrangements they engender, and thecontinuing importance of intellectual propertyissues. Groups of essays address fixed-format pub-lishing, in-house interactives, networked access,museum consortia, museum teamwork, commer-cial partnerships and intellectual property.

Museums and InteractiveMultimediaEdited by Diane LeesISBN 1-8856263-89-X (1993) 436 pp., $20.00

The proceedings of the Second InternationalConference on Hypermedia & Interactivity inMuseums (ICHIM 93) include over sixtypresentations by authors from over twenty nationson the issues of design and implementation ofmuseum interactives.

Hypermedia and Interactivity inMuseums -Out of Print-Edited by David BearmanISBN 1-885626-03-7 (1991) 340 pp.,

Titles on Archives andElectronic Records

Electronic Records Research 1997:Resource Materials (CD-ROM)Edited by David Bearman, Kimberly Barata, andJennifer Trant

CD-ROM (requires web browser, Netscape 3.0 or equiva-lent) 1997. $25.00

This CD contains copies of over 100 papers andexecutable links to hundreds of sources on thedefinition, policy, capture, storage, and migra-tion of electronic records together with anannotated bibliography and the archival Websites of the major research projects at theUniversity of British Columbia and University ofPittsburgh. In addition this essential referencecontains biographical and contact informationfor researchers in these fields worldwide.

Electronic Evidence: Strategies forManaging Records in ContemporaryOrganizationsBy David BearmanISBN 1-885626-08-8 (1994) 314 pp., $30.00

A collection of previously published papers,accompanied by a new essay exploring theevolution of concepts of electronic recordsmanagement. The papers reprinted here wereoriginally published between 1989 and 1993, injournals in the United States, Canada, Portugaland Australia, as well as in a United NationsReport. Includes a detailed index by VictoriaIrons Walch.

Electronic Records ManagementProgram StrategiesEdited by Margaret HedstromISBN 1-885626-07-X (1993) 156 pp., $20.00

Papers prepared for a joint meeting of SAA-CART and NAGARA-CIT in the spring of1993. Includes the results of brainstormingsessions, an essay on program structure optionsby David Bearman and Margaret Hedstrom, andan annotated bibliography by Richard J. Cox.

21-7

Order DetailsEmail [email protected] for details orcontact us at:

Archives & Museum Informatics2008 Murray Ave, Suite DPittsburgh, PA 15217 USAPhone: +1 412 422 8530Fax: +1 412 422 8594Email: [email protected]

Online Order Form athttp://www.archimuse.com/pub.order.html

Bulk Orders, Standing Orders, andClassroom Discounts AvailableShipping and HandlingPrepaid orders from the United States areshipped free of charge. A $5 US shipping andhandling fee applies to invoiced US orders.International orders please add $10 US peritem for shipping and handling.

PaymentAll prices are in US Dollars. Payment can bemade by cheque, money order, bank draft,bank transfer, or credit card (VISA,Mastercard, or American Express). Chequesmust be payable to Archives & MuseumInformatics, in US funds, drawn on a USBank. Money Order or Bank Draft should bepayable to Archives & Museum Informatics.Call or email for Bank Transfer details.

About the CD-ROM

This book is accompanied by a CD-ROM containing versions of all these papersand many others, presented in Hyper Text Mark-Up Language (HTML), the techni-cal language of theWeb. It also includes abstracts of all the papers, demonstrationsand workshops presented at the conference, and the biographies of all speakersand presenters.These electronic versions include color illustrations and links tothe sites discussed and referenced.

You don't have to be connected to the Internet to read the papers on the CD-ROM or to navigate the full background information about the conference.Youwill need your own connection to the Internet to go to the linked museum sitesand to follow the external links in the papers.

To use the CD-ROM, you will need a Web browser (Netscape 4.0 or InternetExplorer 4.0 or higher are recommended). Put the CD in your computer, launchyour browser, and, using the File / Open menu choices, navigate to the index.htmlfile in the main directory of the CD-ROM. Open this file in your browser allother files are linked from there.

Speakers provides a list of all the speakers at the conference and links to theirabstracts, biographies and papers (where available).

Sessions provides an overview of the Museums and theWeb 2000 conferenceprogram and links to abstracts and paper biographies

Best of the Web will take you to the results of the Best of the Web 2000conference, but you have to be connected to the Internet to do this.

If you have any questions or problems using the CD-ROM please [email protected], and we'll do our best to help you.

218

U.S. Department of EducationOffice of Educational Research and Improvement (OERI)

National Library of Education (NLE)

Educational Resources Information Center (ERIC)

NOTICE

Reproduction Basis

410

EElocalloail hafts Magee Center

This document is covered by a signed "Reproduction Release (Blanket)"form (on file within the ERIC system), encompassing all or classes ofdocuments from its source organization and, therefore, does not require a"Specific Document" Release form.

This document is Federally-funded, or carries its own permission toreproduce, or is otherwise in the public domain and, therefore, may bereproduced by ERIC without a signed Reproduction Release form (either"Specific Document" or "Blanket").

EFF-089 (1/2003)