dod software tech news vol. 3- no. 1 - csiac...and versioning (webdav) working group of the internet...

21
Techniques The DoD Source for Software Technology Information. Vol. 3- No. 1 Special STC ‘99 issue! Collaborative Win Win: A System for Negotiating Requirements by Ellis Horowitz, Joo H. Lee, and June Sup Lee Read additional Collaborative Technique articles at: www.dacs.dtic.mil/ awareness/newsletters/ listing.shtml In This Issue: Win Win: A System for Negotiating Requirements ... 1 WebMe Data Visualization Tool ...................................... 2 Web DAV: Collaborative Software Engineering on the Web ........................... 5 Orbit: Collaboration Environment ...................... 10 Collaborative Techniques Resources on the WWW .... 20 DACS Products Orderform ..................... Insert Continued on page 14 Introduction Win Win is a computer program that aids in the capture, negotiation, and coordination of requirements for a large system. It assumes that a group of people, called stakeholders, have signed on with the express purpose of discussing and refining the requirements of their proposed system. The system can be of any type. Win Win contains facilities for: l capturing the desires (win conditions) of the stakeholders l organizing the terminology so that stakeholders are using the same terms in the same way l expressing disagreements or issues needing resolution l offering options as potential solutions l negotiating agreements which resolve the issues l using third party tools to enlighten or resolve issues l producing a requirements document that summarizes the current state of the proposed system l creating documents that support multimedia and hyperlinks l tracing the ways by which requirements decisions were reached l checking the completeness and consistency of requirements Win Win Functionality A Simple Scenario WinWin offers a group of users a great deal of functionality, but as a result some planning is useful before getting started. In this section we offer a simple scenario for how users of WinWin might begin their work.

Upload: others

Post on 27-Jul-2020

11 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

Techniques

The DoD Source for Software Technology Information

Vol 3- No 1Special

STC lsquo99 issue

Collaborative

Win Win A System for Negotiating Requirementsby Ellis Horowitz Joo H Lee and June Sup Lee

Read additionalCollaborative Technique

articles at

wwwdacsdticmilawarenessnewsletters

listingshtml

In This Issue

Win Win A System forNegotiating Requirements 1

WebMe Data VisualizationTool 2

Web DAV CollaborativeSoftware Engineeringon the Web 5

Orbit CollaborationEnvironment 10

Collaborative TechniquesResources on the WWW 20

DACS ProductsOrderform Insert

Continued on page 14

Introduction

Win Win is a computer programthat aids in the capturenegotiation and coordination ofrequirements for a large systemIt assumes that a group ofpeople called stakeholders havesigned on with the expresspurpose of discussing andrefining the requirements of theirproposed system The system canbe of any type Win Win containsfacilities for

l capturing the desires (winconditions) of thestakeholders

l organizing the terminologyso that stakeholders areusing the same terms in thesame way

l expressing disagreements orissues needing resolution

l offering options as potentialsolutions

l negotiating agreementswhich resolve the issues

l using third party tools toenlighten or resolve issues

l producing a requirementsdocument that summarizesthe current state of theproposed system

l creating documents thatsupport multimedia andhyperlinks

l tracing the ways by whichrequirements decisions werereached

l checking the completenessand consistency ofrequirements

Win Win FunctionalityA Simple Scenario

WinWin offers a group of users agreat deal of functionality but asa result some planning is usefulbefore getting started In thissection we offer a simplescenario for how users ofWinWin might begin their work

Lon R Dean
Unclassified and Unlimited Distribution

STN2

The WebMe Data Visualization ToolRoseanne Tesoriero and Marvin V ZelkowitzUniversity of Maryland Fraunhofer Center for Experimental Software Engineering

Introduction

The ubiquity of the World WideWeb and the increased presenceof the Internet in todayrsquosmarketplace have led to anenvironment where software isbeing developed ingeographically distributedlocations This type ofdistributed developmentenvironment raises several newchallenges in the area ofsoftware project management

Measurement data has been usedto control and improve thesoftware process and itsproducts For example theSoftware EngineeringLaboratory of NASArsquos GoddardSpace Flight Center (NASASEL) has been using the QualityImprovement Paradigm alongwith the GoalQuestionMetricparadigm to help evaluatecontrol and improve softwareprocesses and products for overtwenty years [1] The use ofmeasurement data to build andpackage experience is a keyaspect in this approach Oneway in which experience ispackaged is in a baseline modelA baseline model is created byclustering project data fromsimilar projects together todescribe the expected behaviorfor this class of projectsManagement can then compare anew project with this baselinemodel in order to better

understand deviations from thebaseline For example a projectwith too few errors per line ofcode may either represent asuperior development process orinsufficient testing

When dealing with a distributeddevelopment environment theuse of measurement data to buildknowledge and experience forsoftware project managementpresents several new challengesWhen development occurs atphysically separated locations

the data may be geographicallyseparate as well If severalorganizations are involved in thedevelopment (such as withsubcontract management) theorganizational cultures may bedifferent and the data may becollected using differentcollection mechanisms leadingto variations in the formatsCreating baselines and

comparing different projectsbecomes more difficult due to thediffering interpretations to thecollected data When severalorganizations are involved in thedevelopment privacy becomesan issue Each organization doesnot want to give the othersunlimited access to itsproprietary data

The Web MeasurementEnvironment (WebME) has beendeveloped to address thechallenges of using measurementdata for experience-based projectmanagement in a distributeddevelopment environment Thisshort discussion explains how thetechniques incorporated into theWebME system allow projectmanagers to use data from adistributed environment A briefoverview of the WebME systemand a description of the analysistechnique used to build a baselinefrom similar projects are given

WebME solution

The WebME system wasdesigned to build baselinemodels from data collected ingeographically distributedenvironments In particular itcan be used to combine andanalyze time-series data collectedfrom software projects

The architecture of the WebMEsystem is shown in Figure 1

Continued on page 3

Figure 1 WebMe Architecture

STN3

Continued from page 2

The WebME system uses amediated architecture whichhorizontally partitions thestructure of the architecture intothree layers the informationrepositories mediatinginformation servers and the end-user application layer In theWebME context the end-usersubmits a request through a Web

Continued on page 4

Figure 2 Characteristic Curve of Project under Study

Metadata is created by using theWebME scripting language TheWebME scripting languageallows the system configurer tospecify the structure of thearchitecture the data that will beviewable with the system andthe access methods for the dataThe configurer of the systemcreates a script containing the

repositories data collected atgeographically separatedlocations and stored in differentformats can be combinedconsistently

It is important not only tocombine the data in a consistentmanner but to use the data tosupport project managementThe analysis technique supportedin WebME [2] uses data frompast similar projects to build abaseline of expected behavior foran attribute over time First acharacteristic curve is generatedfor the project of interest(Figure 2) The characteristiccurve describes the behavior ofan attribute over the projectlifecycle The baseline model forthe attribute is created using thecharacteristic curves of projectssimilar to the project of interest(Figure 3)

The baseline model can be usedto support project management inseveral ways For example if aproject is performing differentlyfrom the baseline it serves asignal for the project manager InFigure 4 the baseline model ofthe number of errors per weekand the current projects errorsper week are plotted The currentproject has encountered moreerrors than expected The projectmanager should investigate todetermine the cause of thedeviation If the deviation isundesirable corrective actionscan be taken

browser The request is directedto the appropriate WebMEmediator The WebME mediatoruses its metadata to determinewhich information repositories toaccess obtains the data from theinformation repositories andcombines the data into an answerfor the user query The resultsare sent back to the Web browserfor presentation

definitions for the system Thescript is processed checked forconsistency and the results arestored and made available to themediator Because the owners ofthe data are responsible forproviding the access methodsaccess to the data can be limitedBecause the metadata containsinformation about the physicallocation and context informationabout the data at remote

STN4

The WebMe Data Visualization ToolContinued from page 3

Status

The WebME system has been implemented and a prototype version isavailable at ftpftpcsumdedupubselroseannewebme

Figure 3 Baseline Characteristic Curve of Related Projects

Figure 4 Project that is Out of Scoperdquo

Continued on page 5

The baseline model may also beused to determine the impact ofprocess changes If a projectmanager has implemented aprocess change comparing theproject performance against thebaseline can assist in theunderstanding of how theprocess change has affected theattribute behavior of a project(Figure 5) Assume the projectmanager has incorporated codeinspections into the developmentprocess to uncover errors earlierin the development cycleComparing the current projectscharacteristic curve for errorswith the previous baseline modeldemonstrates the impact of theprocess change In Figure 5 theerrors for the current project arediscovered earlier than in thebaseline model

Author Contact Information

Prof Roseanne TesorieroDepartment of Computer

Science 4121 A V WilliamsUniversity of MarylandCollege Park MD 20742

Dr Marvin ZelkowitzFraunhofer Center forExperimental Software

EngineeringUniversity of MarylandCollege Park MD 20742

mvzcsumdedu

STN5

References[1] Basili V M Zelkowitz F McGarry J Page S Waligora and R Pajerski SELrsquos Software Process-ImprovementProgram IEEE Software 12 6 (1995) 83-87

[2] Tesoriero R and M V Zelkowitz ldquoA Model of Noisy Software Engineering Datardquo (Status Report) ACMIEEEInternational Conference on Software Engineering Kyoto Japan (April 1998) 461-464

Continued from page 4

Figure 5 Corrective Action Taken on Project that is ldquoOut of Scoperdquo

Collaborative Software Engineering on the Web Introducing WebDAVE James Whitehead Jr - University of California Irvine

Introduction

Software development today oftentakes place among multiple groupsof software engineers who aregeographically dispersed A keychallenge in supporting thesedistributed software developmentteams is making the softwareremotely accessible and editablevia the Internet addressing issuessuch as overwrite preventionsecurity authentication accesscontrol and reliable operationacross high-latency networkconnections

The Web Distributed Authoringand Versioning (WebDAV)

working group of the InternetEngineering Task Force (IETF)has taken on the challenge ofsupporting collaborative softwareengineering on the Web byextending the core networkprotocol of the Web theHypertext Transfer Protocol(HTTP) [4] to support remotesoftware development A team atthe University of CaliforniaIrvine working under a grantfrom the Defense AdvancedResearch Projects Agency(DARPA) EvolutionaryDevelopment of ComplexSystems program broughttogether interested parties from

academia and industry to formthe IETF working group anddevelop the WebDAVDistributed Authoring Protocol

WebDAV provides many benefits

bull Development teams cancollaboratively developsoftware artifacts in-place onthe Web using locking toprevent overwrite conflictsDue to the distributed natureof the Web these workgroups can have membersfrom within the sameorganization or acrossorganizational boundaries

Continued on page 6

STN6

Introducing WebDAVContinued from page 5

copy of Windows 2000 ServerThe popular Apache web serveralso has WebDAV capability inthe freely available mod_davmodule Additional supportcomes from IBM XeroxNovell DataChannelCyberTeams and DigitalCreations who have allannounced product plans basedon WebDAV WebDAV is clearlya standard that has strongcorporate and open sourcesupport

CapabilitiesThe WebDAV DistributedAuthoring Protocol defines a setof extensions to the HypertextTransfer Protocol (HTTP) for thefollowing capabilities

bull Overwrite preventionKeeping more than oneperson from working on adocument at the same timeThis prevents the ldquolostupdate problemrdquo in whichmodifications are lost as firstone developer then anotherwrites changes withoutmerging the otherdeveloperrsquos work

WebDAV provides facilities forboth shared and exclusivelocking This dual lock supportprovides sufficiently flexiblelocks to accommodate a widerange of collaborations withshared locks best supportingcollaborators who have a lot ofawareness of each otherrsquosactivities and exclusive locksproviding a more stringentguarantee of conflict avoidance

for less aware collaborators orduring periods of highcontention for a softwaredevelopment artifact Locks mayhave a scope of a single artifactor a hierarchy of artifacts suchas a source code tree A lockdiscovery mechanism (aWebDAV property) allowsauthors to find out if any locksexist on a Web-accessibleartifact Since the Web isdesigned so that no lock isrequired to read a Web pagethere is no concept of a readlock

bull Properties Creationremoval and querying ofinformation about Web-accessible artifacts such asits author last modified dateetc Also included is theability to make hypertextlinks between artifacts of anycontent type

WebDAV properties are namevalue pairs where the name is aUniform Resource Locator(URL) and the value is asequence of well-formedExtensible Markup Language(XML) [2] elements UsingURLs as property namesprovides a globally uniqueproperty namespace Sinceproperty names can be URLswhich have a domain name as acomponent of the URL propertynames can be given uniquenesswithout central registration byusing URL property nameschosen from within a domainwhose name is controlled by theparty

Continued on page 7

bull All the types of artifacts in atypical software developmentlifecycle can be edited usingWebDAV includingrequirements designdocuments test cases codeand more So whileWebDAV provides HTML [7]and XML [2] authoringsupport it just as easilysupports authoring of existingword processing spreadsheettext graphics and all otherformats

bull WebDAV and HTTP providea common interface to a widerange of repositories such asconfiguration managementfile systems databasesdocument management etcIn essence WebDAV makesthe Web look like a large-grain network-accessible filesystem But unlike aconventional file system aWebDAV-enabled repositoryprovides Internet access andallows all ldquofilesrdquo to beviewed using a standard Webbrowser

In November of 1998 the IETFapproved the WebDAVDistributed Authoring Protocol[6] clearing the way for broadindustry and open source supportMicrosoft is a major earlysupporter of the protocolproviding WebDAV support inInternet Explorer 5 (in its ldquoWebFoldersrdquo feature) in Office 2000(Word and Excel) and on theserver side in its InternetInformation Services (IIS) serverwhich comes bundled with every

STN7

defining the property So forexample a company whichcontrols a given domain namelike ldquowidgetscomrdquo can chose aproperty name from within thisdomain like ldquowidgetscompropertiescolorrdquo

Using XML to encode the valueof properties provides threemajor benefits First isextensibility Since all contentwithin XML is encoded betweenstart and end tags it is easy toadd additional elements to aproperty by inserting new taggedcontent Internationalization isthe second major benefit SinceXML mandates support for theUTF-8 and UTF-16 encodings ofthe ISO 10646 characterencoding standard as well aslanguage tagging properties canexpress content in the vastmajority of human languagesFinally by using XMLWebDAV properties can supportother metadata activities whichare also based on XML such asthe Resource DescriptionFramework (RDF) underdevelopment at the W3C

bull Name space managementCreation removal andautomatic consistencymaintenance of collectionscontaining sets of softwaredevelopment artifacts Alsothe ability to copy and moveWeb-accessible artifacts andto receive a listing ofartifacts in a collection(similar to a directory listingin a file system)

Work-in-ProgressCurrent work-in-progress withinWebDAV focuses on theseadditional capabilities

bull Version management Theability to store importantrevisions of a softwareartifacts for later retrievalVersion management canalso support collaboration byallowing two or more authorsto work on the same artifactin parallel tracks Automaticversioning recordssuccessive modifications toan artifact made byversioning unaware(ldquodownlevelrdquo) clientsConfiguration managementtracks versioned collectionsof versioned artifacts animportant capability forreverting to previouslyreleased software packagesor for tracking manyresources simultaneously

A new working group buildingon initial design work performedby the WebDAV working groupis currently being formed in theIETF to develop a protocol forWeb versioning andconfiguration managementCalled the Delta-V workinggroup it is already wellunderway towards its goal of afinished protocol specification inmid-2000 A snapshot of work inprogress can be found byexamining the grouprsquos goalsdocument [10] and the currentversioning and configurationmanagement protocolspecification [3]

bull Advanced CollectionsReferential resources act likesymbolic links in filesystems while orderedcollections maintain a client-specified ordering ofresources in a collectionwithout private agreements

bull Access Control The ability tolimit the access rights of agiven authenticated principalon a given artifact remotelyvia the Internet WebDAVassumes the existence of butdoes not specify strongauthentication technologyand today WebDAV servershave server-specific accesscontrols

A strongly related effort toWebDAV is the IETFrsquos DAVSearching and Locating (DASL)working group which isdeveloping an interoperablemeans of searching a repositorywhich is compliant with theWebDAV object model andwhich organizes its artifacts intoURL hierarchies The maincapability of DASL is searching

bull Searching Client specifiedserver-executed queries tolocate artifacts based upontheir property values and textcontent This is an importantcapability for softwaredevelopment supportingqueries like ldquowhere is thisfunction usedrdquo or ldquowhere isthis variable definedrdquo

DASL is working to developextensions to the WebDAVDistributed Authoring Protocol

Continued from page 6

Continued on page 8

STN8

Introducing WebDAVContinued from page 7

specification (and hence toHTTP) for searching WebDAVrepositories DASL has arequirements document [9] andprotocol document [8] which arestill the subject of intense effortwithin the DASL group

Application

Taken together the WebDAVextensions to HTTP provide thestandard needed to make the Weba writable collaborative mediumWhat does this mean for Web-based software developmentAlthough the future isnotoriously hard to predict hereare some likely outcomes of theadoption of WebDAV AsWebDAV technology isdeployed it will initially have itslargest impact on small tomedium sized developmentteams which support WebDAVallowing their developmentpractices to coalesce around alocal intranet gaining itsadvantages for remote access ofsoftware developmentinformation within theorganization Over time ascritical mass grows WebDAVwill also dramatically reduce theaccidental costs of collaborationbetween development teams andbetween developmentorganizations Since WebDAVsupports multiple informationrepositories it acts as a commonbridge across these storesproviding a convenientmechanism for integrating data ina software development

environment Due to thisWebDAV shows significantpromise as an infrastructure fordevelopment of distributedsoftware engineeringenvironments a topic exploredin a recent Communications ofthe ACM article [5]

While WebDAV todaysignificantly reduces the burdenof cross-team collaborativesoftware development theprotocol will really start to shineonce the Web versioningprotocol is finished Thenremote software engineeringteams will be able to remotelyedit software artifacts whilekeeping track of important statesof these documents and whiletracking configurations of theseversioned objects Since majorconfiguration managementvendors such as RationalMerant (Intersolv) Microsoftand IBM are working together todefine this standard broad toolsupport can be expected

One of the nice aspects ofWebDAV technology is thatcompelling WebDAV clients andservers are available today freeof charge ready for initialevaluation The Apachemod_dav module can bedownloaded atwwwwebdavorgmod_dav andworks with Apache servers 134and higher The Apache serversoftware is freely available fordownload at wwwapacheorgOr use one of the WebDAV test

servers listed atwwwwebdavorgprojects Forclient software try InternetExplorer 5rsquos Web Folders feature(IE5 can be downloaded fromwwwmicrosoftcomwindowsieie5defaultasp) or download theWebDAV Explorer a Java-basedclient developed at the Universityof California Irvine fromwwwicsuciedu~webdav

For More Information

Working groups of the InternetEngineering Task Force arecompletely open and may bejoined by subscribing to their E-mail discussion list If you wishto participate in the discussionson WebDAV topics you may jointhe mailing list by E-mailing amessage with the subjectldquosubscriberdquo to w3c-dist-auth-requestw3org

The home page for the WebDAVgroup is at URLwwwicsuciedupubietfwebdavwhich contains links to current working drafts E-mail list archives and backgroundmaterial Another excellentsource of WebDAV informationis the WebDAV Resources pageat wwwwebdavorg maintainedby Greg Stein The related DAVSearching and Locating (DASL)working group has its web pageat URL wwwicsuciedupubietfdasl and a mailing listwhich may be joined by sendinga message with subjectldquosubscriberdquo to www-webdav-dasl-requestw3org

Continued on page 9

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 2: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN2

The WebMe Data Visualization ToolRoseanne Tesoriero and Marvin V ZelkowitzUniversity of Maryland Fraunhofer Center for Experimental Software Engineering

Introduction

The ubiquity of the World WideWeb and the increased presenceof the Internet in todayrsquosmarketplace have led to anenvironment where software isbeing developed ingeographically distributedlocations This type ofdistributed developmentenvironment raises several newchallenges in the area ofsoftware project management

Measurement data has been usedto control and improve thesoftware process and itsproducts For example theSoftware EngineeringLaboratory of NASArsquos GoddardSpace Flight Center (NASASEL) has been using the QualityImprovement Paradigm alongwith the GoalQuestionMetricparadigm to help evaluatecontrol and improve softwareprocesses and products for overtwenty years [1] The use ofmeasurement data to build andpackage experience is a keyaspect in this approach Oneway in which experience ispackaged is in a baseline modelA baseline model is created byclustering project data fromsimilar projects together todescribe the expected behaviorfor this class of projectsManagement can then compare anew project with this baselinemodel in order to better

understand deviations from thebaseline For example a projectwith too few errors per line ofcode may either represent asuperior development process orinsufficient testing

When dealing with a distributeddevelopment environment theuse of measurement data to buildknowledge and experience forsoftware project managementpresents several new challengesWhen development occurs atphysically separated locations

the data may be geographicallyseparate as well If severalorganizations are involved in thedevelopment (such as withsubcontract management) theorganizational cultures may bedifferent and the data may becollected using differentcollection mechanisms leadingto variations in the formatsCreating baselines and

comparing different projectsbecomes more difficult due to thediffering interpretations to thecollected data When severalorganizations are involved in thedevelopment privacy becomesan issue Each organization doesnot want to give the othersunlimited access to itsproprietary data

The Web MeasurementEnvironment (WebME) has beendeveloped to address thechallenges of using measurementdata for experience-based projectmanagement in a distributeddevelopment environment Thisshort discussion explains how thetechniques incorporated into theWebME system allow projectmanagers to use data from adistributed environment A briefoverview of the WebME systemand a description of the analysistechnique used to build a baselinefrom similar projects are given

WebME solution

The WebME system wasdesigned to build baselinemodels from data collected ingeographically distributedenvironments In particular itcan be used to combine andanalyze time-series data collectedfrom software projects

The architecture of the WebMEsystem is shown in Figure 1

Continued on page 3

Figure 1 WebMe Architecture

STN3

Continued from page 2

The WebME system uses amediated architecture whichhorizontally partitions thestructure of the architecture intothree layers the informationrepositories mediatinginformation servers and the end-user application layer In theWebME context the end-usersubmits a request through a Web

Continued on page 4

Figure 2 Characteristic Curve of Project under Study

Metadata is created by using theWebME scripting language TheWebME scripting languageallows the system configurer tospecify the structure of thearchitecture the data that will beviewable with the system andthe access methods for the dataThe configurer of the systemcreates a script containing the

repositories data collected atgeographically separatedlocations and stored in differentformats can be combinedconsistently

It is important not only tocombine the data in a consistentmanner but to use the data tosupport project managementThe analysis technique supportedin WebME [2] uses data frompast similar projects to build abaseline of expected behavior foran attribute over time First acharacteristic curve is generatedfor the project of interest(Figure 2) The characteristiccurve describes the behavior ofan attribute over the projectlifecycle The baseline model forthe attribute is created using thecharacteristic curves of projectssimilar to the project of interest(Figure 3)

The baseline model can be usedto support project management inseveral ways For example if aproject is performing differentlyfrom the baseline it serves asignal for the project manager InFigure 4 the baseline model ofthe number of errors per weekand the current projects errorsper week are plotted The currentproject has encountered moreerrors than expected The projectmanager should investigate todetermine the cause of thedeviation If the deviation isundesirable corrective actionscan be taken

browser The request is directedto the appropriate WebMEmediator The WebME mediatoruses its metadata to determinewhich information repositories toaccess obtains the data from theinformation repositories andcombines the data into an answerfor the user query The resultsare sent back to the Web browserfor presentation

definitions for the system Thescript is processed checked forconsistency and the results arestored and made available to themediator Because the owners ofthe data are responsible forproviding the access methodsaccess to the data can be limitedBecause the metadata containsinformation about the physicallocation and context informationabout the data at remote

STN4

The WebMe Data Visualization ToolContinued from page 3

Status

The WebME system has been implemented and a prototype version isavailable at ftpftpcsumdedupubselroseannewebme

Figure 3 Baseline Characteristic Curve of Related Projects

Figure 4 Project that is Out of Scoperdquo

Continued on page 5

The baseline model may also beused to determine the impact ofprocess changes If a projectmanager has implemented aprocess change comparing theproject performance against thebaseline can assist in theunderstanding of how theprocess change has affected theattribute behavior of a project(Figure 5) Assume the projectmanager has incorporated codeinspections into the developmentprocess to uncover errors earlierin the development cycleComparing the current projectscharacteristic curve for errorswith the previous baseline modeldemonstrates the impact of theprocess change In Figure 5 theerrors for the current project arediscovered earlier than in thebaseline model

Author Contact Information

Prof Roseanne TesorieroDepartment of Computer

Science 4121 A V WilliamsUniversity of MarylandCollege Park MD 20742

Dr Marvin ZelkowitzFraunhofer Center forExperimental Software

EngineeringUniversity of MarylandCollege Park MD 20742

mvzcsumdedu

STN5

References[1] Basili V M Zelkowitz F McGarry J Page S Waligora and R Pajerski SELrsquos Software Process-ImprovementProgram IEEE Software 12 6 (1995) 83-87

[2] Tesoriero R and M V Zelkowitz ldquoA Model of Noisy Software Engineering Datardquo (Status Report) ACMIEEEInternational Conference on Software Engineering Kyoto Japan (April 1998) 461-464

Continued from page 4

Figure 5 Corrective Action Taken on Project that is ldquoOut of Scoperdquo

Collaborative Software Engineering on the Web Introducing WebDAVE James Whitehead Jr - University of California Irvine

Introduction

Software development today oftentakes place among multiple groupsof software engineers who aregeographically dispersed A keychallenge in supporting thesedistributed software developmentteams is making the softwareremotely accessible and editablevia the Internet addressing issuessuch as overwrite preventionsecurity authentication accesscontrol and reliable operationacross high-latency networkconnections

The Web Distributed Authoringand Versioning (WebDAV)

working group of the InternetEngineering Task Force (IETF)has taken on the challenge ofsupporting collaborative softwareengineering on the Web byextending the core networkprotocol of the Web theHypertext Transfer Protocol(HTTP) [4] to support remotesoftware development A team atthe University of CaliforniaIrvine working under a grantfrom the Defense AdvancedResearch Projects Agency(DARPA) EvolutionaryDevelopment of ComplexSystems program broughttogether interested parties from

academia and industry to formthe IETF working group anddevelop the WebDAVDistributed Authoring Protocol

WebDAV provides many benefits

bull Development teams cancollaboratively developsoftware artifacts in-place onthe Web using locking toprevent overwrite conflictsDue to the distributed natureof the Web these workgroups can have membersfrom within the sameorganization or acrossorganizational boundaries

Continued on page 6

STN6

Introducing WebDAVContinued from page 5

copy of Windows 2000 ServerThe popular Apache web serveralso has WebDAV capability inthe freely available mod_davmodule Additional supportcomes from IBM XeroxNovell DataChannelCyberTeams and DigitalCreations who have allannounced product plans basedon WebDAV WebDAV is clearlya standard that has strongcorporate and open sourcesupport

CapabilitiesThe WebDAV DistributedAuthoring Protocol defines a setof extensions to the HypertextTransfer Protocol (HTTP) for thefollowing capabilities

bull Overwrite preventionKeeping more than oneperson from working on adocument at the same timeThis prevents the ldquolostupdate problemrdquo in whichmodifications are lost as firstone developer then anotherwrites changes withoutmerging the otherdeveloperrsquos work

WebDAV provides facilities forboth shared and exclusivelocking This dual lock supportprovides sufficiently flexiblelocks to accommodate a widerange of collaborations withshared locks best supportingcollaborators who have a lot ofawareness of each otherrsquosactivities and exclusive locksproviding a more stringentguarantee of conflict avoidance

for less aware collaborators orduring periods of highcontention for a softwaredevelopment artifact Locks mayhave a scope of a single artifactor a hierarchy of artifacts suchas a source code tree A lockdiscovery mechanism (aWebDAV property) allowsauthors to find out if any locksexist on a Web-accessibleartifact Since the Web isdesigned so that no lock isrequired to read a Web pagethere is no concept of a readlock

bull Properties Creationremoval and querying ofinformation about Web-accessible artifacts such asits author last modified dateetc Also included is theability to make hypertextlinks between artifacts of anycontent type

WebDAV properties are namevalue pairs where the name is aUniform Resource Locator(URL) and the value is asequence of well-formedExtensible Markup Language(XML) [2] elements UsingURLs as property namesprovides a globally uniqueproperty namespace Sinceproperty names can be URLswhich have a domain name as acomponent of the URL propertynames can be given uniquenesswithout central registration byusing URL property nameschosen from within a domainwhose name is controlled by theparty

Continued on page 7

bull All the types of artifacts in atypical software developmentlifecycle can be edited usingWebDAV includingrequirements designdocuments test cases codeand more So whileWebDAV provides HTML [7]and XML [2] authoringsupport it just as easilysupports authoring of existingword processing spreadsheettext graphics and all otherformats

bull WebDAV and HTTP providea common interface to a widerange of repositories such asconfiguration managementfile systems databasesdocument management etcIn essence WebDAV makesthe Web look like a large-grain network-accessible filesystem But unlike aconventional file system aWebDAV-enabled repositoryprovides Internet access andallows all ldquofilesrdquo to beviewed using a standard Webbrowser

In November of 1998 the IETFapproved the WebDAVDistributed Authoring Protocol[6] clearing the way for broadindustry and open source supportMicrosoft is a major earlysupporter of the protocolproviding WebDAV support inInternet Explorer 5 (in its ldquoWebFoldersrdquo feature) in Office 2000(Word and Excel) and on theserver side in its InternetInformation Services (IIS) serverwhich comes bundled with every

STN7

defining the property So forexample a company whichcontrols a given domain namelike ldquowidgetscomrdquo can chose aproperty name from within thisdomain like ldquowidgetscompropertiescolorrdquo

Using XML to encode the valueof properties provides threemajor benefits First isextensibility Since all contentwithin XML is encoded betweenstart and end tags it is easy toadd additional elements to aproperty by inserting new taggedcontent Internationalization isthe second major benefit SinceXML mandates support for theUTF-8 and UTF-16 encodings ofthe ISO 10646 characterencoding standard as well aslanguage tagging properties canexpress content in the vastmajority of human languagesFinally by using XMLWebDAV properties can supportother metadata activities whichare also based on XML such asthe Resource DescriptionFramework (RDF) underdevelopment at the W3C

bull Name space managementCreation removal andautomatic consistencymaintenance of collectionscontaining sets of softwaredevelopment artifacts Alsothe ability to copy and moveWeb-accessible artifacts andto receive a listing ofartifacts in a collection(similar to a directory listingin a file system)

Work-in-ProgressCurrent work-in-progress withinWebDAV focuses on theseadditional capabilities

bull Version management Theability to store importantrevisions of a softwareartifacts for later retrievalVersion management canalso support collaboration byallowing two or more authorsto work on the same artifactin parallel tracks Automaticversioning recordssuccessive modifications toan artifact made byversioning unaware(ldquodownlevelrdquo) clientsConfiguration managementtracks versioned collectionsof versioned artifacts animportant capability forreverting to previouslyreleased software packagesor for tracking manyresources simultaneously

A new working group buildingon initial design work performedby the WebDAV working groupis currently being formed in theIETF to develop a protocol forWeb versioning andconfiguration managementCalled the Delta-V workinggroup it is already wellunderway towards its goal of afinished protocol specification inmid-2000 A snapshot of work inprogress can be found byexamining the grouprsquos goalsdocument [10] and the currentversioning and configurationmanagement protocolspecification [3]

bull Advanced CollectionsReferential resources act likesymbolic links in filesystems while orderedcollections maintain a client-specified ordering ofresources in a collectionwithout private agreements

bull Access Control The ability tolimit the access rights of agiven authenticated principalon a given artifact remotelyvia the Internet WebDAVassumes the existence of butdoes not specify strongauthentication technologyand today WebDAV servershave server-specific accesscontrols

A strongly related effort toWebDAV is the IETFrsquos DAVSearching and Locating (DASL)working group which isdeveloping an interoperablemeans of searching a repositorywhich is compliant with theWebDAV object model andwhich organizes its artifacts intoURL hierarchies The maincapability of DASL is searching

bull Searching Client specifiedserver-executed queries tolocate artifacts based upontheir property values and textcontent This is an importantcapability for softwaredevelopment supportingqueries like ldquowhere is thisfunction usedrdquo or ldquowhere isthis variable definedrdquo

DASL is working to developextensions to the WebDAVDistributed Authoring Protocol

Continued from page 6

Continued on page 8

STN8

Introducing WebDAVContinued from page 7

specification (and hence toHTTP) for searching WebDAVrepositories DASL has arequirements document [9] andprotocol document [8] which arestill the subject of intense effortwithin the DASL group

Application

Taken together the WebDAVextensions to HTTP provide thestandard needed to make the Weba writable collaborative mediumWhat does this mean for Web-based software developmentAlthough the future isnotoriously hard to predict hereare some likely outcomes of theadoption of WebDAV AsWebDAV technology isdeployed it will initially have itslargest impact on small tomedium sized developmentteams which support WebDAVallowing their developmentpractices to coalesce around alocal intranet gaining itsadvantages for remote access ofsoftware developmentinformation within theorganization Over time ascritical mass grows WebDAVwill also dramatically reduce theaccidental costs of collaborationbetween development teams andbetween developmentorganizations Since WebDAVsupports multiple informationrepositories it acts as a commonbridge across these storesproviding a convenientmechanism for integrating data ina software development

environment Due to thisWebDAV shows significantpromise as an infrastructure fordevelopment of distributedsoftware engineeringenvironments a topic exploredin a recent Communications ofthe ACM article [5]

While WebDAV todaysignificantly reduces the burdenof cross-team collaborativesoftware development theprotocol will really start to shineonce the Web versioningprotocol is finished Thenremote software engineeringteams will be able to remotelyedit software artifacts whilekeeping track of important statesof these documents and whiletracking configurations of theseversioned objects Since majorconfiguration managementvendors such as RationalMerant (Intersolv) Microsoftand IBM are working together todefine this standard broad toolsupport can be expected

One of the nice aspects ofWebDAV technology is thatcompelling WebDAV clients andservers are available today freeof charge ready for initialevaluation The Apachemod_dav module can bedownloaded atwwwwebdavorgmod_dav andworks with Apache servers 134and higher The Apache serversoftware is freely available fordownload at wwwapacheorgOr use one of the WebDAV test

servers listed atwwwwebdavorgprojects Forclient software try InternetExplorer 5rsquos Web Folders feature(IE5 can be downloaded fromwwwmicrosoftcomwindowsieie5defaultasp) or download theWebDAV Explorer a Java-basedclient developed at the Universityof California Irvine fromwwwicsuciedu~webdav

For More Information

Working groups of the InternetEngineering Task Force arecompletely open and may bejoined by subscribing to their E-mail discussion list If you wishto participate in the discussionson WebDAV topics you may jointhe mailing list by E-mailing amessage with the subjectldquosubscriberdquo to w3c-dist-auth-requestw3org

The home page for the WebDAVgroup is at URLwwwicsuciedupubietfwebdavwhich contains links to current working drafts E-mail list archives and backgroundmaterial Another excellentsource of WebDAV informationis the WebDAV Resources pageat wwwwebdavorg maintainedby Greg Stein The related DAVSearching and Locating (DASL)working group has its web pageat URL wwwicsuciedupubietfdasl and a mailing listwhich may be joined by sendinga message with subjectldquosubscriberdquo to www-webdav-dasl-requestw3org

Continued on page 9

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 3: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN3

Continued from page 2

The WebME system uses amediated architecture whichhorizontally partitions thestructure of the architecture intothree layers the informationrepositories mediatinginformation servers and the end-user application layer In theWebME context the end-usersubmits a request through a Web

Continued on page 4

Figure 2 Characteristic Curve of Project under Study

Metadata is created by using theWebME scripting language TheWebME scripting languageallows the system configurer tospecify the structure of thearchitecture the data that will beviewable with the system andthe access methods for the dataThe configurer of the systemcreates a script containing the

repositories data collected atgeographically separatedlocations and stored in differentformats can be combinedconsistently

It is important not only tocombine the data in a consistentmanner but to use the data tosupport project managementThe analysis technique supportedin WebME [2] uses data frompast similar projects to build abaseline of expected behavior foran attribute over time First acharacteristic curve is generatedfor the project of interest(Figure 2) The characteristiccurve describes the behavior ofan attribute over the projectlifecycle The baseline model forthe attribute is created using thecharacteristic curves of projectssimilar to the project of interest(Figure 3)

The baseline model can be usedto support project management inseveral ways For example if aproject is performing differentlyfrom the baseline it serves asignal for the project manager InFigure 4 the baseline model ofthe number of errors per weekand the current projects errorsper week are plotted The currentproject has encountered moreerrors than expected The projectmanager should investigate todetermine the cause of thedeviation If the deviation isundesirable corrective actionscan be taken

browser The request is directedto the appropriate WebMEmediator The WebME mediatoruses its metadata to determinewhich information repositories toaccess obtains the data from theinformation repositories andcombines the data into an answerfor the user query The resultsare sent back to the Web browserfor presentation

definitions for the system Thescript is processed checked forconsistency and the results arestored and made available to themediator Because the owners ofthe data are responsible forproviding the access methodsaccess to the data can be limitedBecause the metadata containsinformation about the physicallocation and context informationabout the data at remote

STN4

The WebMe Data Visualization ToolContinued from page 3

Status

The WebME system has been implemented and a prototype version isavailable at ftpftpcsumdedupubselroseannewebme

Figure 3 Baseline Characteristic Curve of Related Projects

Figure 4 Project that is Out of Scoperdquo

Continued on page 5

The baseline model may also beused to determine the impact ofprocess changes If a projectmanager has implemented aprocess change comparing theproject performance against thebaseline can assist in theunderstanding of how theprocess change has affected theattribute behavior of a project(Figure 5) Assume the projectmanager has incorporated codeinspections into the developmentprocess to uncover errors earlierin the development cycleComparing the current projectscharacteristic curve for errorswith the previous baseline modeldemonstrates the impact of theprocess change In Figure 5 theerrors for the current project arediscovered earlier than in thebaseline model

Author Contact Information

Prof Roseanne TesorieroDepartment of Computer

Science 4121 A V WilliamsUniversity of MarylandCollege Park MD 20742

Dr Marvin ZelkowitzFraunhofer Center forExperimental Software

EngineeringUniversity of MarylandCollege Park MD 20742

mvzcsumdedu

STN5

References[1] Basili V M Zelkowitz F McGarry J Page S Waligora and R Pajerski SELrsquos Software Process-ImprovementProgram IEEE Software 12 6 (1995) 83-87

[2] Tesoriero R and M V Zelkowitz ldquoA Model of Noisy Software Engineering Datardquo (Status Report) ACMIEEEInternational Conference on Software Engineering Kyoto Japan (April 1998) 461-464

Continued from page 4

Figure 5 Corrective Action Taken on Project that is ldquoOut of Scoperdquo

Collaborative Software Engineering on the Web Introducing WebDAVE James Whitehead Jr - University of California Irvine

Introduction

Software development today oftentakes place among multiple groupsof software engineers who aregeographically dispersed A keychallenge in supporting thesedistributed software developmentteams is making the softwareremotely accessible and editablevia the Internet addressing issuessuch as overwrite preventionsecurity authentication accesscontrol and reliable operationacross high-latency networkconnections

The Web Distributed Authoringand Versioning (WebDAV)

working group of the InternetEngineering Task Force (IETF)has taken on the challenge ofsupporting collaborative softwareengineering on the Web byextending the core networkprotocol of the Web theHypertext Transfer Protocol(HTTP) [4] to support remotesoftware development A team atthe University of CaliforniaIrvine working under a grantfrom the Defense AdvancedResearch Projects Agency(DARPA) EvolutionaryDevelopment of ComplexSystems program broughttogether interested parties from

academia and industry to formthe IETF working group anddevelop the WebDAVDistributed Authoring Protocol

WebDAV provides many benefits

bull Development teams cancollaboratively developsoftware artifacts in-place onthe Web using locking toprevent overwrite conflictsDue to the distributed natureof the Web these workgroups can have membersfrom within the sameorganization or acrossorganizational boundaries

Continued on page 6

STN6

Introducing WebDAVContinued from page 5

copy of Windows 2000 ServerThe popular Apache web serveralso has WebDAV capability inthe freely available mod_davmodule Additional supportcomes from IBM XeroxNovell DataChannelCyberTeams and DigitalCreations who have allannounced product plans basedon WebDAV WebDAV is clearlya standard that has strongcorporate and open sourcesupport

CapabilitiesThe WebDAV DistributedAuthoring Protocol defines a setof extensions to the HypertextTransfer Protocol (HTTP) for thefollowing capabilities

bull Overwrite preventionKeeping more than oneperson from working on adocument at the same timeThis prevents the ldquolostupdate problemrdquo in whichmodifications are lost as firstone developer then anotherwrites changes withoutmerging the otherdeveloperrsquos work

WebDAV provides facilities forboth shared and exclusivelocking This dual lock supportprovides sufficiently flexiblelocks to accommodate a widerange of collaborations withshared locks best supportingcollaborators who have a lot ofawareness of each otherrsquosactivities and exclusive locksproviding a more stringentguarantee of conflict avoidance

for less aware collaborators orduring periods of highcontention for a softwaredevelopment artifact Locks mayhave a scope of a single artifactor a hierarchy of artifacts suchas a source code tree A lockdiscovery mechanism (aWebDAV property) allowsauthors to find out if any locksexist on a Web-accessibleartifact Since the Web isdesigned so that no lock isrequired to read a Web pagethere is no concept of a readlock

bull Properties Creationremoval and querying ofinformation about Web-accessible artifacts such asits author last modified dateetc Also included is theability to make hypertextlinks between artifacts of anycontent type

WebDAV properties are namevalue pairs where the name is aUniform Resource Locator(URL) and the value is asequence of well-formedExtensible Markup Language(XML) [2] elements UsingURLs as property namesprovides a globally uniqueproperty namespace Sinceproperty names can be URLswhich have a domain name as acomponent of the URL propertynames can be given uniquenesswithout central registration byusing URL property nameschosen from within a domainwhose name is controlled by theparty

Continued on page 7

bull All the types of artifacts in atypical software developmentlifecycle can be edited usingWebDAV includingrequirements designdocuments test cases codeand more So whileWebDAV provides HTML [7]and XML [2] authoringsupport it just as easilysupports authoring of existingword processing spreadsheettext graphics and all otherformats

bull WebDAV and HTTP providea common interface to a widerange of repositories such asconfiguration managementfile systems databasesdocument management etcIn essence WebDAV makesthe Web look like a large-grain network-accessible filesystem But unlike aconventional file system aWebDAV-enabled repositoryprovides Internet access andallows all ldquofilesrdquo to beviewed using a standard Webbrowser

In November of 1998 the IETFapproved the WebDAVDistributed Authoring Protocol[6] clearing the way for broadindustry and open source supportMicrosoft is a major earlysupporter of the protocolproviding WebDAV support inInternet Explorer 5 (in its ldquoWebFoldersrdquo feature) in Office 2000(Word and Excel) and on theserver side in its InternetInformation Services (IIS) serverwhich comes bundled with every

STN7

defining the property So forexample a company whichcontrols a given domain namelike ldquowidgetscomrdquo can chose aproperty name from within thisdomain like ldquowidgetscompropertiescolorrdquo

Using XML to encode the valueof properties provides threemajor benefits First isextensibility Since all contentwithin XML is encoded betweenstart and end tags it is easy toadd additional elements to aproperty by inserting new taggedcontent Internationalization isthe second major benefit SinceXML mandates support for theUTF-8 and UTF-16 encodings ofthe ISO 10646 characterencoding standard as well aslanguage tagging properties canexpress content in the vastmajority of human languagesFinally by using XMLWebDAV properties can supportother metadata activities whichare also based on XML such asthe Resource DescriptionFramework (RDF) underdevelopment at the W3C

bull Name space managementCreation removal andautomatic consistencymaintenance of collectionscontaining sets of softwaredevelopment artifacts Alsothe ability to copy and moveWeb-accessible artifacts andto receive a listing ofartifacts in a collection(similar to a directory listingin a file system)

Work-in-ProgressCurrent work-in-progress withinWebDAV focuses on theseadditional capabilities

bull Version management Theability to store importantrevisions of a softwareartifacts for later retrievalVersion management canalso support collaboration byallowing two or more authorsto work on the same artifactin parallel tracks Automaticversioning recordssuccessive modifications toan artifact made byversioning unaware(ldquodownlevelrdquo) clientsConfiguration managementtracks versioned collectionsof versioned artifacts animportant capability forreverting to previouslyreleased software packagesor for tracking manyresources simultaneously

A new working group buildingon initial design work performedby the WebDAV working groupis currently being formed in theIETF to develop a protocol forWeb versioning andconfiguration managementCalled the Delta-V workinggroup it is already wellunderway towards its goal of afinished protocol specification inmid-2000 A snapshot of work inprogress can be found byexamining the grouprsquos goalsdocument [10] and the currentversioning and configurationmanagement protocolspecification [3]

bull Advanced CollectionsReferential resources act likesymbolic links in filesystems while orderedcollections maintain a client-specified ordering ofresources in a collectionwithout private agreements

bull Access Control The ability tolimit the access rights of agiven authenticated principalon a given artifact remotelyvia the Internet WebDAVassumes the existence of butdoes not specify strongauthentication technologyand today WebDAV servershave server-specific accesscontrols

A strongly related effort toWebDAV is the IETFrsquos DAVSearching and Locating (DASL)working group which isdeveloping an interoperablemeans of searching a repositorywhich is compliant with theWebDAV object model andwhich organizes its artifacts intoURL hierarchies The maincapability of DASL is searching

bull Searching Client specifiedserver-executed queries tolocate artifacts based upontheir property values and textcontent This is an importantcapability for softwaredevelopment supportingqueries like ldquowhere is thisfunction usedrdquo or ldquowhere isthis variable definedrdquo

DASL is working to developextensions to the WebDAVDistributed Authoring Protocol

Continued from page 6

Continued on page 8

STN8

Introducing WebDAVContinued from page 7

specification (and hence toHTTP) for searching WebDAVrepositories DASL has arequirements document [9] andprotocol document [8] which arestill the subject of intense effortwithin the DASL group

Application

Taken together the WebDAVextensions to HTTP provide thestandard needed to make the Weba writable collaborative mediumWhat does this mean for Web-based software developmentAlthough the future isnotoriously hard to predict hereare some likely outcomes of theadoption of WebDAV AsWebDAV technology isdeployed it will initially have itslargest impact on small tomedium sized developmentteams which support WebDAVallowing their developmentpractices to coalesce around alocal intranet gaining itsadvantages for remote access ofsoftware developmentinformation within theorganization Over time ascritical mass grows WebDAVwill also dramatically reduce theaccidental costs of collaborationbetween development teams andbetween developmentorganizations Since WebDAVsupports multiple informationrepositories it acts as a commonbridge across these storesproviding a convenientmechanism for integrating data ina software development

environment Due to thisWebDAV shows significantpromise as an infrastructure fordevelopment of distributedsoftware engineeringenvironments a topic exploredin a recent Communications ofthe ACM article [5]

While WebDAV todaysignificantly reduces the burdenof cross-team collaborativesoftware development theprotocol will really start to shineonce the Web versioningprotocol is finished Thenremote software engineeringteams will be able to remotelyedit software artifacts whilekeeping track of important statesof these documents and whiletracking configurations of theseversioned objects Since majorconfiguration managementvendors such as RationalMerant (Intersolv) Microsoftand IBM are working together todefine this standard broad toolsupport can be expected

One of the nice aspects ofWebDAV technology is thatcompelling WebDAV clients andservers are available today freeof charge ready for initialevaluation The Apachemod_dav module can bedownloaded atwwwwebdavorgmod_dav andworks with Apache servers 134and higher The Apache serversoftware is freely available fordownload at wwwapacheorgOr use one of the WebDAV test

servers listed atwwwwebdavorgprojects Forclient software try InternetExplorer 5rsquos Web Folders feature(IE5 can be downloaded fromwwwmicrosoftcomwindowsieie5defaultasp) or download theWebDAV Explorer a Java-basedclient developed at the Universityof California Irvine fromwwwicsuciedu~webdav

For More Information

Working groups of the InternetEngineering Task Force arecompletely open and may bejoined by subscribing to their E-mail discussion list If you wishto participate in the discussionson WebDAV topics you may jointhe mailing list by E-mailing amessage with the subjectldquosubscriberdquo to w3c-dist-auth-requestw3org

The home page for the WebDAVgroup is at URLwwwicsuciedupubietfwebdavwhich contains links to current working drafts E-mail list archives and backgroundmaterial Another excellentsource of WebDAV informationis the WebDAV Resources pageat wwwwebdavorg maintainedby Greg Stein The related DAVSearching and Locating (DASL)working group has its web pageat URL wwwicsuciedupubietfdasl and a mailing listwhich may be joined by sendinga message with subjectldquosubscriberdquo to www-webdav-dasl-requestw3org

Continued on page 9

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 4: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN4

The WebMe Data Visualization ToolContinued from page 3

Status

The WebME system has been implemented and a prototype version isavailable at ftpftpcsumdedupubselroseannewebme

Figure 3 Baseline Characteristic Curve of Related Projects

Figure 4 Project that is Out of Scoperdquo

Continued on page 5

The baseline model may also beused to determine the impact ofprocess changes If a projectmanager has implemented aprocess change comparing theproject performance against thebaseline can assist in theunderstanding of how theprocess change has affected theattribute behavior of a project(Figure 5) Assume the projectmanager has incorporated codeinspections into the developmentprocess to uncover errors earlierin the development cycleComparing the current projectscharacteristic curve for errorswith the previous baseline modeldemonstrates the impact of theprocess change In Figure 5 theerrors for the current project arediscovered earlier than in thebaseline model

Author Contact Information

Prof Roseanne TesorieroDepartment of Computer

Science 4121 A V WilliamsUniversity of MarylandCollege Park MD 20742

Dr Marvin ZelkowitzFraunhofer Center forExperimental Software

EngineeringUniversity of MarylandCollege Park MD 20742

mvzcsumdedu

STN5

References[1] Basili V M Zelkowitz F McGarry J Page S Waligora and R Pajerski SELrsquos Software Process-ImprovementProgram IEEE Software 12 6 (1995) 83-87

[2] Tesoriero R and M V Zelkowitz ldquoA Model of Noisy Software Engineering Datardquo (Status Report) ACMIEEEInternational Conference on Software Engineering Kyoto Japan (April 1998) 461-464

Continued from page 4

Figure 5 Corrective Action Taken on Project that is ldquoOut of Scoperdquo

Collaborative Software Engineering on the Web Introducing WebDAVE James Whitehead Jr - University of California Irvine

Introduction

Software development today oftentakes place among multiple groupsof software engineers who aregeographically dispersed A keychallenge in supporting thesedistributed software developmentteams is making the softwareremotely accessible and editablevia the Internet addressing issuessuch as overwrite preventionsecurity authentication accesscontrol and reliable operationacross high-latency networkconnections

The Web Distributed Authoringand Versioning (WebDAV)

working group of the InternetEngineering Task Force (IETF)has taken on the challenge ofsupporting collaborative softwareengineering on the Web byextending the core networkprotocol of the Web theHypertext Transfer Protocol(HTTP) [4] to support remotesoftware development A team atthe University of CaliforniaIrvine working under a grantfrom the Defense AdvancedResearch Projects Agency(DARPA) EvolutionaryDevelopment of ComplexSystems program broughttogether interested parties from

academia and industry to formthe IETF working group anddevelop the WebDAVDistributed Authoring Protocol

WebDAV provides many benefits

bull Development teams cancollaboratively developsoftware artifacts in-place onthe Web using locking toprevent overwrite conflictsDue to the distributed natureof the Web these workgroups can have membersfrom within the sameorganization or acrossorganizational boundaries

Continued on page 6

STN6

Introducing WebDAVContinued from page 5

copy of Windows 2000 ServerThe popular Apache web serveralso has WebDAV capability inthe freely available mod_davmodule Additional supportcomes from IBM XeroxNovell DataChannelCyberTeams and DigitalCreations who have allannounced product plans basedon WebDAV WebDAV is clearlya standard that has strongcorporate and open sourcesupport

CapabilitiesThe WebDAV DistributedAuthoring Protocol defines a setof extensions to the HypertextTransfer Protocol (HTTP) for thefollowing capabilities

bull Overwrite preventionKeeping more than oneperson from working on adocument at the same timeThis prevents the ldquolostupdate problemrdquo in whichmodifications are lost as firstone developer then anotherwrites changes withoutmerging the otherdeveloperrsquos work

WebDAV provides facilities forboth shared and exclusivelocking This dual lock supportprovides sufficiently flexiblelocks to accommodate a widerange of collaborations withshared locks best supportingcollaborators who have a lot ofawareness of each otherrsquosactivities and exclusive locksproviding a more stringentguarantee of conflict avoidance

for less aware collaborators orduring periods of highcontention for a softwaredevelopment artifact Locks mayhave a scope of a single artifactor a hierarchy of artifacts suchas a source code tree A lockdiscovery mechanism (aWebDAV property) allowsauthors to find out if any locksexist on a Web-accessibleartifact Since the Web isdesigned so that no lock isrequired to read a Web pagethere is no concept of a readlock

bull Properties Creationremoval and querying ofinformation about Web-accessible artifacts such asits author last modified dateetc Also included is theability to make hypertextlinks between artifacts of anycontent type

WebDAV properties are namevalue pairs where the name is aUniform Resource Locator(URL) and the value is asequence of well-formedExtensible Markup Language(XML) [2] elements UsingURLs as property namesprovides a globally uniqueproperty namespace Sinceproperty names can be URLswhich have a domain name as acomponent of the URL propertynames can be given uniquenesswithout central registration byusing URL property nameschosen from within a domainwhose name is controlled by theparty

Continued on page 7

bull All the types of artifacts in atypical software developmentlifecycle can be edited usingWebDAV includingrequirements designdocuments test cases codeand more So whileWebDAV provides HTML [7]and XML [2] authoringsupport it just as easilysupports authoring of existingword processing spreadsheettext graphics and all otherformats

bull WebDAV and HTTP providea common interface to a widerange of repositories such asconfiguration managementfile systems databasesdocument management etcIn essence WebDAV makesthe Web look like a large-grain network-accessible filesystem But unlike aconventional file system aWebDAV-enabled repositoryprovides Internet access andallows all ldquofilesrdquo to beviewed using a standard Webbrowser

In November of 1998 the IETFapproved the WebDAVDistributed Authoring Protocol[6] clearing the way for broadindustry and open source supportMicrosoft is a major earlysupporter of the protocolproviding WebDAV support inInternet Explorer 5 (in its ldquoWebFoldersrdquo feature) in Office 2000(Word and Excel) and on theserver side in its InternetInformation Services (IIS) serverwhich comes bundled with every

STN7

defining the property So forexample a company whichcontrols a given domain namelike ldquowidgetscomrdquo can chose aproperty name from within thisdomain like ldquowidgetscompropertiescolorrdquo

Using XML to encode the valueof properties provides threemajor benefits First isextensibility Since all contentwithin XML is encoded betweenstart and end tags it is easy toadd additional elements to aproperty by inserting new taggedcontent Internationalization isthe second major benefit SinceXML mandates support for theUTF-8 and UTF-16 encodings ofthe ISO 10646 characterencoding standard as well aslanguage tagging properties canexpress content in the vastmajority of human languagesFinally by using XMLWebDAV properties can supportother metadata activities whichare also based on XML such asthe Resource DescriptionFramework (RDF) underdevelopment at the W3C

bull Name space managementCreation removal andautomatic consistencymaintenance of collectionscontaining sets of softwaredevelopment artifacts Alsothe ability to copy and moveWeb-accessible artifacts andto receive a listing ofartifacts in a collection(similar to a directory listingin a file system)

Work-in-ProgressCurrent work-in-progress withinWebDAV focuses on theseadditional capabilities

bull Version management Theability to store importantrevisions of a softwareartifacts for later retrievalVersion management canalso support collaboration byallowing two or more authorsto work on the same artifactin parallel tracks Automaticversioning recordssuccessive modifications toan artifact made byversioning unaware(ldquodownlevelrdquo) clientsConfiguration managementtracks versioned collectionsof versioned artifacts animportant capability forreverting to previouslyreleased software packagesor for tracking manyresources simultaneously

A new working group buildingon initial design work performedby the WebDAV working groupis currently being formed in theIETF to develop a protocol forWeb versioning andconfiguration managementCalled the Delta-V workinggroup it is already wellunderway towards its goal of afinished protocol specification inmid-2000 A snapshot of work inprogress can be found byexamining the grouprsquos goalsdocument [10] and the currentversioning and configurationmanagement protocolspecification [3]

bull Advanced CollectionsReferential resources act likesymbolic links in filesystems while orderedcollections maintain a client-specified ordering ofresources in a collectionwithout private agreements

bull Access Control The ability tolimit the access rights of agiven authenticated principalon a given artifact remotelyvia the Internet WebDAVassumes the existence of butdoes not specify strongauthentication technologyand today WebDAV servershave server-specific accesscontrols

A strongly related effort toWebDAV is the IETFrsquos DAVSearching and Locating (DASL)working group which isdeveloping an interoperablemeans of searching a repositorywhich is compliant with theWebDAV object model andwhich organizes its artifacts intoURL hierarchies The maincapability of DASL is searching

bull Searching Client specifiedserver-executed queries tolocate artifacts based upontheir property values and textcontent This is an importantcapability for softwaredevelopment supportingqueries like ldquowhere is thisfunction usedrdquo or ldquowhere isthis variable definedrdquo

DASL is working to developextensions to the WebDAVDistributed Authoring Protocol

Continued from page 6

Continued on page 8

STN8

Introducing WebDAVContinued from page 7

specification (and hence toHTTP) for searching WebDAVrepositories DASL has arequirements document [9] andprotocol document [8] which arestill the subject of intense effortwithin the DASL group

Application

Taken together the WebDAVextensions to HTTP provide thestandard needed to make the Weba writable collaborative mediumWhat does this mean for Web-based software developmentAlthough the future isnotoriously hard to predict hereare some likely outcomes of theadoption of WebDAV AsWebDAV technology isdeployed it will initially have itslargest impact on small tomedium sized developmentteams which support WebDAVallowing their developmentpractices to coalesce around alocal intranet gaining itsadvantages for remote access ofsoftware developmentinformation within theorganization Over time ascritical mass grows WebDAVwill also dramatically reduce theaccidental costs of collaborationbetween development teams andbetween developmentorganizations Since WebDAVsupports multiple informationrepositories it acts as a commonbridge across these storesproviding a convenientmechanism for integrating data ina software development

environment Due to thisWebDAV shows significantpromise as an infrastructure fordevelopment of distributedsoftware engineeringenvironments a topic exploredin a recent Communications ofthe ACM article [5]

While WebDAV todaysignificantly reduces the burdenof cross-team collaborativesoftware development theprotocol will really start to shineonce the Web versioningprotocol is finished Thenremote software engineeringteams will be able to remotelyedit software artifacts whilekeeping track of important statesof these documents and whiletracking configurations of theseversioned objects Since majorconfiguration managementvendors such as RationalMerant (Intersolv) Microsoftand IBM are working together todefine this standard broad toolsupport can be expected

One of the nice aspects ofWebDAV technology is thatcompelling WebDAV clients andservers are available today freeof charge ready for initialevaluation The Apachemod_dav module can bedownloaded atwwwwebdavorgmod_dav andworks with Apache servers 134and higher The Apache serversoftware is freely available fordownload at wwwapacheorgOr use one of the WebDAV test

servers listed atwwwwebdavorgprojects Forclient software try InternetExplorer 5rsquos Web Folders feature(IE5 can be downloaded fromwwwmicrosoftcomwindowsieie5defaultasp) or download theWebDAV Explorer a Java-basedclient developed at the Universityof California Irvine fromwwwicsuciedu~webdav

For More Information

Working groups of the InternetEngineering Task Force arecompletely open and may bejoined by subscribing to their E-mail discussion list If you wishto participate in the discussionson WebDAV topics you may jointhe mailing list by E-mailing amessage with the subjectldquosubscriberdquo to w3c-dist-auth-requestw3org

The home page for the WebDAVgroup is at URLwwwicsuciedupubietfwebdavwhich contains links to current working drafts E-mail list archives and backgroundmaterial Another excellentsource of WebDAV informationis the WebDAV Resources pageat wwwwebdavorg maintainedby Greg Stein The related DAVSearching and Locating (DASL)working group has its web pageat URL wwwicsuciedupubietfdasl and a mailing listwhich may be joined by sendinga message with subjectldquosubscriberdquo to www-webdav-dasl-requestw3org

Continued on page 9

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 5: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN5

References[1] Basili V M Zelkowitz F McGarry J Page S Waligora and R Pajerski SELrsquos Software Process-ImprovementProgram IEEE Software 12 6 (1995) 83-87

[2] Tesoriero R and M V Zelkowitz ldquoA Model of Noisy Software Engineering Datardquo (Status Report) ACMIEEEInternational Conference on Software Engineering Kyoto Japan (April 1998) 461-464

Continued from page 4

Figure 5 Corrective Action Taken on Project that is ldquoOut of Scoperdquo

Collaborative Software Engineering on the Web Introducing WebDAVE James Whitehead Jr - University of California Irvine

Introduction

Software development today oftentakes place among multiple groupsof software engineers who aregeographically dispersed A keychallenge in supporting thesedistributed software developmentteams is making the softwareremotely accessible and editablevia the Internet addressing issuessuch as overwrite preventionsecurity authentication accesscontrol and reliable operationacross high-latency networkconnections

The Web Distributed Authoringand Versioning (WebDAV)

working group of the InternetEngineering Task Force (IETF)has taken on the challenge ofsupporting collaborative softwareengineering on the Web byextending the core networkprotocol of the Web theHypertext Transfer Protocol(HTTP) [4] to support remotesoftware development A team atthe University of CaliforniaIrvine working under a grantfrom the Defense AdvancedResearch Projects Agency(DARPA) EvolutionaryDevelopment of ComplexSystems program broughttogether interested parties from

academia and industry to formthe IETF working group anddevelop the WebDAVDistributed Authoring Protocol

WebDAV provides many benefits

bull Development teams cancollaboratively developsoftware artifacts in-place onthe Web using locking toprevent overwrite conflictsDue to the distributed natureof the Web these workgroups can have membersfrom within the sameorganization or acrossorganizational boundaries

Continued on page 6

STN6

Introducing WebDAVContinued from page 5

copy of Windows 2000 ServerThe popular Apache web serveralso has WebDAV capability inthe freely available mod_davmodule Additional supportcomes from IBM XeroxNovell DataChannelCyberTeams and DigitalCreations who have allannounced product plans basedon WebDAV WebDAV is clearlya standard that has strongcorporate and open sourcesupport

CapabilitiesThe WebDAV DistributedAuthoring Protocol defines a setof extensions to the HypertextTransfer Protocol (HTTP) for thefollowing capabilities

bull Overwrite preventionKeeping more than oneperson from working on adocument at the same timeThis prevents the ldquolostupdate problemrdquo in whichmodifications are lost as firstone developer then anotherwrites changes withoutmerging the otherdeveloperrsquos work

WebDAV provides facilities forboth shared and exclusivelocking This dual lock supportprovides sufficiently flexiblelocks to accommodate a widerange of collaborations withshared locks best supportingcollaborators who have a lot ofawareness of each otherrsquosactivities and exclusive locksproviding a more stringentguarantee of conflict avoidance

for less aware collaborators orduring periods of highcontention for a softwaredevelopment artifact Locks mayhave a scope of a single artifactor a hierarchy of artifacts suchas a source code tree A lockdiscovery mechanism (aWebDAV property) allowsauthors to find out if any locksexist on a Web-accessibleartifact Since the Web isdesigned so that no lock isrequired to read a Web pagethere is no concept of a readlock

bull Properties Creationremoval and querying ofinformation about Web-accessible artifacts such asits author last modified dateetc Also included is theability to make hypertextlinks between artifacts of anycontent type

WebDAV properties are namevalue pairs where the name is aUniform Resource Locator(URL) and the value is asequence of well-formedExtensible Markup Language(XML) [2] elements UsingURLs as property namesprovides a globally uniqueproperty namespace Sinceproperty names can be URLswhich have a domain name as acomponent of the URL propertynames can be given uniquenesswithout central registration byusing URL property nameschosen from within a domainwhose name is controlled by theparty

Continued on page 7

bull All the types of artifacts in atypical software developmentlifecycle can be edited usingWebDAV includingrequirements designdocuments test cases codeand more So whileWebDAV provides HTML [7]and XML [2] authoringsupport it just as easilysupports authoring of existingword processing spreadsheettext graphics and all otherformats

bull WebDAV and HTTP providea common interface to a widerange of repositories such asconfiguration managementfile systems databasesdocument management etcIn essence WebDAV makesthe Web look like a large-grain network-accessible filesystem But unlike aconventional file system aWebDAV-enabled repositoryprovides Internet access andallows all ldquofilesrdquo to beviewed using a standard Webbrowser

In November of 1998 the IETFapproved the WebDAVDistributed Authoring Protocol[6] clearing the way for broadindustry and open source supportMicrosoft is a major earlysupporter of the protocolproviding WebDAV support inInternet Explorer 5 (in its ldquoWebFoldersrdquo feature) in Office 2000(Word and Excel) and on theserver side in its InternetInformation Services (IIS) serverwhich comes bundled with every

STN7

defining the property So forexample a company whichcontrols a given domain namelike ldquowidgetscomrdquo can chose aproperty name from within thisdomain like ldquowidgetscompropertiescolorrdquo

Using XML to encode the valueof properties provides threemajor benefits First isextensibility Since all contentwithin XML is encoded betweenstart and end tags it is easy toadd additional elements to aproperty by inserting new taggedcontent Internationalization isthe second major benefit SinceXML mandates support for theUTF-8 and UTF-16 encodings ofthe ISO 10646 characterencoding standard as well aslanguage tagging properties canexpress content in the vastmajority of human languagesFinally by using XMLWebDAV properties can supportother metadata activities whichare also based on XML such asthe Resource DescriptionFramework (RDF) underdevelopment at the W3C

bull Name space managementCreation removal andautomatic consistencymaintenance of collectionscontaining sets of softwaredevelopment artifacts Alsothe ability to copy and moveWeb-accessible artifacts andto receive a listing ofartifacts in a collection(similar to a directory listingin a file system)

Work-in-ProgressCurrent work-in-progress withinWebDAV focuses on theseadditional capabilities

bull Version management Theability to store importantrevisions of a softwareartifacts for later retrievalVersion management canalso support collaboration byallowing two or more authorsto work on the same artifactin parallel tracks Automaticversioning recordssuccessive modifications toan artifact made byversioning unaware(ldquodownlevelrdquo) clientsConfiguration managementtracks versioned collectionsof versioned artifacts animportant capability forreverting to previouslyreleased software packagesor for tracking manyresources simultaneously

A new working group buildingon initial design work performedby the WebDAV working groupis currently being formed in theIETF to develop a protocol forWeb versioning andconfiguration managementCalled the Delta-V workinggroup it is already wellunderway towards its goal of afinished protocol specification inmid-2000 A snapshot of work inprogress can be found byexamining the grouprsquos goalsdocument [10] and the currentversioning and configurationmanagement protocolspecification [3]

bull Advanced CollectionsReferential resources act likesymbolic links in filesystems while orderedcollections maintain a client-specified ordering ofresources in a collectionwithout private agreements

bull Access Control The ability tolimit the access rights of agiven authenticated principalon a given artifact remotelyvia the Internet WebDAVassumes the existence of butdoes not specify strongauthentication technologyand today WebDAV servershave server-specific accesscontrols

A strongly related effort toWebDAV is the IETFrsquos DAVSearching and Locating (DASL)working group which isdeveloping an interoperablemeans of searching a repositorywhich is compliant with theWebDAV object model andwhich organizes its artifacts intoURL hierarchies The maincapability of DASL is searching

bull Searching Client specifiedserver-executed queries tolocate artifacts based upontheir property values and textcontent This is an importantcapability for softwaredevelopment supportingqueries like ldquowhere is thisfunction usedrdquo or ldquowhere isthis variable definedrdquo

DASL is working to developextensions to the WebDAVDistributed Authoring Protocol

Continued from page 6

Continued on page 8

STN8

Introducing WebDAVContinued from page 7

specification (and hence toHTTP) for searching WebDAVrepositories DASL has arequirements document [9] andprotocol document [8] which arestill the subject of intense effortwithin the DASL group

Application

Taken together the WebDAVextensions to HTTP provide thestandard needed to make the Weba writable collaborative mediumWhat does this mean for Web-based software developmentAlthough the future isnotoriously hard to predict hereare some likely outcomes of theadoption of WebDAV AsWebDAV technology isdeployed it will initially have itslargest impact on small tomedium sized developmentteams which support WebDAVallowing their developmentpractices to coalesce around alocal intranet gaining itsadvantages for remote access ofsoftware developmentinformation within theorganization Over time ascritical mass grows WebDAVwill also dramatically reduce theaccidental costs of collaborationbetween development teams andbetween developmentorganizations Since WebDAVsupports multiple informationrepositories it acts as a commonbridge across these storesproviding a convenientmechanism for integrating data ina software development

environment Due to thisWebDAV shows significantpromise as an infrastructure fordevelopment of distributedsoftware engineeringenvironments a topic exploredin a recent Communications ofthe ACM article [5]

While WebDAV todaysignificantly reduces the burdenof cross-team collaborativesoftware development theprotocol will really start to shineonce the Web versioningprotocol is finished Thenremote software engineeringteams will be able to remotelyedit software artifacts whilekeeping track of important statesof these documents and whiletracking configurations of theseversioned objects Since majorconfiguration managementvendors such as RationalMerant (Intersolv) Microsoftand IBM are working together todefine this standard broad toolsupport can be expected

One of the nice aspects ofWebDAV technology is thatcompelling WebDAV clients andservers are available today freeof charge ready for initialevaluation The Apachemod_dav module can bedownloaded atwwwwebdavorgmod_dav andworks with Apache servers 134and higher The Apache serversoftware is freely available fordownload at wwwapacheorgOr use one of the WebDAV test

servers listed atwwwwebdavorgprojects Forclient software try InternetExplorer 5rsquos Web Folders feature(IE5 can be downloaded fromwwwmicrosoftcomwindowsieie5defaultasp) or download theWebDAV Explorer a Java-basedclient developed at the Universityof California Irvine fromwwwicsuciedu~webdav

For More Information

Working groups of the InternetEngineering Task Force arecompletely open and may bejoined by subscribing to their E-mail discussion list If you wishto participate in the discussionson WebDAV topics you may jointhe mailing list by E-mailing amessage with the subjectldquosubscriberdquo to w3c-dist-auth-requestw3org

The home page for the WebDAVgroup is at URLwwwicsuciedupubietfwebdavwhich contains links to current working drafts E-mail list archives and backgroundmaterial Another excellentsource of WebDAV informationis the WebDAV Resources pageat wwwwebdavorg maintainedby Greg Stein The related DAVSearching and Locating (DASL)working group has its web pageat URL wwwicsuciedupubietfdasl and a mailing listwhich may be joined by sendinga message with subjectldquosubscriberdquo to www-webdav-dasl-requestw3org

Continued on page 9

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 6: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN6

Introducing WebDAVContinued from page 5

copy of Windows 2000 ServerThe popular Apache web serveralso has WebDAV capability inthe freely available mod_davmodule Additional supportcomes from IBM XeroxNovell DataChannelCyberTeams and DigitalCreations who have allannounced product plans basedon WebDAV WebDAV is clearlya standard that has strongcorporate and open sourcesupport

CapabilitiesThe WebDAV DistributedAuthoring Protocol defines a setof extensions to the HypertextTransfer Protocol (HTTP) for thefollowing capabilities

bull Overwrite preventionKeeping more than oneperson from working on adocument at the same timeThis prevents the ldquolostupdate problemrdquo in whichmodifications are lost as firstone developer then anotherwrites changes withoutmerging the otherdeveloperrsquos work

WebDAV provides facilities forboth shared and exclusivelocking This dual lock supportprovides sufficiently flexiblelocks to accommodate a widerange of collaborations withshared locks best supportingcollaborators who have a lot ofawareness of each otherrsquosactivities and exclusive locksproviding a more stringentguarantee of conflict avoidance

for less aware collaborators orduring periods of highcontention for a softwaredevelopment artifact Locks mayhave a scope of a single artifactor a hierarchy of artifacts suchas a source code tree A lockdiscovery mechanism (aWebDAV property) allowsauthors to find out if any locksexist on a Web-accessibleartifact Since the Web isdesigned so that no lock isrequired to read a Web pagethere is no concept of a readlock

bull Properties Creationremoval and querying ofinformation about Web-accessible artifacts such asits author last modified dateetc Also included is theability to make hypertextlinks between artifacts of anycontent type

WebDAV properties are namevalue pairs where the name is aUniform Resource Locator(URL) and the value is asequence of well-formedExtensible Markup Language(XML) [2] elements UsingURLs as property namesprovides a globally uniqueproperty namespace Sinceproperty names can be URLswhich have a domain name as acomponent of the URL propertynames can be given uniquenesswithout central registration byusing URL property nameschosen from within a domainwhose name is controlled by theparty

Continued on page 7

bull All the types of artifacts in atypical software developmentlifecycle can be edited usingWebDAV includingrequirements designdocuments test cases codeand more So whileWebDAV provides HTML [7]and XML [2] authoringsupport it just as easilysupports authoring of existingword processing spreadsheettext graphics and all otherformats

bull WebDAV and HTTP providea common interface to a widerange of repositories such asconfiguration managementfile systems databasesdocument management etcIn essence WebDAV makesthe Web look like a large-grain network-accessible filesystem But unlike aconventional file system aWebDAV-enabled repositoryprovides Internet access andallows all ldquofilesrdquo to beviewed using a standard Webbrowser

In November of 1998 the IETFapproved the WebDAVDistributed Authoring Protocol[6] clearing the way for broadindustry and open source supportMicrosoft is a major earlysupporter of the protocolproviding WebDAV support inInternet Explorer 5 (in its ldquoWebFoldersrdquo feature) in Office 2000(Word and Excel) and on theserver side in its InternetInformation Services (IIS) serverwhich comes bundled with every

STN7

defining the property So forexample a company whichcontrols a given domain namelike ldquowidgetscomrdquo can chose aproperty name from within thisdomain like ldquowidgetscompropertiescolorrdquo

Using XML to encode the valueof properties provides threemajor benefits First isextensibility Since all contentwithin XML is encoded betweenstart and end tags it is easy toadd additional elements to aproperty by inserting new taggedcontent Internationalization isthe second major benefit SinceXML mandates support for theUTF-8 and UTF-16 encodings ofthe ISO 10646 characterencoding standard as well aslanguage tagging properties canexpress content in the vastmajority of human languagesFinally by using XMLWebDAV properties can supportother metadata activities whichare also based on XML such asthe Resource DescriptionFramework (RDF) underdevelopment at the W3C

bull Name space managementCreation removal andautomatic consistencymaintenance of collectionscontaining sets of softwaredevelopment artifacts Alsothe ability to copy and moveWeb-accessible artifacts andto receive a listing ofartifacts in a collection(similar to a directory listingin a file system)

Work-in-ProgressCurrent work-in-progress withinWebDAV focuses on theseadditional capabilities

bull Version management Theability to store importantrevisions of a softwareartifacts for later retrievalVersion management canalso support collaboration byallowing two or more authorsto work on the same artifactin parallel tracks Automaticversioning recordssuccessive modifications toan artifact made byversioning unaware(ldquodownlevelrdquo) clientsConfiguration managementtracks versioned collectionsof versioned artifacts animportant capability forreverting to previouslyreleased software packagesor for tracking manyresources simultaneously

A new working group buildingon initial design work performedby the WebDAV working groupis currently being formed in theIETF to develop a protocol forWeb versioning andconfiguration managementCalled the Delta-V workinggroup it is already wellunderway towards its goal of afinished protocol specification inmid-2000 A snapshot of work inprogress can be found byexamining the grouprsquos goalsdocument [10] and the currentversioning and configurationmanagement protocolspecification [3]

bull Advanced CollectionsReferential resources act likesymbolic links in filesystems while orderedcollections maintain a client-specified ordering ofresources in a collectionwithout private agreements

bull Access Control The ability tolimit the access rights of agiven authenticated principalon a given artifact remotelyvia the Internet WebDAVassumes the existence of butdoes not specify strongauthentication technologyand today WebDAV servershave server-specific accesscontrols

A strongly related effort toWebDAV is the IETFrsquos DAVSearching and Locating (DASL)working group which isdeveloping an interoperablemeans of searching a repositorywhich is compliant with theWebDAV object model andwhich organizes its artifacts intoURL hierarchies The maincapability of DASL is searching

bull Searching Client specifiedserver-executed queries tolocate artifacts based upontheir property values and textcontent This is an importantcapability for softwaredevelopment supportingqueries like ldquowhere is thisfunction usedrdquo or ldquowhere isthis variable definedrdquo

DASL is working to developextensions to the WebDAVDistributed Authoring Protocol

Continued from page 6

Continued on page 8

STN8

Introducing WebDAVContinued from page 7

specification (and hence toHTTP) for searching WebDAVrepositories DASL has arequirements document [9] andprotocol document [8] which arestill the subject of intense effortwithin the DASL group

Application

Taken together the WebDAVextensions to HTTP provide thestandard needed to make the Weba writable collaborative mediumWhat does this mean for Web-based software developmentAlthough the future isnotoriously hard to predict hereare some likely outcomes of theadoption of WebDAV AsWebDAV technology isdeployed it will initially have itslargest impact on small tomedium sized developmentteams which support WebDAVallowing their developmentpractices to coalesce around alocal intranet gaining itsadvantages for remote access ofsoftware developmentinformation within theorganization Over time ascritical mass grows WebDAVwill also dramatically reduce theaccidental costs of collaborationbetween development teams andbetween developmentorganizations Since WebDAVsupports multiple informationrepositories it acts as a commonbridge across these storesproviding a convenientmechanism for integrating data ina software development

environment Due to thisWebDAV shows significantpromise as an infrastructure fordevelopment of distributedsoftware engineeringenvironments a topic exploredin a recent Communications ofthe ACM article [5]

While WebDAV todaysignificantly reduces the burdenof cross-team collaborativesoftware development theprotocol will really start to shineonce the Web versioningprotocol is finished Thenremote software engineeringteams will be able to remotelyedit software artifacts whilekeeping track of important statesof these documents and whiletracking configurations of theseversioned objects Since majorconfiguration managementvendors such as RationalMerant (Intersolv) Microsoftand IBM are working together todefine this standard broad toolsupport can be expected

One of the nice aspects ofWebDAV technology is thatcompelling WebDAV clients andservers are available today freeof charge ready for initialevaluation The Apachemod_dav module can bedownloaded atwwwwebdavorgmod_dav andworks with Apache servers 134and higher The Apache serversoftware is freely available fordownload at wwwapacheorgOr use one of the WebDAV test

servers listed atwwwwebdavorgprojects Forclient software try InternetExplorer 5rsquos Web Folders feature(IE5 can be downloaded fromwwwmicrosoftcomwindowsieie5defaultasp) or download theWebDAV Explorer a Java-basedclient developed at the Universityof California Irvine fromwwwicsuciedu~webdav

For More Information

Working groups of the InternetEngineering Task Force arecompletely open and may bejoined by subscribing to their E-mail discussion list If you wishto participate in the discussionson WebDAV topics you may jointhe mailing list by E-mailing amessage with the subjectldquosubscriberdquo to w3c-dist-auth-requestw3org

The home page for the WebDAVgroup is at URLwwwicsuciedupubietfwebdavwhich contains links to current working drafts E-mail list archives and backgroundmaterial Another excellentsource of WebDAV informationis the WebDAV Resources pageat wwwwebdavorg maintainedby Greg Stein The related DAVSearching and Locating (DASL)working group has its web pageat URL wwwicsuciedupubietfdasl and a mailing listwhich may be joined by sendinga message with subjectldquosubscriberdquo to www-webdav-dasl-requestw3org

Continued on page 9

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 7: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN7

defining the property So forexample a company whichcontrols a given domain namelike ldquowidgetscomrdquo can chose aproperty name from within thisdomain like ldquowidgetscompropertiescolorrdquo

Using XML to encode the valueof properties provides threemajor benefits First isextensibility Since all contentwithin XML is encoded betweenstart and end tags it is easy toadd additional elements to aproperty by inserting new taggedcontent Internationalization isthe second major benefit SinceXML mandates support for theUTF-8 and UTF-16 encodings ofthe ISO 10646 characterencoding standard as well aslanguage tagging properties canexpress content in the vastmajority of human languagesFinally by using XMLWebDAV properties can supportother metadata activities whichare also based on XML such asthe Resource DescriptionFramework (RDF) underdevelopment at the W3C

bull Name space managementCreation removal andautomatic consistencymaintenance of collectionscontaining sets of softwaredevelopment artifacts Alsothe ability to copy and moveWeb-accessible artifacts andto receive a listing ofartifacts in a collection(similar to a directory listingin a file system)

Work-in-ProgressCurrent work-in-progress withinWebDAV focuses on theseadditional capabilities

bull Version management Theability to store importantrevisions of a softwareartifacts for later retrievalVersion management canalso support collaboration byallowing two or more authorsto work on the same artifactin parallel tracks Automaticversioning recordssuccessive modifications toan artifact made byversioning unaware(ldquodownlevelrdquo) clientsConfiguration managementtracks versioned collectionsof versioned artifacts animportant capability forreverting to previouslyreleased software packagesor for tracking manyresources simultaneously

A new working group buildingon initial design work performedby the WebDAV working groupis currently being formed in theIETF to develop a protocol forWeb versioning andconfiguration managementCalled the Delta-V workinggroup it is already wellunderway towards its goal of afinished protocol specification inmid-2000 A snapshot of work inprogress can be found byexamining the grouprsquos goalsdocument [10] and the currentversioning and configurationmanagement protocolspecification [3]

bull Advanced CollectionsReferential resources act likesymbolic links in filesystems while orderedcollections maintain a client-specified ordering ofresources in a collectionwithout private agreements

bull Access Control The ability tolimit the access rights of agiven authenticated principalon a given artifact remotelyvia the Internet WebDAVassumes the existence of butdoes not specify strongauthentication technologyand today WebDAV servershave server-specific accesscontrols

A strongly related effort toWebDAV is the IETFrsquos DAVSearching and Locating (DASL)working group which isdeveloping an interoperablemeans of searching a repositorywhich is compliant with theWebDAV object model andwhich organizes its artifacts intoURL hierarchies The maincapability of DASL is searching

bull Searching Client specifiedserver-executed queries tolocate artifacts based upontheir property values and textcontent This is an importantcapability for softwaredevelopment supportingqueries like ldquowhere is thisfunction usedrdquo or ldquowhere isthis variable definedrdquo

DASL is working to developextensions to the WebDAVDistributed Authoring Protocol

Continued from page 6

Continued on page 8

STN8

Introducing WebDAVContinued from page 7

specification (and hence toHTTP) for searching WebDAVrepositories DASL has arequirements document [9] andprotocol document [8] which arestill the subject of intense effortwithin the DASL group

Application

Taken together the WebDAVextensions to HTTP provide thestandard needed to make the Weba writable collaborative mediumWhat does this mean for Web-based software developmentAlthough the future isnotoriously hard to predict hereare some likely outcomes of theadoption of WebDAV AsWebDAV technology isdeployed it will initially have itslargest impact on small tomedium sized developmentteams which support WebDAVallowing their developmentpractices to coalesce around alocal intranet gaining itsadvantages for remote access ofsoftware developmentinformation within theorganization Over time ascritical mass grows WebDAVwill also dramatically reduce theaccidental costs of collaborationbetween development teams andbetween developmentorganizations Since WebDAVsupports multiple informationrepositories it acts as a commonbridge across these storesproviding a convenientmechanism for integrating data ina software development

environment Due to thisWebDAV shows significantpromise as an infrastructure fordevelopment of distributedsoftware engineeringenvironments a topic exploredin a recent Communications ofthe ACM article [5]

While WebDAV todaysignificantly reduces the burdenof cross-team collaborativesoftware development theprotocol will really start to shineonce the Web versioningprotocol is finished Thenremote software engineeringteams will be able to remotelyedit software artifacts whilekeeping track of important statesof these documents and whiletracking configurations of theseversioned objects Since majorconfiguration managementvendors such as RationalMerant (Intersolv) Microsoftand IBM are working together todefine this standard broad toolsupport can be expected

One of the nice aspects ofWebDAV technology is thatcompelling WebDAV clients andservers are available today freeof charge ready for initialevaluation The Apachemod_dav module can bedownloaded atwwwwebdavorgmod_dav andworks with Apache servers 134and higher The Apache serversoftware is freely available fordownload at wwwapacheorgOr use one of the WebDAV test

servers listed atwwwwebdavorgprojects Forclient software try InternetExplorer 5rsquos Web Folders feature(IE5 can be downloaded fromwwwmicrosoftcomwindowsieie5defaultasp) or download theWebDAV Explorer a Java-basedclient developed at the Universityof California Irvine fromwwwicsuciedu~webdav

For More Information

Working groups of the InternetEngineering Task Force arecompletely open and may bejoined by subscribing to their E-mail discussion list If you wishto participate in the discussionson WebDAV topics you may jointhe mailing list by E-mailing amessage with the subjectldquosubscriberdquo to w3c-dist-auth-requestw3org

The home page for the WebDAVgroup is at URLwwwicsuciedupubietfwebdavwhich contains links to current working drafts E-mail list archives and backgroundmaterial Another excellentsource of WebDAV informationis the WebDAV Resources pageat wwwwebdavorg maintainedby Greg Stein The related DAVSearching and Locating (DASL)working group has its web pageat URL wwwicsuciedupubietfdasl and a mailing listwhich may be joined by sendinga message with subjectldquosubscriberdquo to www-webdav-dasl-requestw3org

Continued on page 9

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 8: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN8

Introducing WebDAVContinued from page 7

specification (and hence toHTTP) for searching WebDAVrepositories DASL has arequirements document [9] andprotocol document [8] which arestill the subject of intense effortwithin the DASL group

Application

Taken together the WebDAVextensions to HTTP provide thestandard needed to make the Weba writable collaborative mediumWhat does this mean for Web-based software developmentAlthough the future isnotoriously hard to predict hereare some likely outcomes of theadoption of WebDAV AsWebDAV technology isdeployed it will initially have itslargest impact on small tomedium sized developmentteams which support WebDAVallowing their developmentpractices to coalesce around alocal intranet gaining itsadvantages for remote access ofsoftware developmentinformation within theorganization Over time ascritical mass grows WebDAVwill also dramatically reduce theaccidental costs of collaborationbetween development teams andbetween developmentorganizations Since WebDAVsupports multiple informationrepositories it acts as a commonbridge across these storesproviding a convenientmechanism for integrating data ina software development

environment Due to thisWebDAV shows significantpromise as an infrastructure fordevelopment of distributedsoftware engineeringenvironments a topic exploredin a recent Communications ofthe ACM article [5]

While WebDAV todaysignificantly reduces the burdenof cross-team collaborativesoftware development theprotocol will really start to shineonce the Web versioningprotocol is finished Thenremote software engineeringteams will be able to remotelyedit software artifacts whilekeeping track of important statesof these documents and whiletracking configurations of theseversioned objects Since majorconfiguration managementvendors such as RationalMerant (Intersolv) Microsoftand IBM are working together todefine this standard broad toolsupport can be expected

One of the nice aspects ofWebDAV technology is thatcompelling WebDAV clients andservers are available today freeof charge ready for initialevaluation The Apachemod_dav module can bedownloaded atwwwwebdavorgmod_dav andworks with Apache servers 134and higher The Apache serversoftware is freely available fordownload at wwwapacheorgOr use one of the WebDAV test

servers listed atwwwwebdavorgprojects Forclient software try InternetExplorer 5rsquos Web Folders feature(IE5 can be downloaded fromwwwmicrosoftcomwindowsieie5defaultasp) or download theWebDAV Explorer a Java-basedclient developed at the Universityof California Irvine fromwwwicsuciedu~webdav

For More Information

Working groups of the InternetEngineering Task Force arecompletely open and may bejoined by subscribing to their E-mail discussion list If you wishto participate in the discussionson WebDAV topics you may jointhe mailing list by E-mailing amessage with the subjectldquosubscriberdquo to w3c-dist-auth-requestw3org

The home page for the WebDAVgroup is at URLwwwicsuciedupubietfwebdavwhich contains links to current working drafts E-mail list archives and backgroundmaterial Another excellentsource of WebDAV informationis the WebDAV Resources pageat wwwwebdavorg maintainedby Greg Stein The related DAVSearching and Locating (DASL)working group has its web pageat URL wwwicsuciedupubietfdasl and a mailing listwhich may be joined by sendinga message with subjectldquosubscriberdquo to www-webdav-dasl-requestw3org

Continued on page 9

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 9: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN9

Continued from page 8

About the Author

Emmet James Whitehead iscurrently a PhD student at theUniversity of California IrvineHe has been the Chair of theInternet Engineering Task ForceWorld Wide Web DistributedAuthoring and Versioning(WebDAV) Working GroupMarch 1997-present Withassistance from the World Wide

References[1] T Berners-Lee R Fielding L Masinter ldquoUniform Resource Identifiers (URI) Generic Syntaxrdquo Draft StandardRequest for Comments (RFC) 2396 MITLCS UC Irvine Xerox August 1998 ftpftpisieduin-notesrfc2396txt

[2] T Bray J Paoli C M Sperberg-McQueen ldquoExtensible Markup Language (XML) 10rdquo W3CRecommendation REC-xml February 1998 wwww3orgTRREC-xml

[3] C Kaler J Amsden G Clemm ldquoVersioning Extensions to WebDAVrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-versioning-01 January 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-versioning-01txt

[4] R Fielding J Gettys J Mogul H Frystyk T Berners-Lee ldquoHypertext Transfer Protocol mdash HTTP11rdquoInternet Proposed Standard Request for Comments (RFC) 2068 UC Irvine DEC MITLCS January 1997wwwicsuciedupubietfhttprfc2068txt

[5] R Fielding E J Whitehead Jr K M Anderson G A Bolcer P Oreizy R N Taylor ldquoWeb-Based Developmentof Complex Information Productsrdquo Communications of the ACM Vol 41 No 8 August 1998 pages 84-92

[6] Y Y Goland E J Whitehead Jr A Faizi S R Carter D Jensen ldquoExtensions for Distributed Authoring onthe World Wide Web mdash WEBDAVrdquo Proposed Standard Request for Comments (RFC) 2518 February 1999ftpftpisieduin-notesrfc2518txt

[7] D Raggett A Le Hors I Jacobs ldquoHTML 40 Specificationrdquo W3C Recommendation REC-html40 April 24 1998wwww3orgpubWWWTRREC-html40html

[8] S Reddy D Jensen S Reddy R Henderson J Davis A Babich ldquoDAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-reddy-dasl-protocol-04 November 1998ftpftpisieduinternet-draftsdraft-reddy-dasl-protocol-04txt

[9] J Davis S Reddy J Slein ldquoRequirements for DAV Searching and Locatingrdquo Internet-Draft Work-in-Progress draft-ietf-dasl-requirements-00 February 1999ftpftpisieduinternet-draftsdraft-ietf-dasl-requirements-00txt

[10] J Stracke J Amsden ldquoGoals for Web Versioningrdquo Internet-Draft Work-in-Progress draft-ietf-webdav-version-goals-00 February 1999 ftpftpisieduinternet-draftsdraft-ietf-webdav-version-goals-00txt

[11] J Slein F Vitali E J Whitehead Jr D Durand ldquoRequirements for Distributed Authoring and Versioningon the World Wide Webrdquo Xerox University of Bologna UC Irvine Boston University Internet InformationalRequest for Comments (RFC) 2291 February 1998 wwwicsuciedupubietfwebdavrequirementsrfc2291txt

Web Consortium WebDAVcreated this working group to bethe focus of coordination for thedevelopment of interoperabilitywork on tools which supportdistributed authoring andversioning via the WWW

Mr Whiteheadrsquos current researchfocuses on hypertext andsoftware architecture

Author contact Information

E James Whitehead JrDept of Information and

Computer ScienceUniversity of California Irvine

Irvine CA 92697-3425

ejwicsuciedu wwwicsuciedu~ejw

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 10: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN10

Introduction

The Information Age of thiscentury has allowedcommunication technology toboldly go well beyond those ofthe past Cellular phonesbeepers personal workstationscomputer networks and theInternet have all contributed tothe Information Age Some oftodayrsquos technologies are aimed atproviding collaboration supportfor the workday activities ofdistributed groups Theseactivities may includecommunication cooperationproblem solving andornegotiation and may be directedtowards almost any applicationincluding telemedicineCommand and Control andsoftware developmentevolutionWhile existing collaborationenvironments have come a longway in providing capabilities tosupport group work issues suchas usability flexibility andinteroperability remain TheOrbit work aims to overcomesuch obstacles through thedevelopment of a next generationcollaboration environmentThree key elements contribute toOrbitrsquos approach for a nextgeneration collaborationenvironment

1) a theoretical frameworkbased on groupwork incomplex domains

2) a multi-functional openarchitectural framework forsupport of collaboration

Orbit A Next Generation Collaboration EnvironmentJames Milligan amp Carla Burns Air Force Research Laboratory Information Technology Division

thought of as a focal point aroundwhich to define structure andrelate the relevant peopleobjects tools and resourcesgermane to a particularcollaboration activity Thelocales foundations principlecaptures the basic domainstructuring and furnishing ofwork Locale foundations aretherefore about a) providingadequate media and mechanismsin available domains to supportthe sharing of objects tools andresources b) supporting agrouprsquos notion of membershipand related processes and c)facilitating appropriate privacyand access mechanisms

The civic structure principledeals with facilitating interactionwith the wider communitybeyond an individualrsquosimmediate workgroups andlocales It includes the lifecycleprocesses that support theemergence and dissolution oflocales and the structuring of theworld of locales in the broadersense This is where externalinfluences beyond the locales ofdirect interest can be considered(eg organizational professionalfinancial and political)

The trajectory principle isconcerned with all of thetemporal aspects of the grouprsquoslocale its associated individualsand entities It also deals withthe phasing articulation andmanagement of interactions Themutuality principle concerns the

activities and the theoreticalframework and

3) powerful data visualizationfacilities which improve userunderstanding

The TheoreticalFramework of OrbitLocales

Orbitrsquos theoretical framework isbased upon the work ofsociologist Anselm Strauss andhis notion of social worlds [1 2]Straussrsquo social world modeladdresses the understandinganalysis and reasoning aboutgroup work in complex domainsThe Orbit work takes many ofthe social concepts in Straussrsquomodel and applies them tosupport collaboration activitiesthrough networks of computersusing locales Locales arevirtual places for group worksituated in the computernetwork Straussrsquo work wasselected as the theoretical basisfor Orbit due its support of thefollowing group work conceptsdisplaced action as the centralfocus of work admittedflexibility and contingency ofwork and equal weight of formaland informal aspects of work

The locales frameworkintegrates five principlesnecessary for distributed groupcollaboration These principlesare locale foundationsmutuality individual viewsinteraction trajectories and civicstructures A locale can be

Continued on page 11

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 11: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN11

degree to which awareness andpresence must be supported inthe collaborative work Themutuality principle is importantfor both synchronous andasynchronous interactionsalthough the medium ofexpression might be verydifferent

Finally the individual viewprinciple looks at the differentindividual views of the samelocale and the individual viewsover multiple locales While theremay be a group definition of thelocale the individuals in the groupmay all have a different view ofor interest in the locale based ontheir current level of involvementMoreover few individuals havethe luxury of being able to focuson a single task exclusively Theyusually belong to multiple socialworlds and work on manydifferent tasks concurrently withvarying (and shifting) degrees ofintensity

The Orbit EnvironmentMulti-functional andOpen Architecture

The Orbit environment providescollaboration tools that supportthe locales theoreticalframework which is guided bythe five principles discussedabove It allows spontaneousgeneration and evolution ofnetworks of locales Each localeis dedicated to a particularpurpose and furnished with theartifacts and tools required fordistributed workgroups toeffectively accomplish the task

at hand When groups are work-ing they need the following [3]

bull The family of artifacts thatmake up the formal layer oftheir work activitiesExamples include programfiles medical records yellowstickies etc

bull The tools that are used tomanipulate these artifactsExamples include compilerseditors pens etc

bull Resources for effectivecommunication that grantmembers of the social worldthe ability to communicateappropriately to the task athand

bull Automation of mundane taskssuch as change notification

bull The ability to navigate Thatis to seamlessly switchamong multiple on-going

tasks interrelate them asappropriate and find tasksand people as needed

The environment emphasizesflexible coordination andcommunication and can beeasily integrated with workgrouprepositories artifacts and tools aswell as other workgroupmanagement systems Keyfeatures of the environmentinclude a ubiquitouscollaborative desktop wide-areascalable collaborationinfrastructure synchronous andasynchronous group supportuser-controlled projection ofpresence and awareness andpervasive audiovideocapabilities Figure 1 providesan illustration of the Orbitenvironment The Orbit userinterface shown in the figure

Figure 1 The Orbit Collaboration Environment

Continued from page 10

Continued on page 12

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 12: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN12

consists of a locale navigator anda locale workspace Thenavigator permits configurationof locales objects documentsand shared audiovideoconferencing The workspaceprovides capabilities for viewingand manipulating sharedrepository artifacts

Orbit is based on a three-layerconceptual model thatimplements locales andindividual views based on theexternal objects in the lowestlayer The locales layer groupsobjects and tools together andprovides support for presenceawareness and trajectoryinformation The individualviews layer provides the userwith vision into multiple localessimultaneously and with varyingdegrees of intensity The systemprovides a ubiquitouscollaborative desktop throughwhich users can perform allshared and individual tasksFigure 2 presents an illustrationof Orbitrsquos conceptual architecture

The Data VisualizationCapabilities of OrbitVirtue

In the physical world individualsreact directly with their dailyenvironments The Orbitenvironment provides powerfuldata visualization capabilitiesthat immerse the individual intohisher domain This componentis called Virtue To date Virtue

has been applied to the systemengineering arena by providing adirect manipulation interface forexploring software structureevolution and behaviordynamics Key capabilities ofthe Virtue component includeshared 3D views of softwarestructure and history virtualtools for exploring datavisualizations multimediaannotations for spatio-temporalmarking tactile feedback forgrasping and manipulation andvoice command forunencumbered interaction Theactions of each Virtue user arereflected in all other coupledvirtual spaces Future work willexplore the application of Virtueto other application domainssuch as telemedicine andCommand and Control

Acknowledgments

The Orbit and Virtue work iscurrently being sponsored by theDARPA Evolutionary Design ofComplex Systems (EDCS)Program and the Air ForceResearch Laboratory (AFRL)Information TechnologyDivision Additional sources ofsponsorship for the work havebeen provided by the DARPAIntelligent Collaboration ampVisualization and SoftwareEnvironments Programs the USArmy Construction EngineeringLaboratory the National ScienceFoundation Sun MicrosystemsHewlett-Packard LaboratoriesDigital Equipment CorporationIntel and Fujitsu Open SystemsSolutions

OrbitContinued from page 11

Figure 2 Orbit Conceptual Architecture

Continued on page 13

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 13: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN13

Continued from page 12

Additional Information

Further information on Orbit can be found at wwwdstceduauworlds

Further information on Virtue can be found atwww-pablocsuiuceduProjectVRVirtueOverviewhtml

About the Authors

James Milligan has been a computer scientist with Air ForceResearch Laboratory (AFRL) for 14 years where he has directedresearch and development activities in software engineering He iscurrently the AFRL Program Manager for the Evolutionary Design ofComplex Software (EDCS) Program He received his BS inComputer Science from the State University of New York (SUNY)School of Technology and is currently finalizing his MS in SoftwareDevelopment and Management through Rochester Institute ofTechnology

Carla Burns has been a computer scientist with Air Force ResearchLaboratory (AFRL) for 13 years where she has directed research anddevelopment activities in software engineering in the areas ofrequirements engineering rapid prototyping and scenario generationShe received her BS in Mathematics from Clarkson University and aMS in Computer Science from the State University of New York(SUNY) Institute of Technology

Author Contact Information

James Milligan Carla Burns

Air Force Research Laboratory Air Force Research LaboratoryAFRLIFTD AFRLIFTD

525 Brooks Road 525 Brooks RoadRome NY 13441-4505 Rome NY 13441-4505

milliganjrlafmil burnscrlafmil

References[1] Fitzpatrick G WJ Tolone et al (1995) Work Locales and DistributedSocial Worlds

[2] Strauss A (1993) Continual Permutations of Action New YorkAdeline De Gruyter

[3] University of Illinois (1998) ldquoDraft Final Technical Report forAtlantisrdquo Air ForceContract No F30602-94-0161

[4] Orbit Brochure Kaplan Simon Reed Daniel (1998) EvolutionaryDesign of Complex Systems Demonstration Days

Here are the people thatbring you this finepublication the

DoD Software Tech NewsEditorial Board Members

Lon R Dean - EditorDoD Software Tech NewsITT Industries Systems Div

Paul Engelhart -DACS COTR

US Air Force ResearchLaboratory -

Information DirectorateIFTD

Elaine FedchakITT Industries Systems Div

Morton A HirschbergInformation Science ampTechnology Directorate

US Army Research Laboratory

Thomas McGibbon ndashDACS Director

ITT Industries Systems Div

Marshall PotterDDRampE (IT)

Dan Snell -DACS Deputy Director

ITT Industries Systems Div

Nancy L SunderhaftITT Industries Systems Div

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 14: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN14

1 An owner of the project isidentified He identifies theother people who willparticipate in thisnegotiation

2 The owner starts WinWincreates the new project andenters the names of all usersThese people are calledstakeholders

3 One stakeholder isdesignated to define or tailoran existing set of terms forthe proposed system Heenters them in WinWin

4 One stakeholder isdesignated to define or tailoran existing taxonomy for theproposed system He entersthe taxonomy in WinWin

5 Stakeholders review anditerate the terms andtaxonomy

6 One now begins thenegotiation process whichcontinuously loops throughsteps 6 and 7

6a stakeholders create WinConditions expressing theirpreferences andor

6b stakeholders create Issuesthat they believe exist andor

6c stakeholders createproposed Agreements

7 Stakeholders review newlyentered artifacts withexisting artifacts

7a a new conflict is observedso a new Issue is created

7b stakeholders developOptions to address Issues

7c stakeholders create newAgreements and vote onAgreements in-Progress

Steps 6 and 7 continue until allWin Conditions are covered allIssues are resolved and allAgreements are passed WinConditions and Issues that nolonger are relevant Options thatare unused and Agreements thathave failed are marked asINACTIVE Inactive artifactsare not shown by default butthere is a way to restore them

A More ComplicatedExample of Negotiation(Illustrated in Figure 2 on page 15)

Suppose we have two winconditions involved in an issueThe issue has one option whichis adopted by Agreement1Agreement1 is voted on passes

WiWinContinued from page 1

and in turn covers Win Condition1 and 2 Now suppose a new wincondition is entered which causesAgreement1 to become invalidWhat are the actions that shouldresult

1 The owner of Agreement1changes its status toINACTIVE This causesOption1 to be unadoptedIssue1 to be unresolved andWin Conditions 1 and 2 tobe uncovered

2 A new issue is draftedIssue2 which involves WinConditions 1 2 and 3Issue2 replaces Issue1

3 Options to resolve Issue2 aregenerated and Option2 ischosen to create anagreement Agreement2Agreement2 replacesAgreement1

4 Agreement2 initiates a votewhich eventually passesThis causes Option2 to beadopted Issue2 resolvedand Win Conditions 1 2and 3 to be covered

WinC Involves

Issue

AddressesOption

AdoptsAgreement

Figure 1 A Simple Scenario

Continued on page 15

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 15: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN15

Continued on page 16

Continued from page 15

Rationale Graph

One of the essential elements inany negotiation is a record of thearguments that were used infavor or against a particular issueor option WinWin assists in thecapture and retention of all sucharguments through its processmodel described earlier and arationale graph In technicalterms a rationale graph is thetransitive closure of the set ofnodes that are reachable from anAgreement In effect thisincludes all proposed Optionswhether adopted or not allIssues which are eventuallyaddressed and all WinConditions This graph isdisplayed by the program as anindented list By tracing through

the web of interconnections onemay completely resurrect thearguments that were used whichled to the adoption of theAgreement

In the Figure 3 is a picture ofartifact customer-AGRE-1 Thisis an Agreement artifact On theright hand side you see theArtifact Set window Thisagreement points to an optioncustomer-OPTN-1 which in turnpoints to an issue customer-ISSU-1 The agreement alsopoints to a win condition user-WINC-3 At some later point inthe process stakeholders willvote on this Agreement Once avote is started all pointers toartifacts are frozen as theartifacts must maintain the

identical form throughout thevoting process Once a vote iscomplete the Agreement eitherpasses or fails

Another form of rationalesupport in WinWin is theRationale field This field isplaced next to the bodydescription of an artifact andcan be seen in Figure 4 Thestakeholder may explicitlyprovide his rationale for aparticular artifact by entering astatement in that field

Win Win Attachments

WinWin recognizes that theremay be auxiliary tools thatstakeholders desire to useduring the course of anegotiation For example onemight use a spreadsheet to

analyze the financial impacts of agiven Option Or one might use aprogram such as COCOMO toestimate the effort and schedulerequired for a particular decisionWinWin provides a capability toattach such programs and theiroutputs to any artifact in thesystem This is called theAttachment field By making theAttachment field be a part ofevery artifact stakeholders mayassociate these programelaborations at a desirable levelof granularity The Attachmentfield allows for an arbitrarynumber of attachments Eachattachment includes the name ofthe program and its associateddata set

Figure 2 A More Complicated Negotiation

Issue1 Option1 Agreement1

WinC1

WinC2

WinC3

Issue2 Option2 Agreement2

Step A Initially two Win Conditions give rise to an Issuewhich is resolved by an Option and an Agreement is drafted

Step B A new Win Conditions is entered which causes theAgreement to become invalid

Step C Issue2 replaces Issue1 and Agreement2 replaces Agreement1

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 16: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN16

Figure 5 shows an example of theAttachments field of an artifactOn the right hand side of thewindow towards the bottom aretwo input lines labeled Tool andFile Here is placed the name ofthe program to be attached plusany associated data file Whenthe apply button is clicked theprogram name and data file areplaced in the windowimmediately above with all otherattachments

Mode of Operation

As a tool for requirementscapture and negotiation WinWinassumes that stakeholders will bepotentially working at differentlocations and at different timesThus WinWin supports adistributed asynchronous modeof operation A stakeholder maysign onto the system at any timeThere may or may not be otherstakeholders using the systemThe stakeholder may examine theMessages a record of all changesmade to the WinWin database bythe stakeholders These messagesare ordered by date and eachstakeholder has the option tomaintain or discard any or all ofthe Messages

Figure 6 on page 18 shows somesample output of the Messageswindow Each artifact is namedby its unique identifier eg user-TERM-1 is the artifact thatbelongs to the stakeholder nameduser and it is an artifact of typeTERM Each line in theMessages window refers to a

Continued on page 17

WinWinContinued from page 15

Figure 3 Agreement Artifact

Figure 4 Rationale Field

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 17: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN17

Continued from page 16

Continued on page 18

unique action performed by thestakeholder For example thefirst line in the figure indicatesthat a new TERM artifact hasbeen created on 022798 at time1922 Other messages indicatethe date and time that the artifactwas modified including thename of the field that wasmodified At the bottom of thiswindow there are three buttonsThe Delete button will removethe highlighted Message Cancelcauses the Message window todisappear OK causes the artifactin the highlighted line to appear

Win Win Versions

Win Win was developed in C X-Windows and Motif and runs onSolaris HP-UX and Linuxoperating systems It is availablefrom httpsunsetusceduWinWinwinwinhtmldownload

There is a version of Win Winthat has been developed usingJava This version can be runfrom the CSE web site byinvoking the URL httpsunsetuscedujavawinwinwinwinhtml

You may download the Javaclass files and install Java WinWin at your own site To do thisinvoke the URLhttpsunsetuscedujwinshtml

Win Win API

We have developed a library offunctions which can be used tocreate programs that interactdirectly with WinWin This is

referred to as the WinWinApplication ProgrammersInterface or WinWin API Inaddition to the library wedistribute a test program Thisprogram should be run afterWinWin is installed to make surethe API is functioning properlyIn addition we have providedsource for the test program sopeople interested in using the APIcan imitate this program

Analysis of the Win WinProcess Model(Illustrated in Figure 7 on page 19)

Figure 7 shows a state transitiondiagram that describes thevarious states of the WinWindatabase as negotiation proceedsNodes describe the possible

states of the database whiletransitions are actions taken bythe WinWin system or by thestakeholders

Acknowledgments

Thanks go to everyone at theCenter for Software Engineeringwho participated in the designand development of WinWinThat includes Prasanta BoseMing June Lee AhmedAbdullah Cristina Gacek AnneCurran and Barry Boehm Alsoto our industrial affiliates whocontributed time and effort intesting out earlier versions Theircontributions are gratefullyacknowledged

Figure 5 Attachments Field

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 18: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN18

References[1] B Boehm and H In ldquoAids for Identifying Conflicts Among Quality Requirementsrdquo One of best papers atInternational Conference on Requirements Engineering (ICRE96) Colorado April 1996 and IEEE SoftwareMarch 1996

[2] P Bose ldquoConceptual Design Model based Requirements Analysis in the WinWin Framework for ConcurrentRequirements Engineeringrdquo IWSSD 1995

[3] P Bose ldquoA Model for Decision Maintenance in the WinWin Collaboration Frameworkrdquo Knowledge BasedSoftware Engineering (KBSE 95) 1995

[4] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements Negotiation and Renegotiation Aids ATheory-W Based Spiral Approachrdquo Proceedings of the 17th International Conference on Software Engineering(ICSE-17) Seattle April 1995

[5] M J Lee ldquoFormally Modeling the WinWin Requirements Negotiation Systemrdquo Doctoral Consortium SecondIEEE International Symposium on Requirements Engineering (RErsquo95) York March 1995

[6] M J Lee and Barry Boehm Knowledge Summary USC Center for Software Engineering Focused Workshopon Next Generation Software Processes and Their Support USC Center for Software Engineering 1994

[7] B Boehm and P Bose ldquoA Collaborative Spiral Software Process Model Based on Theory Wrdquo ThirdInternational Conference on the Software Process 1994

[8] B Boehm P Bose E Horowitz and M J Lee ldquoExperimental Results from a Prototype Next-Generation ProcessSupport Systemrdquo TRW Systems Integration Group Technology Review Summer 1994 Volume 2 Number 1

[9] B Boehm P Bose E Horowitz and M J Lee ldquoSoftware Requirements As Negotiated Win ConditionsrdquoProceedings of the First International Conference on Requirements Engineering (ICRE94) IEEE Computer SocietyPress Colorado Springs Colorado April 1994

[10] CSE Web Site and all software can be found at httpsunsetuscedu

Figure 6 Message Window

Win WinContinued from page 17

Continued on page 19

Author Contact Information

Ellis HorowitzCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089

horowitzuscedu

Joo H LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089jooleesunsetuscedu

June Sup LeeCenter for Software Engineering

amp Computer Science - USCLos Angeles CA 90089juneleesunsetsucedu

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 19: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

STN19

Figure 7 WinWin Transition States

Continued from page 18

No Unresolved Issues

No Uncovered WinC

No Open Agreements

WinC Entered

Uncovered WinC Exists

Stakeholders Notified

Stakeholders Review

WinC and Agreement

Comments Imply Consensus

Open Agreement enteredcovering WC and usingOptions Vote Initiated

Stakeholders Notified

Vote Completed

VoteSuccessful

Issue ExplainingProblem is Entered

Options ResolvingIssue are Entered

Stakeholders Notified

Option is Chosen toResolve the Issue

Vote Unsuccessful

Stakeholders NotifiedComments Imply a Concern

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 20: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

DoD DACS Products amp Services Order Form

Name PositionTitle

Organization Acronym

Address

City State Zip Code

Country E-mail

Telephone Fax

Product Description Format Quantity Price Total

The DACS Information Package Including 2 recent Software Tech News newsletters and

several DACS Products amp Services Brochures Documents FREE FREE

Empirical Data Architecture Research Facility (ARF) Error Dataset Disk $ 50 NASA Software Engineering Laboratory (SEL) Dataset CD-ROM $ 50 NASA AMES ErrorFault Dataset Disk $ 50 Software Reliability Dataset Disk $ 50 DACS Productivity Dataset Disk $ 50

Technical Reports A Business Case for Software Process Improvement Document $ 25 ROI from Software Process Improvement Spreadsheet Disk $ 40 A History of Software Measurement at Rome Laboratory Document $ 25 An Analysis of Two Formal Methods VDM and Z Document $ 25 An Overview of Object-Oriented Design Document $ 25 Artificial Neural Networks Technology Document $ 25 A Review of Formal Methods Document $ 25 A Review of Non-Ada to Ada Conversion Document $ 25 A State of the Art Report Using Defect Tracking amp Analysis Document $ 50 A State of the Art Report Software Design Methods Document $ 25 A State of the Art Review Distributable Database Technology Document $ 25 Electronic Publishing on the World Wide Web

An Engineering Approach Document $ 5 Object Oriented Database Management Systems (Revisited) Document $ 50 Software Analysis and Testing Technologies Document $ 25 Software Design Methods Document $ 25 Software Prototyping and Requirements Engineering Document $ 25 Software Interoperability Document $ 25 Software Reusability Document $ 25 A State of the Art Report Understanding amp Improving

Technology Transfer in Software Engineering Document $ 50

Bibliographic Products Rome Laboratory Research in Software Measurement Document $ 25 DACS Custom Bibliographic Search Disk $ 40 DACS Software Engineering Bibliographic Database (SEBD) CD-ROM $ 50

Number of TotalItems Ordered Cost

Method of Payment Check Mastercard Visa

Credit Card _________________________________________ Expiration Date _____________________________

Name on Credit Card ___________________________________ Signature __________________________________

Mail this form to DACS Customer Liaison Telephone (315) 334-4905DoD Data amp Analysis Center for Software Fax (315) 334-4964PO Box 1400 Rome NY 13442-1400 E-mailcust-liasndacsdticmil

This form is also on-line at wwwdacsdticmilformsorderformshtml

Note All Disksare available in

PC or Mac

FREE withSpreadsheet

SALEItem

NEW

NEW

NEW

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml

Page 21: DoD Software Tech News Vol. 3- No. 1 - CSIAC...and Versioning (WebDAV) working group of the Internet Engineering Task Force (IETF) has taken on the challenge of supporting collaborative

DoD Data amp Analysis Center for SoftwarePO Box 1400Rome NY 13442-1400

Return Service Requested

First-Class MailUS Postage

Software Tech News on the World Wide WebThis newsletter in its entirety available on the web at wwwdacsdticmilawarenessnewsletterslistingshtml

Other Collaborative Techniques Web Resources

DoD DACS Collaborative Techniques Topic Area - wwwdacsdticmil

Association for Computing Machinery (ACM) Special Interest Group onSupporting Group Work (SIG GROUP) - wwwacmorgsiggroup

Collaborative Strategies - wwwcollaboratecom

Columbia University Prototype WebDAV - wwwcscolumbiaedu~eaddywebdavhtml

Computer Supported Collaborative Work (CSCW) Systems -wwwcswaikatoacnzcsResearchcscw and wwwcsclivacuk~team-itindexhtml

ICSE 98 Workshop on Software Engineering over the Internet A List of all accepted amp presentedsubmissions - httpserncpscucalgaryca~maurerICSE98WSICSEWSSubmissionshtml

Orbit Collaborative Environment - wwwdstceduauworlds

University of California Irvine WebDAV - wwwicsuciedu~webdavUniversity of California Irvine IETF WEBDAV Working Group -wwwicsuciedupubietfwebdav

Web-Based Distributed Authoring and Versioning (WebDAV) Resourses - wwwwebdavorg

WebME - the WebME system talked about in this newsletter is available in prototype form atftpftpcsumdedupubselroseannewebme

WebSoft - using the Web as the infrastructure and external integration mechanism for a globalsoftware engineering environment - wwwicsuciedupubwebsoft

WinWin Spiral Model - httpsunsetusceduWinWinwinwinhtml