best practices in collaborative ontology engineering...best practices in collaborative ontology...
TRANSCRIPT
Best Practices in Collaborative Ontology Engineering
The Service Ecosystem Ontology in the THESEUS/TEXO project
Daniel Oberle
Senior Researcher
SAP Research, CEC Karlsruhe
Germany
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 2
1. SAP Research
2. THESEUS/TEXO
3. Service Ecosystem Ontology
4. Conclusion
Agenda
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 3
About SAP Research
Our Vision Our Mission
Be a world-class thoughtleadership partner to SAPand SAP’s customers &partners
Prepare the groundwork for SAP’sfuture growth
by acting as SAP’s IT trend scoutidentifying emerging IT trends
by researching and developing instrategically important SAPbusiness areas as well as
by leveraging entrepreneurialinventive talent
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 4
SAP Research Process– From Idea to Innovative Product
BasicResearch
Qualified Idea
Focus SAP ResearchAppliedResearch
IndustrialResearch
MarketEvaluation
Productionor Spin Off
Demonstrator Prototype Customer Pilot Innovative Project
Invent Validate Transfer
Collaborative Research
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 5
SAP’s Research & Development Network
Bangalore
Tokyo
Budapest
Sofia
Montreal
PretoriaBrisbane
Belfast
Karlsruhe
Palo Alto
Walldorf
DarmstadtDresden
SAP LabsSAP Research
SophiaAntipolis
Tel AvivShanghai
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 6
The CEC Karlsruhe
Campus-based Engineering Center
Since 1999
Birthplace of SAP Research
About 100 employees(Senior) ResearchersPhD studentsStudents
Interfacing with “Semantic Karlsruhe”PhD supervisionBilateral cooperationCooperation in publicly funded projects
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 7
1. SAP Research
2. THESEUS/TEXO
3. Service Ecosystem Ontology
4. Headaches/Outlook
Agenda
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 8
THESEUS Programme 2007-2011
TheseusProgramme
CTC (Core Technology Cluster) – Lead:
Programme Coordination:
TexoTexo“The Future
BusinessValue
Network”
Lead
ProcessusProcessus„The
SemanticEnterprise“
Lead
OrdoOrdo„Organizingyour digital
Life“
Lead
ContentusContentus„Management
of culturalassets in
digital form“
Lead
AlexandriaAlexandria„The world is
youraudience“
Lead
MedicoMedico„Scalable
Image Searchin Medicine“
Lead
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 9
TEXO
Services willbecome tradablecomposed of services of different providersbe offered, delivered & executedautomatically & supported by IT
Internet of ServicesWorldwide, trusted Service Ecosystem ofService Providers, Consumers and BrokersFor buying, selling, repurposing andcomposing services for different needsNew way of organizing the interactionbetween partner ecosystem & customerbase
€
€
€
€
Map data
Consumers &Citizens
Industry
Government
Services in a connected world
Compute
Open Plattform
Market Research
Census data
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 10
Technical Aspects
The Internet of Services and Business Webs
ServiceArchitectures
SemanticTechnologies Web2.0
Application(de)composition
Rich servicedescriptions
Informationmash-up
BusinessServices
ServiceDiscovery
InformationIntegration
Business Aspects
Attractiveness BusinessValue
Networkeffects
Businesscollaboration
Innovation BusinessModels
BusinessWeb
BusinessWeb
BusinessWeb
BusinessWeb
Internet of Services
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 11
1. SAP Research
2. THESEUS/TEXO
3. Service Ecosystem Ontology1. Motivation2. Walkthrough3. Collaborative Modeling4. Modeling Guidelines
4. Conclusion
Agenda
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 12
Motivation:Comprehensive Service Description
ServiceEcosystemOntology
Legal
Business Model
Feedback
QualityTechnical
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 13
Upper Level
Foundational Ontology
Avoid modeling from scratch
High-quality starting point
Ontology Design Patterns
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 14
Choice of Upper Level
UnclearUnclearPerdurantismBothBothEndurantism andPerdurantism
UnclearUnclearPossibilismPossibilismActualismPossibilism vs.Actualism
UnclearUnclearUnclearMultiplicativeReductionistMultiplicative vs.reductionist
UnclearDescriptiveRevisionaryDescriptiveRevisionaryDescriptive vs.revisionary
SUMOOpenCycOCHREDOLCEBFORequirement \Alternative
Ont
olog
ical
Cho
ices
YesYesNoYesYesMaturity
YesYesYesYesYesModularization
NoNoNoYesNoOntology DesignPatterns
YesYesYesYesYesOWL DL Version
SUMOOpenCycOCHREDOLCEBFORequirement \Alternative
Ext
rinsi
c P
rope
rties
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 15
dolce:Non-Agentive Non-Physical Endurant
dolce:Parameter dolce:Role dolce:Course
dolce:defines
dolce:Entity
dolce:Endurant
……
Upper Level - DOLCE
dolce:Perdurant dolce:Quality dolce:Abstract
dolce:Description
……
……
dolce:Region
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 16
Core Service Description
Core Service Description module contains information common to every service
Service, Service Description, Service Provider, Service Consumer, etc.
Ontological Foundations of Service Science [Ferrario, Guarino 2008]
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 17
dolce:Parameter dolce:Role dolce:Course
dolce:defines
dolce:Endurant
Core Service Description
dolce:Perdurant
dolce:Description
QoSParameter
ServiceConsumer ServiceProviderspecifies
ServiceDescription offeredByconsumedBy
Service
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 18
Core Modules
Different aspects are devoted their own ontology module
Each module is contributed by a different domain expert
Modules import each other and can be omitted
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 19
Core Modules
QoSParameter
ServiceConsumer ServiceProviderspecifies
ServiceDescription offeredByconsumedBy
GeneralTermsRegulation
obeyssubjectTo
dolce:Description
PriceModel
Company
NaturalPersonCore Service Descriptionhttp://www.texo-project.info/service#
Business Model http://www.texo-project.info/business#
DOLCE http://www.loa-cnr.it/ontologies/dul#
Legal http://www.texo-project.info/legal#
exhibits
dolce:Organizationdolce:PhysicalAgent
representedBy
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 20
Industry Modules
Introduce industry-specific taxonomies of service categories
Formal and natural language documentation
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 21
Industry Modules
QoSParameter
ServiceDescription specifiesRegulationobeys
Core Service Description http://www.texo-project.info/service#Legalhttp://www.texo-project.info/legal#
Automotive Industry Module http://www.texo-project.info/industry/automotive#
EcoCalculatorServiceDescriptionEcoNormobeys
specifies
ErrorRate ResponseTime Availability
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 22
Instances
Specific service descriptions are represented as instances
Instances are created by service providers
Instances can be spread and interlinked on the Internet of Services
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 23
Instances
Automotive Industry Module http://www.texo-project.info/industry/service#
EcoCalculatorServiceDescriptionEcoNormobeys
specifies
ErrorRate ResponseTime Availability
http://eur-lex.europa.eu/ http://www.ecochexx.com http://www.broker.com http://www.ratings.org
EcoLabel2007 EcoCheXx
EcoCheXxPricing
rdf:typerdf:type rdf:type rdf:type rdf:type
obeys
exhibits
EcoCheXxRating
appraises
EcoCheXxErrorRate
EcoCheXxResponseTime
EcoCheXxAvailability
specifies
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 24
1. SAP Research
2. THESEUS/TEXO
3. Service Ecosystem Ontology1. Motivation2. Walkthrough3. Collaborative Modeling4. Modeling Guidelines
4. Conclusion
Agenda
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 25
Toolkit
Collaboration Server
ContentMgmt
Plug-In
Toolkit
ContentMgmt
Plug-In
Toolkit
ContentMgmt
Plug-In
LegalExpert
Business ModelExpert
RatingExpert
GovernanceTeam
Modeling Task Force
Collaborative Modeling
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 26
Content Mgmt Plug-In
[Palma, Haase, Corcho, Gómez-Pérez, Ji: An Editorial Workflow Approach For Collaborative Ontology Development. ASWC 2008]
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 27
Modeling Guidelines – Naming conventions
Classes must be named in English in CamelCasee.g. ServiceDescription
Relations must be named in English in CamelCase with non capital lettere.g. offeredBy
Descriptions for relations should be meaningful verbse.g., "describes," "obeys," etc.
Model the inverse relation and declare the inversity in the editore.g., offers and offeredBy
Classes and relations from imported ontology modules must be specialized with thecurrent module's namespace. This is necessary to enable a sound modularizationwhere the individual modules can also "live on their own" without the importedmodules.
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 28
Modeling Guidelines – Documentation
Explanatory and verbose descriptions must be provided in English and German foreach class and relation in line with the formal axiomatization
e.g. EcoCalculatorServiceDescriptionrdfs:label de : Ökokalkulatorrdfs:label en : Eco-calculator
rdfs:comment en: “An Eco-calculator service description is a service description with the following constraints- it obeys an eco regulation- it specifies all of the following three quality of service parameters:
- availability- response time- error rate”
rdfs:comment de: ‘…‘
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 29
1. SAP Research
2. THESEUS/TEXO
3. Service Ecosystem Ontology
4. Conclusion
Agenda
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 30
Why is this relevant to IOHN?
[Integrated Operations and the Oil & Gas Ontology, OLF + PCA]
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 31
Some Headaches …
Maturity of tools
Expertise missing
Modeling expensive
Danger of over-engineering
Experts do not really model
Large effort to collaborate and streamline
How to evaluate whether design is realistic?
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 32
Q&A
Questions?
© SAP 2009 / Daniel Oberle : Best Practices in Collaborative Ontology Engineering Slide 33
Thank you!