Joint OpenOntologyRepository-OntologySummit2008 Panel Discussion Session - Thu 3-April-2008    (1CZL)

Conference Call Details    (1CZU)

Attendees    (1D0P)

Background    (1D0S)

Two parallel initiatives are ongoing in the community, revolving around the theme of "Open Ontology Repository". On the one hand, a working group under the auspices of the OpenOntologyRepository Initiative, and on the other, the discourse (and essentially a discussion group that culminates in a two-day workshop) conducted as the main focus for OntologySummit2008.    (1D0T)

It is at the intersection of these two initiatives that this panel discussion session is being held. The OpenOntologyRepository team is taking the opportunity to have some of its members who are bringing technology and infrastructure to the table to present them side-by-side, and to discuss how these can all fit nicely together. The OntologySummit2008 folks, on the other hand would want to take the opportunity to survey (at least a subset of) the technology & infrastructure landscape to gain insight into the state-of-art in Ontology Registry and Repository.    (1D0U)

Besides hearing from the panelists, we are setting aside ample time after their briefings, for some good Q&A and discussions among all who are participating in this session.    (1D0V)

Refer to details at the respective project homepages of the two initiatives at: . & .    (1D0W)

Agenda & Proceedings:    (1D0X)

Title: An Open Ontology Repository: Rationale, Expectations & Requirements    (1D16)

Abstracts:    (1D17)

Resources:    (1D1M)

Questions, Answers & Discourse:    (1D21)

Questions and Discussion captured from the chat session:    (1D29)

 VNC2: Welcome to the: Joint OpenOntologyRepository-OntologySummit2008 Panel Discussion Session
 Subject: An Open Ontology Repository: Rationale, Expectations & Requirements - Session-2    (1DHN)
 PeterYim: to all the "anonymous' participants, Please change your name from 'anonymous' using 
           the Settings button to your real name (in WikiWord format)    (1DHO)
 PeterBenson: At the 3rd IEEE Conference on Standardization held on October 24, 2003 
              Timothy Schoechle from the International Center for Standards Research at 
              the University of Colorado, Boulder presented an interesting paper titled 
              Digital Enclosure: The Privatization of Standards and Standardization, in which 
              he made the following statement:
              In the field of ICT (Information and Communication Technology) such standards 
              specify everything from the prongs on plugs and cables to the software protocols 
              that make the Internet work.  Historically, these standards have been set largely 
              by volunteer participants in committees that operate within a wide range of 
              environments, institutional rules and social practices; but in general they have 
              espoused a traditional commitment to general principles of democratic deliberation, 
              consensus, public accessibility and balanced stakeholder representation.  
              The historical practice is now being challenged by newer, more private organizations 
              that do not necessarily have a commitment to the same principles. 
              we must be careful that the OOR clearly highlights any restrictions on the use of 
              any ontology included in the OOR.    (1DHP)
 RaviSharma: Dr. Doug Lenat Is there an attempt in OpenCyc or ResearchCyc to categorise the type 
             and complexity of relationships that improve the usability of the millions of 
             assertions, any analysis done?    (1DHQ)
 DougLenat: Answer to Ravi: Yes, we actually chose the OpenCyc relations to be the most useful ones 
            from full Cyc.  And we would be happy to revisit that and add more as other people 
            point out useful ones from Cyc that we haven't included.    (1DHR)
 LeoObrst: Question to Doug: Given Cyc's long experience with such matters, can you provide the 
           OOR group with what you would suggest as a "small set of inter-ontology alignment relations"? 
           Which are necessary and which are desirable?    (1DHS)
 DougLenat: Response to Leo: Yes, I would be happy to provide the set of (surprisingly few) predicates 
            we use to state those inter-ontology correspondences.  I will send that out and/or post it 
            today or tomorrow.    (1DHT)
 BillBug: Dr. Lenat - I strongly agree with your identifying the need for synonymy mapping.  
          A few questions - 
          (1) given Dr. Zeng's presentation, do you think SKOS works for this purpose?  
          (2) Should SKOS be used in OWL, where SKOS would essentially provide an orthogonal 
              semantics for managing an ontology lexicon? 
          (3) Should an OOR adopt a particular standard for providing these synonymous mappings?    (1DHU)
 DougLenat: Response to BillBug: Take a look at what I just promised for Leo, and then see if 
            that answers your question.  I was pleased to see what SKOS does in this particular 
            area, and need to look at it in more detail.  We need e.g. to represent cases where 
            one concept is merely strongly related to one in another ontology, not equivalent 
            to it, and if you're not careful there is a slippery slope where you end up wanting 
            all of the predicates you use WITHIN your ontology because after all those are the 
            predicates you deemed worthy/important to interrelate the terms in your ontology, 
            so it only makes sense that you might want to use them among ontologies.  
            Somewhere between that (everything) and n=1 (equivalentConceptInOntology) is the 
            sweet spot of the curve.    (1DHV)
 PatHayes: Doug: please include me in that inter-ontology mailing, thanks.    (1DHW)
 DougLenat: Will do    (1DHX)
 PeterYim: to Dr. Lenat: do you see enough differentiation even for the OOR team implementation 
           effort (note, not the OntologySummit2008 intelectual pursuit) to exists. If so, how 
           should this effort line up with the LarKC    (1DHY)
 RexBrooks: Question for Deke: Are you aware of the City GML effort?    (1DHZ)
 LeoObrst: Question for Deke: The NBIMS (slide 8, for example) seems to consist of multiple taxonomies, 
           is that right? Do you need relations among the nodes too?    (1DI0)
 PatHayes: Question for Marcia: Can you enlarge on the point on your slide 18? What do you mean by 
           'non-symmetrical' here?    (1DI1)
 MarciaZeng: Pat, for example, under 'tax', the sub-classes or narrower terms in different versions 
             from different countries are different even they all fit in the same multi-lingual thesaurus.    (1DI2)
 PatHayes: Thanks, Marcia.    (1DI3)
 RaviSharma: Dr. MarciaZeng - What would happen to the content of current ontologies if we were to 
             exclude relationships associated with Thesauri and Taxonomies (the natural language 
             free flow usage e.g. less logic based). How much would be left? Further, is there a 
             value chain concept of Information content from complexity or concatenation of relationships?              
             Further many times I get a feeling that knowledge and understanding floodgates would open 
             if we could transform (transliterate or map) concepts and understanding from major cultural 
             and lingual groups alone?    (1DI4)
 MarciaZeng: Dr. RaviSharma,  
             1. I think even though the relationships are not consistently logic and highly structured, 
                many hierarchical relationships best presented in the domain which do not have a 
                widely accepted 'taxonomy' like in biology.   So, if you only want to use the work for 
                controlling synonyms and disambiguate terms, you may miss lots of good hierarchical structures. 
             2. Yes, it is difficult to map among the multilingual and multi-cultural vocabularies, 
                e.g., about traditional medicine or even geographic regions.    (1DI5)
 PatHayes: Question for Denise: can you enlarge on your distinction between metadata and metainformation? 
           I'm not familiar with the second term.    (1DI6)
 RaviSharma: Ms. MalaMehrotra  As we have in astrophysical coordinate systems, or in translating 
             among different inertial and other reference coordinate systems, are solvers and translators 
             similarly being conceptualized among different ontology representations of TIME or 
             INTERVAL or DURATION?    (1DI7)
 RaviSharma: Ms. MalaMehrotra - Further, is the value of human collaboration in terms of enriching 
             the value of domain based relationships for future automated processing and is there a 
             roadmap for the same with use cases?    (1DI8)
 RaviSharma: Dr. RobRaskin  Is slide 5 a hierarchical relationship among ontologies or is it taxonomy 
             of science disciplines? Especially since the type of relationships are not indicated among them?    (1DI9)
 LeoObrst: Question to Rob: Is the "Data Ontology" the same as the SWEET ontologies, or is the former part 
           of SWEET? And do you find you need "rules"?    (1DIA)
 LeoObrst: Question to Rob: What kind of methodology did you use in the creation of the SWEET ontologies, 
           i.e., requirements-driven (competency questions?) and then what kind of review process, and 
           gateway/quality criteria?    (1DIB)
 PeterYim: to Dr. Raskin ... given your prevailing work in SWEET and (which is an OOR in 
           its own right), I would like to invite you to join us in the OOR-team (which is cast as an 
           implementation effort, independent of the OntologySummit2008 effort, which focuses in the 
           intellectual discussion of the Open Ontology Repository & Resgistry subject matter.) see: 
 PeterYim: the same invitation goes to anyone else who want to contribute to the OOR-implementation effort    (1DID)
 LeoObrst: Question to Marcia: I think you are saying that the OOR will need both terminological resources 
           and ontologies/conceptual resources? Do you think SKOS can provide a terminological framework 
           for term taxonomies/thesari and perhaps a term->concept indexing into ontologies?    (1DIE)
 MarciaZeng: Leo: yes I think the terminological resources could be part of the conceptual resources.  
             Some 'ontologies' claimed on the Web are no better than a taxonomy.    (1DIF)
 MarciaZeng: Leo: SKOS probably can handle well thesauri, but, for how to present mapping results, 
             SKOS may not be perfect so far. It is working on that direction.    (1DIG)
 ToddSchneider: Pat, Moving towards 'truth' should be seen as the 'to-be' that an OOR will evolve too.    (1DIH)
 PatHayes: Todd, I think that many of us would strongly disagree. Which was exactly my main point.    (1DII)
 MichelleRaymond: Panelists - It was noted that the Architecture for an Ontology Repository may differ 
                   depending on the Ontologies held (and? the relationships between held Ontologies).  
                   If this is so, I suspect this is not simply based on the complexity of the Ontologies 
                   (dictionary, taxonomy, relational data map, ontology, ontologies with guiding upper 
                   ontology...)  A) Can we have one Arch. approach?  If the statement before is accepted, 
                   what are the characteristics that direct you toward selection of the Architecture 
                   structure? Examples?    (1DIJ)
 PatHayes: Michelle: depends what you mean by 'architecture'. As far as management/metadata issues go, 
           I think there can be a common uniform framework. But if that means all using the same ontology 
           language, for example, then I'd say no, for an open ontology.    (1DIK)
 ToddSchneider: Can we use the TOGAF notion of 'architecture'?    (1DIL)
 MichelleRaymond: Reply to Todd re TOGAF - My understanding of TOGAF is only at the higher level of 
                   different "layers" for enterprise level structures and rules, data structuring and 
                   storage, components and application availability and connectivity and ? I think the 
                   "network" level.  These are important and part of the framework for any architecture.  
                   Where a thread might help is in describing the notation of TOGAF and suggestions as 
                   to its usefulness.  Does the TOGAF force a specific direction of centralized, 
                   decentralized, federated or other core concept for architecture data locations and 
                   management practices?    (1DIM)
 ToddSchneider: Leo, Would it be possible to extend your list of requirements by culling the presentations 
                from these last two meetings?    (1DIN)
 PeterYim: to Todd, answering your question for Leo ... YES, of course (this is what these panel discussion 
                sessions were designed for)    (1DIO)
 MichelleRaymond: Reply to ToddSchneider's question to Leo - Absolutely, the Content Committee is culling 
                these presentations and question threads.  However, it is very helpful for individuals to 
                push forward their personal favorite Ah-ha's and questions as new discussion threads on the 
                OntologySummit2008 list.    (1DIP)
 KenBaclawski: Leo asked at the last session if the panelists could prioritize their requirements.  
               Specifically he asked about the top three services of an OOR.    (1DIQ)
 DougLenat: See my slide 9: the 4 things we would want from a good host    (1DIR)
 DougLenat: Commons (not GNU) license; provenance kept; agreement on at least a few key relations; 
            agreement on inter-ontology relations.    (1DIS)
 PatHayes: doug: why not the GNU licence? Too "open" ??    (1DIT)

Audio Recording of this Session    (1D2B)

Join us at the next OOR-Panel Discussion - see: ConferenceCall_2008_04_10    (1D34)