Joint OpenOntologyRepository-OntologySummit2008 Panel Discussion Session - Thu 10-April-2008    (1D3J)

Conference Call Details    (1D3W)

Attendees    (1D4R)

Background    (1D4U)

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.    (1D4V)

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.    (1D4W)

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.    (1D4X)

Refer to details at the respective project homepages of the two initiatives at: http://ontolog.cim3.net/cgi-bin/wiki.pl?OpenOntologyRepository . & . http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2008    (1D4Y)

Agenda & Proceedings:    (1D4Z)

Topic: Developing an Ontology of Ontologies for OOR    (1D58)

Panelists' Presentation Titles and Abstracts:    (1D5A)

A Framework for an Ontology of Ontologies - MichaelGruninger    (1D5D)

      [ slides ]    (2Q8J)

Under the appellation of "ontology" are found many different types of artifacts created and used in different communities to represent entities and their relationships for a wide range of purposes.The Ontology Summit 2007 "Ontology, Taxonomy, Folksonomy: Understanding the Distinctions," was an attempt to bring together various communities having a different understanding of what is an ontology, and to foster dialog and cooperation among these communities.    (1DC7)

The results of the Summit identified a limited number of key dimensions along which ontologies can be characterized and to provide operational definitions for these dimensions.    (1DOD)

This talk will give an overview of the framework that is constituted by these dimensions.    (1DOE)

The OMV Ontology Metadata Vocabulary - PeterHaase    (1D5E)

      [ slides ]    (2Q8K)

In my presentation I will talk about OMV, a proposal for an Ontology Metadata Vocabulary that has been originally developed within the Knowledge Web project and is now maintained by the OMV consortium. I will explain the coverage and scope of OMV, some basic design rationales as well as applications. In particular, I will talk about uses and tool support for OMV developed in the NeOn project.    (1DC8)

Ontology metadata in NCBO's BioPortal - NatashaNoy    (1D5F)

      [ slides ]    (2Q8L)

The National Center for Biomedical Ontology (NCBO; see http://bioontology.org) is developing BioPortal, an open-source repository of ontologies, terminologies, and thesauri of importance in biomedicine. BioPortal is accessible at http://bioportal.bioontology.org. Users can access the BioPortal content interactively via their Web browsers or programmatically via Web services. An alpha version of the next BioPortal release, which is available at http://alpha.bioontology.org supports facilities for visualizing ontologies, for creating mappings between ontologies, and for adding "marginal notes" through which the user community can comment on ontologies and discuss modeling decisions.    (1DC9)

I will discuss the metadata requirements for BioPortal and our approach for representing metadata about ontologies, maintaining the metadata as the ontologies evolve, and encouraging users to provide metadata both on the ontologies that they author and the ones that they use in their own projects.    (1DCA)

Metadata Support for OMG's Emerging Ontology and Vocabulary Management Initiative - ElisaKendall    (1DHK)

      [ slides ]    (2Q8M)

As presented in EvanWallace's talk in the OOR-Panel Session (ConferenceCall_2008_03_27), the OMG is exploring what it would take to host a repository to manage standard ontologies and vocabularies. An initial request for information was issued at the recent OMG Technical Meeting in Washington, DC. (http://www.omg.org/cgi-bin/doc?ontology/08-03-02.pdf, Ontology and Vocabulary Management RFI). Searchable metadata is a critical aspect of this anticipated facility, and as such, we are actively investigating and seeking input on ontology management metadata.    (1DHL)

Potential starting points include, but are certainly not limited to: the OMV Ontology Metadata Vocabulary augmented via lessons learned in developing the BioPortal, current work in the ISO JTC1 SC32 WG2 effort to update the ISO 11179 Metadata Registry standard to include concept systems, vocabularies, and ontologies, and lessons learned in provenance and trust research by Dr. DeborahMcGuinness and her colleagues. I will briefly present some of the requirements and challenges we have identified for a successful ontology management capability at OMG, and look forward to active discussion and input from this community.    (1DHM)

( more details coming! ) - see: developing details    (1D5B)

Resources    (1D5J)

Questions, Answers & Discourse:    (1D5V)

Questions and Discussion captured from the chat session:    (1D63)

 PeterYim: Welcome to the Joint OpenOntologyRepository-OntologySummit2008 Panel Discussion Session - Thu 10-April-2008    (1D64)
 * Subject: Developing an Ontology of Ontologies for OOR    (1DPG)
 * Session co-chairs:    (1D3L)
   o Professor BarrySmith (NCOR; University at Buffalo) &
   o Professor MichaelGruninger (University of Toronto)    (1DPH)
 * Panelists:
   o Professor MichaelGruninger (U of Toronto) - "A Framework for an Ontology of Ontologies"    
   o Dr. PeterHaase (AIFB, U of Karlsruhe; Project NeOn) - "The OMV Ontology Metadata Vocabulary"
   o Dr. NatashaNoy (Stanford-BMIR) - "Ontology metadata in NCBO's BioPortal"
   o Ms. ElisaKendall (OMG; Sandpiper Software) - "Metadata Support for OMG's Emerging Ontology and Vocabulary Management Initiative"    (1DPI)
 ArturoSanchez: Question for MichaelGruninger: why limit the formalisms used to express semantics to Model Theory (slide 9)?    (1DPJ)
 ArturoSanchez: For MichaelGruninger and PeterHaase: how would you characterize OMV according to the 
                OntologySummit2007 model?    (1DPT)
 RexBrooks: For MichaelGruninger: Could we have some examples of Semi-Structured Ontologies?    (1DPK)
 RexBrooks: For MichaelGruninger: I would also like to have some examples of descriptive v. prescriptive 
            ontologies, and how we can capture that in metadata, to avoid inadvertent misuse.    (1DPN)
 JohnGraybeal: For MichaelGruninger: On slide 17, is it really the Design Methodology that you care about, 
               or the validation/review characteristics of the result?  The important thing would seem to be 
               whether an ontology has been established as consistent or conformant along various axes, 
               not how it got to that state.    (1DPO)
 ToddSchneider: In general we should assume we will not know all of the 'items' that will need to be 
                represented in the metadata associated with an artifact and design for this.    (1DPP)
 JohnGraybeal: For MichaelGruninger: The characterization of WordNet during slide 11 (Level of Structure) 
               lost me a little, I'd like to hear that again. (An example table, showing how a number of 
               ontologies fits into these various dimensions, would be quite helpful.)    (1DPM)
 ArturoSanchez: For JohnGraybeal: see http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007_Population_Session 
                and search for Population-spreadsheet, various ontologies where mapped to the model 
                during the summit    (1DPQ)
 JohnGraybeal: For ArturoSanchez: This is brilliant, though the copy I found is largely empty. 
               Where can I find one that integrates all the results?    (1DPU)
 ArturoSanchez: For JohnGraybeal: if you look under "Breakout Group Task:" you will find the 
                various results from all the groups. In particular, the link to the document 
                about WordNet is http://ontolog.cim3.net/file/work/OntologySummit2007/workshop/Population-spreadsheet_WordNet-as-ontology_20070423b.doc    (1DPX)
 PeterYim (added subsequently): to JohnGraybeal: you can also look under 
           http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007_Communique_Session#nid108F  
           ... check out the various breakout groups' work products as well as 
           OlivierBodenreider's "Framework Dimensions Sorted" piece 
           at: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007_Communique_Session#nid10PD    (1DQI)
 KenBaclawski: With one of my students, I have evaluated 40 large ontologies using the OntologySummit2007 
               Framework Dimensions.  I have written up some conclusions about the results and could 
               present them at some point.    (1DPR)
 RexBrooks: For PeterHaase: Small note on "Party," don't you think we need a category for "Agent" 
            which can be a subtype of either "Organisation" or "person," as well as a "party" in its 
            own right, e.g. a spider/webcrawler bot?    (1DPS)
 PeterHaase: to RexBrooks: I believe the concept Party captures what you propose as Agent    (1DPV)
 PeterHaase: we did not want to call it agent as the term is so overloaded    (1DPW)
 RexBrooks: Thanks PeterHaase. It may require a qualifier or another term to be able to specify 
            the Party as some kind of software agent v. an organisation or person. I will need 
            to take a good look at OMV.    (1DQ1)
 MalaMehrotra: For NatashaNoy: Are the mappings being stored as metadata? Slide 5.    (1DPY)
 NatashaNoy: For MalaMehrotra: Yes, mappings are also metadata (and have the same issues, 
             such as versioning, contributions from communities, etc.)    (1DPZ)
 MalaMehrotra: For NatashaNoy: SO what language - or what level of expressivity - are you 
               all requiring in the BioPortal for mappings?    (1DQ0)
 ArturoSanchez: For NatashaNoy: what is the definition of "validity"? without a formal definition 
                it is not possible to check for such property ...    (1DQ3)
 ArturoSanchez: For NatashaNoy: so, by allowing ontology versioning by creating brand new ontologies 
                from existing ones, that means interoperability among versions might not be possible 
                ... any comments?    (1DQ4)
 JohnGraybeal: Awesome lineup, Peter. Must leave now, will catch up later.    (1DQ7)
 ToddSchneider: For PeterHaase: Does OMV support a distributed structure (e.g. P2P)?    (1DQ8)
 RexBrooks: For PeterHaase: Would OMV be capable of being used as a "Classification Scheme" 
            in ebXML RIM, if you happen to know?    (1DQ2)
 PeterHaase: to RexBrooks: The OMV API for the registry that I mentioned is actually an 
             "OMV Specialized ebXML Registry"    (1DQ5)
 PeterHaase: so we used the extension mechanisms of ebXML    (1DQ6)
 RexBrooks: Thanks PeterHaase, that is terrific for the groups I work with.    (1DQ9)
 PeterHaase: we have P2P-like applications that make use of OMV and have a P2P extension to describe 
             the peers, but there is nothing P2P specific in OMV itself    (1DQA)
 PeterHaase: what kind of structures were you thinking about?    (1DQB)
 PeterHaase: regarding the ebXML registry, you may want to look at "  D6.4.1: Realisation & 
             early evaluation of NeOn service-oriented registry and repository " under 
             http://www.neon-project.org/web-content/index.php?option=com_weblinks&catid=17&Itemid=67    (1DQC)
 PeterHaase: what we also do have is the separation between logical identifiers of 
             ontologies (ontology URI) and their physical location, this can be used to support 
             replication and caching    (1DQD)
 ToddSchneider: To PeterHaase: In that case (i.e. the ontologies are maintained locally), 
                is OMV being used more for a registry? If so, can there be multiple registries 
                (i.e.distributed)? For example, different COP's maintaining their own 'instances' 
                of thier OMV registry for their domain/COP.    (1DQE)
 PeterHaase: yes, we have both registries and repositories using OMV. They can be distributed, 
             in fact, the Oyster system I mentioned is such a distributed system    (1DQF)
 PeterHaase: it also supports the propagation and synchronization of the metadata    (1DQG)
 RexBrooks: To all panelists: This is really terrrific as the groups I work with head directly 
            into building registry-repositories and ontologies. Thanks All!    (1DQH)
 RexBrooks: My question is for PeterHaase: Could you elaborate on the difference, with examples, 
            if you can, between prescriptive and descriptive ontologies?    (1DQJ)
 PeterHaase: I am afraid I cannot, MichaelGruninger brought this distinction up, perhaps he can 
             better answer [--ed. which MichaelGruninger did answer to verbally. Refer to the audio archive.]    (1DQK)
 EvanWallace: I don't understand this distinction in this context.    (1DQL)
 AnnWrightson: I'm uncomfortable with the use of "structured" to apparently mean "logically clean". 
               Surely these are different notions? An "unstructured" ontology in the expression-sense 
               stated in slide 11 could have in fact a considerable amount of structure in the usual sense 
               of the word (eg follow a component-assembly-structure defined in EXPRESS) 
               [--ed. which MichaelGruninger did answer this verbally when AnnWrightson posed the 
                question again (below). Refer to the audio archive.]    (1DPL)
 AnnWrightson: My question: To MichaelGruninger, regarding Structure. I'm thinking in particular of 
               the HL7v3 RIM (Reference Information Model) that is used as a theory-of-what-there-is 
               (a semi-formal upper ontology) to both describe and prescribe the kinds of things 
               can be said in a largish family of derivative standards. The RIM has internal structure, 
               and the related set of standards has structure (that could be expressed eg using 
               STEP concepts). How would/should this kind of structure be handled in our dialogue, 
               distinct from the quality of being "structured" that is present in MichaelGruninger's slide 11?    (1DQM)
 EvanWallace: For some content in some ontology language forms, some metadata could be    (1DQN)
 EvanWallace: automatically generated (DL expressivity, satisfiability, etc.).    (1DQO)
 EvanWallace: I believe that some OMV tools do this already for OWL DL ontologies.    (1DQP)
 PeterHaase: yes, this is what e.g. Watson does, it generates/extracts most metadata automatically    (1DQQ)
 EvanWallace: PeterHaase: Does OMV or Watson provide any support for Common Logic or some other FOL form?    (1DQR)
 PeterHaase: in principle it can capture metadata about ontologies in any ontology language    (1DQT)
 AnnWrightson: Following on from the response just now, and thinking back to the NASA presentation 
               on ConferenceCall_2008_03_20: though we shouldn't expect the OOR to provide or perform 
               configuration management of ontologies, I guess there should be guidance concerning 
               provision of versioning infomration that will support users who want to reuse ontologies 
               within a strong configuration management or product line development environment.    (1DQS)
 ChaimZins: Just a short remark: We need to diferentiate between the topic "Ontology of ontologies" 
            and metadata issues related to the field of ontology study. 
            [--ed. to which BarrySmith did answer, verbally. Refer to the audio archive.]    (1DQU)
 ArturoSanchez: For PeterHaase: Good point ... so, metadata management is informed (so to say) by 
                the semantics of your model, using which concepts such as "validity", "extensions", 
                "editing", "refinement" can be equally formally defined (in theory, at least ) Thanks.    (1DQV)
 ToddSchneider: To BarrySmith: Perhaps we need a variant or sub-repository of the Open Repository that 
                would support a peer review process to meet higher quality standards?    (1DQW)
 PeterHaase: to ArturoSanchez: yes, in theory and to some extent already in practice    (1DQX)
 ChaimZins: I agree with your remark.    (1DQY)
 EvanWallace: BarrySmith's comment just underlines my opinion that the term Ontology    (1DQZ)
 ChaimZins: The first question is "what is ontology?"    (1DR0)
 EvanWallace: of Ontologies is not a good term for describing ontology metadata needed for a repository.    (1DR1)
 EvanWallace: Cute terms are great for slideset titles but bad for describing discussion topics.    (1DR2)
 ArturoSanchez: Ontology of Ontologies in this context is: using the metadata approach to represent the 
                knowledge about ontologies. Therefore, applications will be inherently limited by the models 
                used to characterized ontologies. For instance, consider the example for which the ontology 
                used to characterize ontologies in the repository has less "expressive power" than ontologies 
                it is trying to define    (1DR3)
 MichelleRaymond: AnnWrightson mentions that (paraphased) 'in product line development environments 
                  strong configuration management is required.'  Currently, there are industry product 
                  providers that very much MUST point to the exact ontology AND the associated metadata 
                  for that ontologies when providing a product release.  To do this now, sometimes 
                  the "real" standard can't be linked in, as there is concern that it has remained static 
                  or at least a needed level of consistency.  Thus, the vendor includes a copy of 
                  the ontology and associated metadata, files, documents and the like in their 
                  product release and does not encourage customers to visit the "real" source.  This can be 
                  a legal issue of "fair use" when a copy is made.  It is also disassociating the product 
                  from the further value potentially available in the "real" ontology and the repository 
                  that holds it.  I believe we need to address this issue both in firm rules that SHALL NOT 
                  allow revision of a "status-final" ontology and that we must further consider requiring 
                  some meta-data associated with an ontology as also being "status=final."  Comments about 
                  my statements are very much desired as is turning this into a conversational thread.    (1DR4)
 EvanWallace: We should start a thread about versioning on one of the email lists which captures both 
              AnnWrightson's and MichelleRaymond's comments on today's chat session.  Would one of you 
              do it or should I? 
              --ed. see: follow-up discussion thread started by EvanWallace    (1DR5)

Audio Recording of this Session    (1D65)

Join us at the OntologySummit2008 workshop, April 28~29, 2008; either on-site at NIST, Gaithersburg, Maryland, USA or participate remotely! - see deails here.    (1D6Y)