ppy/oor_session-transcript_unedited_20121030a.txt ---------- Chat transcript from room: oor_20121106 2012-11-06 GMT-08:00 [PST] ---------- PeterYim: welcome to the = OpenOntologyRepository: OOR "Strategy and Funding" Discussion Session - Tue 2012_11_06 = session page: http://ontolog.cim3.net/cgi-bin/wiki.pl?OOR/ConferenceCall_2012_11_06 . == Proceedings: == . [07:29] PeterYim: Hi Ken! [07:31] PeterYim: I don't see you on the voice conference bridge yet, Ken [07:31] KenBaclawski: I am having some trouble connecting. [07:33] BobSmith: hi all [07:33] PeterYim: Ken is in! [07:43] MichaelGruninger: What is the role of OOR sessions: 1) distributed software development of OOR as an artifact. (2) Forum for general issues related to ontology repositories, such as modularity, provenance, versioning. [07:45] MichaelGruninger: With respect to content, OOR will be most successful when it plays a role in ontology best practices. For example, encouraging authors of ontology papers in conferences (such as FOIS, WOMO, FOMI) and journals (Applied Ontology, Artificial Intelligence, ...) to submit their ontologies to an OOR instance. [07:45] ToddSchneider: Project Management [07:46] PeterYim: == brainstorming - on the steps (ii) (iii) & (iv) of a strategic planning exercise - now ... http://ontolog.cim3.net/cgi-bin/wiki.pl?OOR/ConferenceCall_2012_11_06#nid3HMW [07:46] BobSmith: What is the primary purpose of an OOR Strategy? (Assumes variety of audience: Such as Big BUILDING and City DATA MODELS with the Natl Institute of Building Standards's National Building Information Modeling Standard 3.0 (just starting) [07:47] MichaelGruninger: OOR needs to more actively engage other projects that are related to ontology repositories [07:47] KenBaclawski: @Michael I would add these: (3) OOR as a standard API for the purposes of interoperability and federation (i.e., operationally), (4) Increasing content. [07:47] PeterYim: ALL: please approach this waring your hat as a Designer, Planning and Convening the OOR Initiative [07:48] CoryCasanave: Eventualy this would need to be done under some standards org like OMG, W3C or Oasis [07:50] BobSmith: A "current" and perhaps temporary "Threat" to Seaboard Cities such as NYC Atlantic City, Huntington Beach, San Francisco is Sea Level Rise and rethinking City Building Codes and infrastructure requirements at Federal, State, Regional, and Local levels. Some States, such as Calif., have a Chief Enterprise Architect and a Chief GIS Officer with potential great needs for OGC - like ontology references [07:50] KenBaclawski: Related standards already exist: MDR, OMV; and others are in progress: MFI. No doubt there are others. It should be a standard built on existing standards. [07:52] ToddSchneider: Cory, couldn't we continue will our development and then at some point in the future approach a standards body? [07:53] anonymous morphed into John Sowa [07:54] PeterYim: personally, I see OOR as a community driven "collabroative" effort in (i) infrastructure, (ii) software/system development, and (iii) content [07:56] FrankOlken: Technically, W3C is not an accredited standards organization like ANSI, ISO, or OMG. [07:56] John Sowa: The best standards are almost always a clean up of some de facto standard. A good example is ECMAScript, which cleaned up JavaScript. [07:56] PeterYim: let's use (A) infrastructure, (B) software/system development, and (C) content, instead [08:00] PeterYim: who could be our "role models" for (A), (B) and (C)? [08:00] John Sowa: I don't believe that any technology should be proposed as a standard until there are some successful applications. The standards organizations once tried the idea of "proactive standards". But without some practical applications, it's a bad idea to propose any design as a standard. [08:01] PeterYim: MikeDean has suggested, previously, that the Apache Foundation would be a good role model for (B) [08:01] KenBaclawski: John, would you regard Bioportal as an example of a successful repository project? [08:04] FrankOlken: Many of the standards organizations (e.g., ANSI L8 and ISO SC32 for Metadata Registry Std) have preexisting commitments to possibly overlapping standards (such as ISO/IEC 11179 or various OMG stds) which they may want us to conform to. [08:08] MichaelGruninger: I'm not sure that OOR should be considered as a standards project. We are providing infrastructure that uses existing standards for ontology representation and relationships. [08:08] ToddSchneider: Have to go. Cheers. [08:09] MichaelGruninger: We only need to guarantee that ontologies in different repositories can be uniformly accessed and described; standards for these aspects already exist [08:10] PeterYim: update: incorporating Gruninger and Baclawski's inputs above: look at the OOR Initiative as: (A) infrastructure, (B) software/system development, (C) content, (D) a standard API, and (E) a community Forum for general issues related to ontology repositories ... (in no particular order) [08:11] CoryCasanave: In summary: I support continuing with the implementations track and that implementation work may best be done under Apache or something similar. These implementations will use existing standards but may also indicate the need for others. We should consider collaboration with an SDO and think about the roadmap to do so. [08:11] MichaelGruninger: To some extent, OOR should be involved in a bit of a culture change -- rather than everyone developing their ontologies in isolation, OOR should provide the infrastructure that will finally deliver the promise of ontology sharing and reuse. [08:12] MikeBennett: @Michael agreed. That seems to suggest two distinct threads of activity: 1. syntactic; 2. semantic. [08:12] John Sowa: I support Till's idea. The [08:13] John Sowa: If and when Bioportal is extended to support Ontoiop + Common Logic + other standards, then it would be a good example. [08:14] CoryCasanave: @Michael: Much of the tooling and infrastructure is centered on developing ontologies rather than reusing them. [08:14] CoryCasanave: @Michael: So yes, focusing on the collaboration and reuse is very important [08:15] John Sowa: There is no difference between development, maintenance, and reuse. All large development projects build on and extend other implementations. [08:15] MikeBennett: I'm already self muted. [08:16] FrankOlken: I am now self muted. [08:19] John Sowa: I have also spoken with some people in medical informatics, who have been very unhappy about the "pressure" they have felt to use ontologies in Bioportal. That is a serious danger of premature standards -- they can block innovation. [08:20] TillMossakowski: OOR has a nice infrastructure diagram (see http://ontolog.cim3.net/cgi-bin/wiki.pl?OpenOntologyRepository_Architecture/Candidate03 ), the only drawback being that it has not been implemented so far. The only existing implementation seems to be BioPortal, but BioPortal does not follow the OOR architecture. To remedy this situation, we have started ontohub.org (sources at https://github.com/ontohub/ontohub ), which starts with an architecture simpler than OOR's, but with the goal to be eventually extended to that architecture. The semantic background is provided by the OntoIOp ISO standard initiative, providing a smeantics for federated ontologies involving different languages like OWL, RDF, Common Logic, UML. Ontohub.org could be federated with BioPortal in the future. [08:26] MichaelGruninger: References to standardization in OOR are premature and/or misdirected. The only standards that are relevant (for ontology representation, ontology relationships, ontology metadata) already exist or are under development. We only need to guarantee that ontologies can be imported/exported between repositories and that there are protocols for software applications to uniformly access ontologies within ontology repositories. [08:27] TillMossakowski: is there is list of these standards (also explaining their use for OOR)? [08:27] TillMossakowski: is there a list of these standards (also explaining their use for OOR)? [08:27] MichaelGruninger: Ontology representation language standards: RDF, OWL, Common Logic [08:28] MichaelGruninger: Ontology relationship standards (under development): OntoIOp [08:28] MichaelGruninger: Ontology metadata standards: OMV (?) [08:31] TillMossakowski: probably also REST or SOAP for communication among repository components (although I am not sure whether these are standards...) [08:31] PeterYim: if we were to tweak our strategy, and bring us from moving toward irrelevance to moving towards the center of attention - I suggest we re-focus our OOR efforts to revolve around (hot themes like) (I) BigData-BigSystems (ref. http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012 ), and (II) Ontology-based Standard (ref. http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologyBasedStandards ) [08:32] PeterYim: ALL: please suggest how that "re-focus" or tweak would look like [08:32] PeterYim: ALL: also, there may be similar themes that are just as compelling, please suggest what they are? [08:32] TillMossakowski: sorry, have to leave now, will read chat later, so please go on capturing the discussion... [08:33] PeterYim: thanks for joining us, Till. Before you go, please list times on Tuesdays that might work well for you (1 to 1.5 Hr slots) [08:33] PeterYim: @Till - we do hope you can join us regularly on future OOR meetings [08:34] TillMossakowski: yes, sure. Generally anything before 6 pm CET works for me on Tuesdays. [08:35] FrankOlken: By John's logic SQL would have adopted some sort of backward compatibility to CODASYL databases. [08:35] PeterYim: @Till - thanks [08:38] FrankOlken: Rules folks also have RuleML. [08:45] FrankOlken: As I recall, Datalog is more expressive than SQL, i.e., with respect to recursive queries. [08:46] FrankOlken: Typed Datalog would be interesting. I am not sure this is right setting to develop (or standardize) such a thing. [08:49] MikeBennett: I would strongly support (2) at least [08:54] PeterYim: proposal: we re-focus our OOR efforts to revolve around: (I) BigData-BigSystems (ref. http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012 ), and (II) Ontology-based Standard (ref. http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologyBasedStandards ) [08:55] PeterYim: ^"Standards" (instead of "Standard") [08:58] John Sowa: ... the typed Datalog would include the SQL WHERE-clause as a proper subset. It would also include SPARQL as a proper subset -- except for the FILTER and OPT abominations. [09:00] PeterYim: we have consensus on the above, but will need to elaborate and work on the details on how adopting (I) & (II) would impace (A) thru (E) [09:02] KenBaclawski: The meeting time was moved to 7:30am PST. [09:04] PeterYim: [consensus] now regular OOR meetings will be on Tuesdays, starting 7:30am PST / 10:30am EST / 4:30pm CET / 15:30 GMT/UTC (like today, which is one hour earlier than before) - session duration: ~1.5 Hours. [09:05] PeterYim: no meeting next Tuesday, as ISWC is in session [09:06] PeterYim: therefore, our next OOR meeting will be on Tue Nov-20, starting 7:30am PST / 10:30am EST / 4:30pm CET / 15:30 GMT/UTC ... we will continue today's conversation, and will drill down, go more granular, until we get to specifics on tactics and action plans given the re-focused short/medium-term direction [09:06] PeterYim: great meeting! [09:07] PeterYim: -- session ended: 9:03am PST -- [09:08] List of attendees: BobSmith, CoryCasanave, FrankOlken, John Sowa, KenBaclawski, MichaelGruninger, MikeBennett, MikeDean, PeterYim, TillMossakowski, ToddSchneider, anonymous ---------