ppy/oor_chat-transcription_20101015a_unedited.txt


Chat transcript from room: oor_20101015
2010-10-15 GMT-08:00
[06:08] PeterYim: . Welcome to the OpenOntologyRepository: OOR Team Conference Call - Fri 2010-10-15 .
[06:08] PeterYim: session page is at: http://ontolog.cim3.net/cgi-bin/wiki.pl?OOR/ConferenceCall_2010_10_15
[06:08] PeterYim: .
[06:10] anonymous morphed into JouniTuominen
[06:10] KenBaclawski: Here is the architecture diagram (same as on the wiki): http://ontolog.cim3.net/work/OpenOntologyRepository/2010-09-17_OOR-Dev-Take-4/OOR-Architecture--KenBaclawski_20100917.png
[06:11] KenBaclawski: This is the current OOR API (in Java) based on BioPortal: http://ontolog.cim3.net/work/OpenOntologyRepository/2010-09-17_OOR-Dev-Take-4/OOR-API--KenBaclawski_20100917.java
[06:22] PeterYim: see " Simplified BSD License" at: http://opensource.org/licenses/bsd-license.php
[06:29] PeterYim: == Key requirements of the new OOR architecture? ==
[06:29] MichaelGruninger: support for multiple ontology specification languages
[06:30] PeterYim: I think it must accommodate both the existing codebase (BioPortal) and new contributions
[06:30] KenBaclawski: The use cases are certainly important. The posted architecture is based on the use cases.
[06:30] MichaelGruninger: services of the OOR should be independent of the ontology specification language
[06:30] YuriyMilov2: Java based
[06:30] TillMossakowski: It should be flexible enough in order to easily integrate new formats and services
[06:32] PeterYim: as per OntologySummit2008_Communique ... we want to be on a Service Oriented Architecture
[06:32] MichaelGruninger: support for ontology metdata
[06:32] PeterYim: metadata should be an extension of OMV too
[06:48] YuriyMilov2: I agree to discuss the Java implementation later
[06:48] PeterYim: == Who are the immediate software contributors that we will need to accommodate? ==
[06:49] PeterYim: NCBO / BioPortal
[06:49] KenBaclawski: My team at Northeastern
[06:49] MichaelGruninger: Cameron Ross is leading the architectural design of COLORE
[06:50] TillMossakowski: at Bremen, we have a BioPortal clone and the Hets system
[06:52] TillMossakowski: the BioPortal clone is athttp://ontologies.informatik.uni-bremen.de/
[06:54] PeterYim: MikeDean: contributions from Raytheon-BBN
[06:55] TillMossakowski: We might want to use tntbase.org (but we have close contact to them)
[06:56] PeterYim: NeOn project? .... (of course, OMV is from the NeOn project, and that is definitely in)
[06:58] JouniTuominen: ONKI ontology service http://www.onki.fi, project page http://www.seco.tkk.fi/services/onki/
[06:58] PeterYim: I'll continue the conversation with MathieuD'Aquin, AndreasHarth and EnricoMotta
[06:58] PeterYim: (above) ref. NeOn involvement
[06:59] BartGajderowicz: My contribution wold be library development
[06:59] BartGajderowicz: so code part
[06:59] TillMossakowski: I am back again. Have been cut off, while the phone call still seemed to continue...
[07:00] PeterYim: == Enumerating our options and candidate architectural approaches ==
[07:01] BartGajderowicz: As part of my thesis, I have been working on ontology matching, and have developed an application, which I would like to contribute to the OOR architecture
[07:03] TillMossakowski: slide no. 7 of Immanuel's presentation http://ontolog.cim3.net/file/work/OpenOntologyRepository/2010-09-17_OOR-Dev-Take-4/SOA-for-OOR--ImmanuelNormann_20100917.pdf
[07:03] YuriyMilov2: Is Gool;e's AppEngine an appropriate platform? I'd like to add some stuff from there
[07:04] YuriyMilov2: I mean the Google of course :)
[07:06] TillMossakowski: I think a good thing is define a RESTful interface of services. We have implemented one for Hets, see http://trac.informatik.uni-bremen.de:8080/hets/wiki/RESTfulInterface . I am not meaning to use this one, but just want to give an idea of what a RESTful interface is.
[07:07] PeterYim: ref. BioPortal high level architecture - http://ontolog.cim3.net/cgi-bin/wiki.pl?OpenOntologyRepository_Architecture#nid2J32
[07:07] KenBaclawski: I have posted an architecture which is based on the use cases, but I am open to suggestions especially concerning how to accommodate multiple specification languages for ontologies.
[07:10] JouniTuominen: +1 for RESTful interface. We have developed LOOS API (Linked Open Ontology Services, in spirit of LOD) for ONKI, see http://www.yso.fi/onkirest . Actually the intefrace it's not that RESTful, more like a method-based HTTP API :)
[07:12] KenBaclawski: @Till Concerning the slide in Immanuel's presentation, this is not an SOA architecture diagram. In fact the entire diagram is just two diagrams: the GUI and the core services.
[07:13] KenBaclawski: @Till I have the experience of developing web services for a military contract, and SOAP/WSDL was much easier to develop than REST.
[07:14] MichaelGruninger: Architecture for COLORE can be found in slide 5 in the presentation at http://ontolog.cim3.net/file/work/OpenOntologyRepository/2010-09-10_CL-support-for-OOR/An-OOR-Implementation-for-COLORE--CameronRoss_20100910.pdf
[07:16] TillMossakowski: We have a bit of experience with SOAP, but have not implemented one project with both REST and SOAP.
[07:16] KenBaclawski: The current trend in SOA is for REST and SOAP to converge, so the distinction may disappear.
[07:16] TillMossakowski: Can WSDL be used together with either of them?
[07:24] KenBaclawski: Yes, it is possible for a single service to support both SOAP and REST at the same time.
[07:28] KenBaclawski: In my previous post about Immanual's presentation, I mean the diagram is two components, not two diagrams.
[07:30] PeterYim: (back to key tenets) .... robustness, scalability, etc. for the repository is definitely important
[07:31] YuriyMilov2: http://www.memo.in.th/wp-content/uploads/2010/09/Service-Oriented-Cloud-Computing-Architecture.pptx
[07:35] YuriyMilov2: SOCCA supports easy application migration from one cloud to another and service redeployment to different clouds by separating the roles of service logic provider and service hosting/cloud providers. It promotes an open platform on which open standards, ontology are embraced.
[07:53] PeterYim: decision: delegate to committee
[07:53] PeterYim: committeee to run with an open transparent process, and will keep syncing all of us up
[07:53] PeterYim: meetings to be announced so others can join as observers
[07:54] PeterYim: only code contributors will participate in making the final decisions ... other "architecture only" suggestions will only be provided as references to the committee
[07:55] PeterYim: the precepts agreed earlier as "Key requirements of the new OOR architecture" will drive the committee's work and decisions
[07:57] PeterYim: on the committee: rep from NCBO, NEU, UToronto, Bremen, Raytheon-BBN, NeOn ... (to be decided) ONKI, ...
[07:58] PeterYim: (also to be decided) YuriyMilov
[07:59] PeterYim: Architecture Committee will work up a architecture first ... and then develop the API
[08:00] PeterYim: each part in the committee to nominate their committee member ... Peter to contact those who are not here today
[08:01] PeterYim: will defer discussion on development platform
[08:11] PeterYim: we will continue to discuss regular meeting times on the [oor-forum] list ... please provide input into the doodle poll if you haven't already - goto: http://doodle.com/a2snxkpkd8hhxek4
[08:12] PeterYim: as discussed previously, this is almost about time we start this - have alternate Friday OOR team meetings - one for OOR high level issues, content, admin etc., and the other for deep-water technical meetings.
[08:13] PeterYim: ALL: agreed ... we can possibly even have alternate (Friday) start-times for each of these meeting sets
[08:13] PeterYim: thanks everyone ... very productive meeting indeed!
[08:13] PeterYim: -- session ended 8:10am PDT --