[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: OSLC CORE TC Minutes April 13, 2017
|
Scribe
Attendees
Regrets
Resolutions
Actions
Chat transcript from room: oslc
[07:05] Martin Sarabura (PTC): Martin to scribe [07:05] Martin Sarabura (PTC): No minutes to vote to accept [07:05] Martin Sarabura (PTC): Nick no progress [07:05] Martin Sarabura (PTC): Jim: CS pretty much done, just one minor edit [07:06] Martin Sarabura (PTC): Chet next steps email sent to Jim [07:08] Martin Sarabura (PTC): Need min 3 statements of use, one of which must come from member [07:09] Martin Sarabura (PTC): Requirements are minimal, TC can decide what is an adequate statement of use [07:09] Martin Sarabura (PTC): TCs report to the Member Section, steering committee has vote on whether CS can be submitted as standard [07:10] Martin Sarabura (PTC): TC should decide on what is an acceptable Statement of Use (SOU) and then accept such SOUs [07:11] Martin Sarabura (PTC): SOU could be an assertion saying "we implement OSLC", other is a reference implementation of OSLC v3 with test suite that tests each feature of the spec [07:11] Martin Sarabura (PTC): Sam Padgett already set up an Eclipse Lyo implementation [07:12] Martin Sarabura (PTC): Not appropriate as a foundation for a production quality server but is probably fine for testing purposes [07:12] Martin Sarabura (PTC): Has not been maintained for the past year+ [07:14] Martin Sarabura (PTC): Only a few test cases created but pattern is there - not clear whether it follows the clauses. MUST/MAY/SHOULD [07:14] Martin Sarabura (PTC): MAY/SHOULD are equivalent for testing purpose [07:15] Martin Sarabura (PTC): Unlikely we will be able to test all clauses but not required for OASIS [07:16] Martin Sarabura (PTC): Jim recommends we adopt a statement of minimum criteria [07:19] Martin Sarabura (PTC): Suggested compromise: Because V3 is compatible with V2 to ensure seamless introduction of the new standard to existing implementations, and because V3 shift to LDP is useful for future implementations but
there is a lot of V2 work out there, criteria is that implementer of V3 assert that their V2 implementation is compatible with the V3 spec. [07:20] List of attendees: David Honey (Persistent/IBM), Harish K (Software AG), Jim Amsden (IBM), Martin Sarabura (PTC), Nick Crossley (IBM), Nick Crossley (IBM)1, anonymous [07:20] Martin Sarabura (PTC): This puts onus on V2 implementer to make that assertion [07:21] Martin Sarabura (PTC): Proposal is consistent with Chet's guidance [07:21] Martin Sarabura (PTC): Risk is relatively low of significant errors in the spec [07:22] Nick Crossley (IBM): An implementer may also assert that their V3 implementation is compatible with the V3 spec. [07:22] Nick Crossley (IBM): We are just saying there is (as yet anyway) no formal V3 test suite that an implementer MUST pass. [07:22] Martin Sarabura (PTC): Jim has discussed with SC and they agree these SOUs are acceptable [07:24] Martin Sarabura (PTC): Nick: Not clear how complete and/or correct the V2 test suite on Lyo is [07:25] Martin Sarabura (PTC): Jim has run it against the v2 implementation [07:25] Martin Sarabura (PTC): on lyo [07:25] Martin Sarabura (PTC): There is the OSLC v3 RI on Lyo, no current test suite [07:26] Martin Sarabura (PTC): Would require significant updating, may even be incompatible with v2 [07:27] Martin Sarabura (PTC): Domain specs are relatively independent of v2 vs v3 in core [07:28] Martin Sarabura (PTC): Martin to write up formal proposal, to be voted on next meeting [07:28] Martin Sarabura (PTC): Nick and Jim could come up with similar for Change mgmt spec [07:29] Martin Sarabura (PTC): Nick: Will accept any organization's statement without proof, or MAY accept ... and TC will review? [07:30] Martin Sarabura (PTC): Ultimately it is up to the committee to forward to SC anyway so it's the latter [07:31] Martin Sarabura (PTC): Jim: ability to integrate with somebody else's V2 implementation is a powerful SOU [07:31] Martin Sarabura (PTC): Third party could assert that they have used implementations successfully [07:31] Martin Sarabura (PTC): Demonstrated interopera b ility makes a stronger case but is not required [07:32] Martin Sarabura (PTC): Risk is that someone could assert without testing and it turns out to be limited in some way that doesn't expose a problem [07:34] Martin Sarabura (PTC): Any client/server pair demonstrates interoperability - doesn't demonstrate completeness of course [07:35] Martin Sarabura (PTC): Is anybody interested in picking up OSLC v3 implementation on Lyo? Maybe Axel Reichwein, perhaps academic project? [07:36] Jim Amsden (IBM): https://wiki.eclipse.org/Lyo/ToolchainModellingAndCodeGenerationWorkshop [07:37] Martin Sarabura (PTC): Jim used this to generate an OSLC implementation recently [07:38] Martin Sarabura (PTC): Jim now technical lead for Lyo [07:39] Martin Sarabura (PTC): Tool chain editor allows you to create model of OSLC domain, creates EMF representation of model, unfortunately not resource shapes [07:41] Martin Sarabura (PTC): Allow you to model capabilities such as which dialogs, query support, etc. Creates a web application that you can implement - Jetty server [07:41] Martin Sarabura (PTC): Generates code into Eclipse web project [07:42] Martin Sarabura (PTC): Biggest problem is it doesn't use Turtle vocabularies, and is a static generative approach. [07:43] Martin Sarabura (PTC): Change of topic [07:43] Martin Sarabura (PTC): Publish shapes and vocabularies [07:44] Martin Sarabura (PTC): One method is to set up redirects at open-services.net to OASIS documents [07:44] Martin Sarabura (PTC): Other method is to simply copy content to open-services.net [07:45] Martin Sarabura (PTC): Nick: Need to have a stable open-services.net forever [07:46] Martin Sarabura (PTC): Jim: SC will own it going forward. Some discussion in sc re new content, we could host the spec there in addition to the content [07:48] Martin Sarabura (PTC): Document for the redirect approach that Arthur Ryman wrote up, Nick to send link [07:50] Martin Sarabura (PTC): If we updated open-services.net vocabularies, will we break anything? [07:50] Martin Sarabura (PTC): Nick says nothing, all we do is add terms [07:51] Martin Sarabura (PTC): We do need all formats (not just rdf/xml) [07:52] Martin Sarabura (PTC): GET with accept header of application/rdf+xml, and you get it in a slightly different format, and somebody using XPath to parse it, what will we break? [07:52] Martin Sarabura (PTC): David: Indicates implementation is fragile [07:53] Martin Sarabura (PTC): Nick: Correct XPath should be OK. [07:53] Martin Sarabura (PTC): We have run shape checker on those files [07:55] Martin Sarabura (PTC): We originally generated the v2 pages automatically; this was fed into the v3 rdf [07:56] Martin Sarabura (PTC): Jim to look into how we should publish. Since it's stable now we can do it [07:57] Martin Sarabura (PTC): Could have normative Turtle in OASIS, refer to it from open-services.net [07:57] Martin Sarabura (PTC): Nick: No content negotiation, doesn't work [07:57] Martin Sarabura (PTC): Refuses to interpret what comes back from OASIS as anything but HTML [07:58] Martin Sarabura (PTC): So, let's just copy it over to open-services.net [07:59] Martin Sarabura (PTC): Nick still has xslt script (ANT build script) to do conversion between formats [08:00] Martin Sarabura (PTC): Don't need html generation since that comes with respec [08:01] Martin Sarabura (PTC): Doesn't use namespaces - should be readable [08:01] Martin Sarabura (PTC): Can get Jena to print out right namespaces [08:01] Martin Sarabura (PTC): RDF validate may do it [08:02] Martin Sarabura (PTC): meeting adjourned |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]