OSLC CORE TC Minutes June 23, 2016

["Sarabura, Martin" <msarabura@ptc.com>]

Scribe

• Nick Crossley (IBM)

Attendees

• Bill Chown (Mentor Graphics)
• David Honey (IBM)
• Harish Krishnaswamy (Software AG)
• Jad El-khoury (KTH)
• Jean-Luc Johnson (Airbus)
• Jim Amsden (IBM)
• Martin Sarabura (PTC)
• Nick Crossley (IBM)

Regrets

Resolutions

Actions

·         Nick to update ReSpec template and send email pointing to it

• Nick to compare his local copies of TRS and ILDP documents to those found online (see references in minutes)
• David to raise issues for known gaps such as what to do if there are two create events for the same resource, or support for selective indexing.
• Nick to raise issue around support for delegated access control
• Jim to raise issue regarding support for notifications via MQTT or other such publish/subscribe method

Chat transcript from room: oslc

[07:03] Nick Crossley (IBM): Nick to scribe

[07:04] Nick Crossley (IBM): List of attendees: Bill Chown (Mentor Graphics)2, David Honey (IBM), Jad El-khoury (KTH), Jean-Luc Johnson (Airbus), Jim Amsden (IBM), Martin Sarabura (PTC), Nick Crossley (IBM)

[07:04] Martin Sarabura (PTC): https://wiki.oasis-open.org/oslc-core/Meetings/Telecon2016.06.09

[07:04] Nick Crossley (IBM): Minutes approved

[07:05] Nick Crossley (IBM): Martin: update from previous meeting: collecting list of emails for notifying about Core CSPRD

[07:06] Nick Crossley (IBM): Martin: have received some feedback from Chet - mostly minor, but some need to be fixed before publication

[07:07] Nick Crossley (IBM): Jim: some of these changes are in areas we thought we had already made compliant with OASIS formats, so there is some concern we are iterating without making progress

[07:07] Nick Crossley (IBM): Jim may try to make ReSpec changes himself

[07:08] Nick Crossley (IBM): Martin to follow up on date for ballot; the date must be manually inserted into the drafts!

[07:09] Nick Crossley (IBM): Next on agenda: TRS and ILDP

[07:10] Nick Crossley (IBM): Jim: Is there a master template for ReSpec documents?

[07:11] Nick Crossley (IBM): Nick: there is a template in ReSpec itself. Needs tidying up and bringing up to date.

[07:11] Nick Crossley (IBM): Action item: Nick to update template and send out email pointing to it.

[07:11] Harish K (Software AG): sorry team for joining late.

[07:12] Harish K (Software AG): joining the audio shortly

[07:12] Martin Sarabura (PTC): Hi Harish!

[07:12] Nick Crossley (IBM): Jim intends to copy current TRS into ReSpec format, and similarly ILDP

[07:13] Jim Amsden (IBM): http://open-services.net/wiki/core/TrackedResourceSet-2.0/

[07:14] Nick Crossley (IBM): For ILDP, Jim des Rivieres sent me his latest copy

[07:14] Martin Sarabura (PTC): http://open-services.net/wiki/core/IndexableLinkedDataProvider-2.0/

[07:15] Nick Crossley (IBM): Nick: I have not checked to see if those last two are the same

[07:19] Nick Crossley (IBM): Martin: what is the difference between the term 'context' in configuration context vs. access context?

[07:20] Nick Crossley (IBM): Nick: Configuration context is a dynamic thing - a user works in a configuration context, and that context is used to resolve any references to versioned artifacts. Access context is a more static thing; it relates a set of artifacts that have, or should have, the same access rights.

[07:21] Nick Crossley (IBM): Jim: Access context and TRS Patch should be moved into the primary TRS spec.

[07:21] Nick Crossley (IBM): TRS Patch should stay as is, since LDP Patch is not progressing.

[07:22] Nick Crossley (IBM): David: much of the rest of ILDP is non-normative guidance - should it be a separate document?

[07:22] Nick Crossley (IBM): Jim: Making this a Committee Note makes it easier to update - recommends doing that

[07:23] Nick Crossley (IBM): Shapes and discovery also need to move to normative spec

[07:24] Nick Crossley (IBM): Martin: since the volume of non-normative guidance is quite large, it is better suited to a separate document

[07:24] Nick Crossley (IBM): David: some details need to be addressed - for example, there's no definition of what happens if there are two create events for the same resource?

[07:25] Nick Crossley (IBM): Jim: David should raise an issue for that, and any other known gaps in the spec

[07:25] Nick Crossley (IBM): David: another gap may be around support for selective indexing

[07:25] Nick Crossley (IBM): Nick has an issue to raise on support for delegated access control

[07:26] Nick Crossley (IBM): Last two points are action items

[07:26] Nick Crossley (IBM): Or last three

[07:26] Nick Crossley (IBM): Martin: Any other topics to be raised?

[07:28] Nick Crossley (IBM): David: describes performance issue with TRS in its current form - polling may not be the ideal mechanism if you want near-real-time updates. Should we consider adding a push-based notification?

[07:29] Nick Crossley (IBM): Jim: David should also raise this as an issue, and we should consider use of MQTT or other systems.

[07:29] Nick Crossley (IBM): David: have done some experiments with MQTT.

[07:31] Nick Crossley (IBM): Martin: note that adding parts to the normative spec - even MAY statements - is that we'd have to enlarge the scope of the reference implementation to include them, and are we prepared to take on that work?

[07:36] Nick Crossley (IBM): Discussions about guaranteed delivery, or lack of guarantees, transaction boundaries, etc.

[07:43] Nick Crossley (IBM): What about TRS 1.0 compatibility?

[07:44] Nick Crossley (IBM): Bill and Martin lean toward leaving out requirements for 1.0 compatibility, and see who complains.

[07:45] Nick Crossley (IBM): Jim: remind everyone that we are still trying to get enough proposers for the OSLC Domains TC. If you know of anyone who would support this, please let Jim know.

[07:46] Nick Crossley (IBM): Martin: why June 7th for Core 3.0? We had a ballot at that time, and it passed, so that is the right date.

[07:48] Nick Crossley (IBM): Jim wants to get in touch with Jad for the code generator in Eclipse Lyo - would like to get rid of some technical debt. Bugzilla adapter is broken. GET on creation factory URL goes to Bugzilla, gets back HTML, but that HTML is badly formed, and SAX parser fails, even with validation supposedly turned off.

[07:48] Nick Crossley (IBM): Update to new Jetty means significant restructuring is required.

[07:50] Nick Crossley (IBM): Jim and Jad discussed pros and cons of updating to new versions of libraries. Difficult to predict client requirements. Jim thinks we should update to be compatible with latest libraries.

[07:51] Nick Crossley (IBM): Jad: believe he has a version with Bugzilla working.

[07:52] Nick Crossley (IBM): Meeting adjourned