Minutes for OASIS OSLC CCM TC on July 31st

[Nick Crossley <ncrossley@us.ibm.com>]

The minutes for the OASIS OSLC CCM TC meeting on July 31st are now available at https://wiki.oasis-open.org/oslc-ccm/Meetings/Telecon2014.07.31.

Agenda

• Roll Call
• Scribe nomination
• Review minutes of Minutes of 22 May 
• Change Management and Core status update (reordered to be first)
• Configuration Management topics - see below
• Any other business
• Adjourn

Configuration Management

• Review recent changes to the draft spec: skew and ordering of contributions, returning info on secondary matches, marking acceptability of contributions, compact rendering

Resolutions

• Approved minutes of previous meeting: Minutes of 19 June 

Minutes

Chat transcript from room: oslc-ccm 2014-07-31 GMT-08:00

[07:06] Nick Crossley (IBM): Calendar link: https://www.oasis-open.org/apps/org/workgroup/oslc-ccm/events.php?day=1406815541#linkbar

[07:06] List of attendees: Bill Chown (Mentor Graphics), David Honey (IBM), Nick Crossley (IBM), Peter Hack (IBM), Sam Padgett (IBM), Steve Speicher (IBM), anonymous

[07:08] Peter Hack (IBM): Peter is scribe.

[07:09] Nick Crossley (IBM): https://wiki.oasis-open.org/oslc-ccm/Meetings/Telecon2014.06.19

[07:10] Peter Hack (IBM): Minutes from last meeting are approved.

[07:12] Peter Hack (IBM): Status update on Change Management and Core moved up in agenda.

[07:14] Peter Hack (IBM): Steve reports slow progress due to summer schedules. Some actions spec updates that need to be reflected in CCM.

[07:15] Nick Crossley (IBM): https://tools.oasis-open.org/version-control/browse/wsvn/oslc-ccm/trunk/specs/config-mgt.html

[07:16] Peter Hack (IBM): Notable changes in spec annotated with "please review".

[07:17] Steve Speicher (IBM): Renders better as http:// http://tools.oasis-open.org/version-control/browse/wsvn/oslc-ccm/trunk/specs/config-mgt.html

[07:19] Peter Hack (IBM): Configurations can specify subtypes for accepts and acceptedBy.

[07:19] Peter Hack (IBM): If anything is accepted, it must be a global configuration.

[07:22] Peter Hack (IBM): Nick proposes that 64-bytes should be sufficient length for contribution strings and restricted to ASCII characters.

[07:22] Peter Hack (IBM): A string is used instead of a number to allow for easy ordered insertions.

[07:23] Peter Hack (IBM): Nick will add a "should" to the description to account for the length.

[07:25] Peter Hack (IBM): Nick to clarify ordering.

[07:28] Peter Hack (IBM): For example, Rational Team Concert might use rtcStream for the type of a stream configuration, whose "accepts" might have componentBaseline, while the componentBaseline type might use rtcStream in the acceptedBy.

[07:30] Peter Hack (IBM): In general, to be configuration-aware, clients must always be prepared to provide a configuration context which is why that can be provided in multiple ways. Can be included for instance, in all GET requests.

[07:32] Bill Chown (Mentor Graphics): Sorry to have to drop off

[07:32] Peter Hack (IBM): However, this is a challenge for compact rendering - it's not clear where to include that. The reference to the previews and the icons must be resolved to the requested context.

[07:33] Peter Hack (IBM): Therefore, the client can pass the results directly to the browser.

[07:33] Peter Hack (IBM): Please give thought to that change - if provider has to include that information in this case, does that open the door to requiring that for other cases as well?

[07:33] Peter Hack (IBM): Nick to look for missing text after dangling "or".

[07:35] Peter Hack (IBM): One alternative - have client modify results before passing to browser.

[07:36] Peter Hack (IBM): Steve - Using a URI template would be a possibility but current proposal sounds OK.

[07:37] Peter Hack (IBM): Ordering of contributions exists so we can resolve component skew.

[07:39] Peter Hack (IBM): Most SCM systems can't handle version skew - and it's expensive to do. This is why the spec addresses component skew but not version skew.

[07:45] Peter Hack (IBM): For all matches searches, the result may contain a Link to a pointer to a configurationMatches resource (for providers that implements the optional complete matching behavior).

[07:49] Peter Hack (IBM): Component lookup is not defined yet because Actions have not been specified - still to come.

[07:53] Peter Hack (IBM): Spec currently requires that server run query for configurationMatches and must persist that for some length of time since requests must be made one component at a time; David expressed some concern over this approach, and would prefer some way to return all the match info in the original request.

[07:56] Peter Hack (IBM): Nick will add a section on caching.

[07:59] Peter Hack (IBM): No plans yet around a reference implementation or test suite - still to come.

[08:02] Peter Hack (IBM): Meeting closed.