Minutes of OASIS OSLC CCM TC meeting on 2015-07-30

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

The draft minutes for the OASIS OSLC CCM TC meeting on 2015-07-30 are available at https://wiki.oasis-open.org/oslc-ccm/Meetings/Telecon2015.07.30, and are also included below.

These minutes will be reviewed at this week's meeting on Thursday - see the agenda at https://www.oasis-open.org/apps/org/workgroup/oslc-ccm/event.php?event_id=39040.


Meetings/Telecon2015.07.30

Event details
• https://www.oasis-open.org/apps/org/workgroup/oslc-ccm/event.php?event_id=39038 
Chat room: http://webconf.soaphub.org/conf/room/oslc-ccm 

Agenda
• Roll Call
• Scribe nomination
• Review minutes of previous meeting
• Review changes to Configuration Management draft
• Any other business
• Adjourn

Minutes
Chat transcript from room: oslc-ccm 2015-07-30 GMT-08:00
[07:06] List of attendees: David Honey (IBM), Deric Merino (PTC), Martin Sarabura (PTC), Nick Crossley (IBM), Peter Hack (IBM), Jim Amsden (IBM) (telecon only)
[07:07] David Honey (IBM): David volunteered as scribe.
[07:07] Nick Crossley (IBM): Minutes of last meeting: https://wiki.oasis-open.org/oslc-ccm/Meetings/Telecon2015.07.16
[07:08] David Honey (IBM): Minutes approved.

Review changes to Configuration Management draft
[07:10] Nick Crossley (IBM): The draft specification has been split into multiple HTML pages to make the standard easier to approach, and for ease of update, and ease of reference by teams implementing only versioning but not configuration management. The working group is open to feedback on this split; if it is considered to be a hindrance to readability we can consolidate back into a single HTML page.
Use of the boolean property oslc_config:mutable is now deprecated. Instead, baselines and streams are indicated by mandatory new types oslc_config:Baseline and oslc_config:Stream (both subclasses of oslc_config:Configuration).
The new draft removes references to the Automation and Actions specifications, and removes requirements for POST to the LDPC of configurations per component. In their place, the new draft adds an LDPC of streams per baseline, and an LDPC of baselines per stream.
The configuration match triples returned inline with a versioned resource as result of a GET with the oslc_config:allMatches header have been moved into a new ConfigurationMatches resource, still returned inline; this avoids modifying the representation of a baseline, which might have disturbed digital signatures, etc.
The new draft provides machine-readable resource shapes, using Respec to generate the property tables in the rendered HTML pages.
[07:11] David Honey (IBM): Proposal: Don't separate global CM provider info from normal CM provider info.
[07:12] David Honey (IBM): Still needs doc on client responsibilities - to be written.
[07:14] David Honey (IBM): Consensus is that proposed structure is good.
[07:15] David Honey (IBM): Marker RDF types: oslc_config:Baseline and oslc_config:Stream. Avoids confusion about what "mutable" means. Can be used in accepts and acceptedBy.
[07:18] David Honey (IBM): Previous spec defined an LPDC for a component. A GET got all the configurations of that component. A POST could be supported but client had to include appropriate RDF content to create a stream or baseline.
[07:19] David Honey (IBM): New spec has additional LDPCs for a baseline's streams and a stream's baselines.
[07:19] David Honey (IBM): Provides simpler creation factories by POST to the LPDC.
[07:20] David Honey (IBM): GET on LDPCs returns the baselines created from a stream, or the streams created from a baseline. A more specific query that is likely to be useful to clients.
[07:23] David Honey (IBM): Possible concerns: Will clients who have no prior experience of LPDCs find this burdensome?
[07:24] David Honey (IBM): Jim: OSLC is adopting LPDCs generally. Eclipse Lyo could provide a reference implementation that deals with link headers etc.
[07:27] David Honey (IBM): Change to how all matches are returned.
[07:28] David Honey (IBM): Previous spec: Add header, returned resource had extra triples in response.
[07:29] David Honey (IBM): New spec: Header references an inline resource, and this can be different subject from that used for the normal data for the resource.
[07:32] David Honey (IBM): Use case for multiple matches: Can advise the user the resolved version is ambiguous. User might focus on a narrower configuration to resolve the ambiguity.
[07:39] David Honey (IBM): rdf:nil for oslc_config:selects indicates no version is selected in a specified configuration.
[07:43] Nick Crossley (IBM): Open issues to raise: Error status values What must or must not be in a creation POST When can or must a server ignore config header or query string When can or must a client provide or not provide a config header or query string previousBaseline - handling of delete of intermediate baseline
[07:44] David Honey (IBM): New section 12 Long Operations was added since previous review request. Some open issues on that design.
[07:44] David Honey (IBM): https://tools.oasis-open.org/version-control/browse/wsvn/oslc-ccm/trunk/specs/config-mgt/config-resources.html
[07:45] David Honey (IBM): Should the progress and result be separate resources (as currently spec'd), or be a single resource that holds progress or result.
[07:45] David Honey (IBM): Welcome feedback on this.
[07:46] David Honey (IBM): OSLC automation spec seems heavyweight for use with long running operations.
[07:47] David Honey (IBM): Current spec uses some OSLC automation vocabulary, not the OSLC request class.
[07:48] David Honey (IBM): Would welcome feedback on this.
[07:57] David Honey (IBM): Will review section 12 and feedback on it in the next OSLC CM TC meeting.
[07:58] David Honey (IBM): Meeting adjourned.