Date:
23 June 2011 Time:
7:00 AM Pacific, 10:00 AM Eastern, 3:00 PM UK, 4:00 PM Frankfurt, 5:00 PM Haifa, 7:30 PM Bangalore Call In Number: (emailed)
Participation request: contact
JimConallen
Agenda
- Review new open issue regrading link type properties
- Plan the finalization activities
- Review scenarios, ensure compatibility with 2.0 spec.
Attendance
Regrets: Alanna Zito
Atendees: Clyde D Iscuspit, Daniel Berg, John Crouchley, Sandeep Kohli, Steve Speicher, Jim Conallen
Minutes
Jim C. opened up the meeting with the desire to start the last steps of finalization. Steve, a vetran of delivering OSLC specs, clarified the process. Jim and Steve will offline create the proper vocabulary document for OSLC AM resource defintions (in our case just AM Resource and Link Type Resource). After that an additional review by the core team might be in order, and with finished covenants we essentially just declare final.
Jim, in what many might have thought a rather windy explanation, raised a new open issue regarding the use of the property types in the link type resource. Jim explained, the use of the link type resource, and compared it to the primary approach for obtaining human readable information about link type predicated (just GETting them in rdf form). Since purl.org and w3c.org provide machine readable resolutions on the URIs they define, we should be consistent with how they do it. So instead of using dcterms:title and description we use rdfs:label and rdfs:comment as they do.
Given this changeis late in the game, it will not significantly affect the only know implementers (Design Manager, System Architect, and Rio).
Everyone on the call was ok with this change.
The next topic was a review of the summary of provider and client requirements. The only interesting discussion came in the section of authentication requirements. John raised the viewpoint of a third party client. It would be eaiser for them to know that at least one form of authentation would be available in all providers, citing OAuth as the most common one so far. Steve also pointed out that the CM specification has OAuth support as a SHOULD requirement, unlike the current MAY in the AM specification.
We agreed to share this suggestion to the larger team via the list server, to get more comments. But the consensus from this group was to elevate provide OAuth authentication to a SHOULD level.