Notes on the CPP/CPA Meeting on Dec 6th
by
G. Neelakantan Kartha
(with some additions by Dale Moberg)
In Attendance:
Pallavi Malu
Hima Mukamala
Brian Hayes
Marty Sachs
Dale Moberg
Jean Zheng
David Fischer
Arvola Chan
Jamie Clark
Selim Aissi
Tony Fletcher
Neelakantan Kartha
Administrative
There was some discussion about the time and location of the next face to face. Dale suggested that 2.5 days should be enough time, with 0.5 additional days allotted to negotiation sub-team. January28-30 was suggested as a possible time and San Francisco as a possible location. Dale hopes to firm this up soon.
Dale said that people interested in BTP issues should subscribe to the ebxml-cppa-btp mailing list.
Sub-team Reports
Brian talked about two types of retries---retries happening due to business failure and retries happening due to transmission failure. [The UMM experts indicate that Recurrence pertained to the lower level, reliable messaging resends of the same message. Arvola pointed out that this is also indicated by the BP attribute "IsGuaranteedDeliveryRequired." It should be possible to treat the Recurrence as a most a hint, and simply make the Retry value recorded in the CPA, the definitive RM parameter. If a BPSS treated the RM local failure report as part of the failure condition at the BP level, would any bad interactions arise? None were identified. Dale added.]
Hima made the CPPA conformance paragraphs available to the Oasis IIC TC.
Arvola reported on the recent voting meeting of the MSH group. The major decision that was made was that the target date for the final version of the spec was determined as January 11, 2002. The finalized version of the spec will be circulated by January 4. [Arvola reported that the perMessage changes seemed stable. The MessageOrder debate continues to swirl, and it is not clear how to describe capabilities, constraints, requirements, or agreements in this area.]
Other Discussions
There was a long discussion on various versions of persistDuration. At the end, there was s consensus that persistDuration in the CPPA refers only to reliable messaging. Marty wanted some lines in the specification to explicitly state that persistDuration does not refer to the notion of conversations, and that the implementers might need to store data belonging to a conversation for a period longer than the persistDuration.
Dale made an inventory of the elements beind added or changed for the 1.1 version, and needs volunteers to help on some of the blank spaces for authors and reviewers. The first inventory draft was included as an attachment on the announcement of this teleconference's agenda, but updated versions will be coming out weekly to provide a quick assessment of where we are in revisions.
|