Minutes of meeting from Monday 12th September

[Kevin Conner <Kevin.Conner@arjuna.com>]

Apologies for the delay in getting these minutes out

MARTIN: Last meeting was not quorate.  Need to approve the minutes from
   the last two meetings.
MARK: motion to adopt minutes from last two meetings, Greg second.
  No Objections.  Passed.

Review Action Items

a) Greg posted his proposal, discuss later in meeting.

b) Not yet done.
ACTION ITEM: Greg to write up a proposal for how to address the issue of
   asynchrony in the ACID spec by the end of the week.

c) Now a duplicate of a)

d) & e)
MARTIN: Delta is the obligation to disclose within certain time frames.
   No obligation to convert over so suggest we carry on as is.

No disagreement
Close d) & e)

Discussion continued
DOUG: New mechanism is more straight forward, not positive doing nothing
   is the way to proceed.
MARTIN: Once the TC has agreed, it has to go out to a full organisational
   ballot.  UDDI took 3-4months.  Just balancing benefits against time.
PETE: UDDI needed amendments, effort on our part is small apart from legal
   investigations.  TC calls for a vote, members vote and it must be a
   unanimous decision.
GREG: Oracle requires significant legal investigation, to avoid would be nice.

f) Done.  Proposal for issues 276 and 278 sent to the mailing list.

6) WS-Context

MARTIN: Public review came back with no comments.  There are no outstanding
   issues and no plans to change the specification.

MARK: Motion to move WS-Context to Committee Specification.
ERIC: Second
DOUG: Shouldn't Jamie conduct this vote?
MARTIN: Yes, that's correct.

MARTIN: Motion that the TC admins put out an email for a vote for Committee Specification.
No objections, motion passed.

ACTION ITEM: Martin to send out an email requesting that the TC admins conduct a
   ballot to request moving WS-Context to Committee Specification.

7) WS-CF issues

Issue 278:
KEVIN: The current extensibility mechanism allows for elements to be added to the structure
   at the same level as the CAF specific structure.  This is currently achieved by using
   the Schema ANY construct.  This mechanism introduces an ambiguity as derivations are
   made to the types.  A better solution would be to remove this ambiguity by wrapping
   these extensibility elements within an element specific to each type.

GREG: Is this to be applied to CF and above?
KEVIN: Yes, applying to CF and derived specifications is sufficient to remove any ambiguities.

Discussion over containing extensions in elements, broad agreement.

KEVIN: Motion to implement the extensibility mechanism in CF and derived specifications
   by providing an element within which all extensibility elements are to be placed.
MARK: Second
No discussion, no objections, motion passed.

Issue 279:
MARK: Issue is to add a separate operation called addParticipants.  This method would
   contain a list of participants to be enlisted, returning a list of current responses.
   Should this method be atomic or the same as current, i.e. return on first failure.
   Propose it should be first failure returns.
ERIC: Makes sense.
DOUG: As we are not changing semantics do we need the current addParticipant?
MARK: Happy to change the current addParticipant to support a list of participants.

MARK: Motion that the editors make necessary changes to allow list of participants to
   be passed with a list of URIs returned.  Failure will be on first failure.
ERIC: Second
No discussion, no objections, motion passed.

Issue 276:
KEVIN: Proposal sent to mailing list today.
MARTIN: should revisit at next meeting.

KEVIN: Also raised issues 280 and 281

ACTION ITEM: Martin to put issues 276, 280 and 281 on agenda.

Discussion on Action Item a)

(from earlier)
GREG: Motion to add XML Schema definition for an assertion stating that the service supports
   the use of an acid context.
MARK: Second
ERIC: Is this the entire policy?
GREG: Yes
ERIC: There's nothing about requires.
GREG: Not sure which is better.
ERIC: Amend to include requires

(discussion starts again)
ERIC: What happens if we require acid semantics?
GREG: Why would that be the case?
ERIC: STDL Spec had cases where applications done in the context of a TX require
   TX context as part of the call, they could end up with a separate TX.  An example
   would be moving data from one DB to another.

Discussion about Requires/Supports

GREG: Agrees that there are use cases for requires.

ACTION ITEM: Greg to send out a proposal on issue 273 by COB Wednesday.

Interop
KEVIN: Mark has made suggestions for the documents.  Most of his suggestions have
   been implemented and the response has been sent to the TC for all to read.
MARTIN: What is the state of the demo?
KEVIN: Nothing been done since the F2F, John Fuller has an implementation of context,
   I have a partial implementation.  Have just returned from LOA so things should start
   moving again.

AOB
MARTIN: Suggest there should be no more F2F for the time being.  If necessary we can
   hold longer or additional concalls.
ERIC: No objections to postponing.

Next meeting in two weeks.





ACTION ITEMS from the meeting:

Greg to write up a proposal for how to address the issue of asynchrony in the
   ACID spec by the end of the week.
Martin to send out an email requesting that the TC admins conduct a ballot to
   request moving WS-Context to Committee Specification.
Martin to put issues 276, 280 and 281 on agenda.
Greg to send out a proposal on issue 273/policies by COB Wednesday.


	Kev

-- 
Kevin Conner
Arjuna Technologies Ltd.