Minutes and meeting reminder - MetaModel team

[Karsten Riemer <Karsten.Riemer@East.Sun.COM>]

Reminder: The weekly BP/CC metamodel meeting will be tomorrow 10/24 at noon
EST, 9 am PST.

Agenda: 
  Focus on FSV layer.
  Review of revised and/or alternative XML level metamodel.
  Review of TPA requirements on XML level metamodel.
  Process for review of metamodel.

To access the call, dial 888-699-0348 domestically and +1 732-336-6000
internationally, with  a PIN of 1642#.

Minutes from Metamodel meeting 10/17/2000

Agenda:
  Review of proposal for XML level metamodel.
  Scheduling of FSV review.
  Process for review of metamodel.

Attendees:

Bob Haugen
Core Casanave
Mike Rowell
Jean-Jacques Dubrais
Anne Hendry
Antoine Lonjon
Paul Levine
Karsten Riemer
Sharon Kadlec
John Petit

1. Review of XML level proposal:

Jean-Jacques Dubrais presented the structure of the proposed DTD:

The DTD is intended to cover the semantics of the BOV layer, specifically the
sequencing of Commercial Transactions within BusinessCollaborationProtocols,
and the message exchange withing Commercial Transactions. The sequencing
semantics is intended to cover complex sequencing needs. It has a number of
possible 'rules' such as 'follows', 'optional', 'split'.  A commercial
transaction always has one request, and a set of possible responses, of which
at most one will actually occur. Sequencing on the requesting side may be
based on the message received.  The DTD also covers exceptions, for instance
handling of out-of-sequence messages.

Next steps for this proposal are mapping to XMI and mapping to FSV.
(Editorial: The proposal contained a small section on TPA, which can be
thought of as a place holder for FSV level semantics)

Jim Clark stated that he felt the proposal did not reflect the current ebXML
(Edifecs) UML profile semantics for the BOV level. He felt that new semantics
were introduced and others were left out. He felt that the proposal was
message centric, not process centric.

Cory Casanave stated that whatever the content of the models at the UML and
XML levels, the the shared part of the semantics must isomorphic, and that we
should not have two different ways of expressing the same semantics.

Jim Clark stated that the XML level model should be a pure subset of the UML
level model.

Bob Haugen stated that we must not rely solely on (traditional) sequence, but
that the model must support pre-conditions as a way of 'ordering' commercial
transactions. There was general agreement on this point, and it was decided to
provide a new revision of the proposal that reflects this.

Jim Clark offered to prepare an alternative proposal.

Antoine Lonjon and J.J. Dubrais will prepare updated version of their
proposal.

Karsten stated that the objective is a simple, minimal but sufficient XML
level representation of BOV, which can drive the XML TP layers and serve as
the 'configuration' file for a Business Service Interface.

(Editorial: After conversations with Klaus Naujok on the layers, I agree with
Klaus that it is the FSV layer we need to focus on. So next meeting will try
to refocus this effort at the FSV level. This proposal at the BOV level,
however was very useful in flushing out a lot of issues in the different
approaches)

2. Scheduling of FSV level review.

This had previously been decided as an agenda item for the Monday BP meetings.
(Editorial: Paul Levine has talked with Jim, and this will be covered through
a Lunch&Learn session at Tokyo)

3. Process for review of metamodel.

Karsten asked that all known parts of the metamodel be put out for open access
by the BP/CC metamodel team.