minutes of today's conference call

[Martin W Sachs <mwsachs@us.ibm.com>]

The TP team held a conference call today.

Attendees:

Marty Sachs, Jean-Jacques Dubray, Dale Moberg, Henry Lowe, Ravi Kacker,
Chris Ferris, Jamie Clark, Antoine Lonjon.

We discussed the use of context with PartyId.

   Chris explained that a lot of peole feel that the PartyId should be a
   URI, which eliminates the need for a context indicator.  However others
   are concerned that a URI has to be registered with IANA and therefore
   prefer an explicit context identifier.  Chris pointed out that multiple
   layers of context identifier may be required.  For example, identifiers
   for different naming systems may be listed in one of more directories
   and there is no standard for the value of an identifier for a given
   naming system.  Thus the context identifier must identify both the
   directory and the naming system within  the directory.  On the other
   hand, providing a context identifier permits to parties to a CPA to
   agree between themselves what the value of the context identifier means.
   Regarding the use of URIs, Chris gave the following example:  Suppose
   that Dun and  Bradstreet obtains a URI from IANA to use with DUNS
   numbers.  Then the URI to identify a given party is the DUNS URI
   concatenated with the party's DUNS number.

   Ravi asked about routing to subunits of a party, which can be done with
   the last 4 digits of the DUNS number.  The answer is that nothing
   changes.  The last 4 digits of the URI based on the DUNS number would
   specify the routing.

   Henry said that CORBA uses URIs in conjunction with its identifiers
   (IOR) and has had no problems.

   Ravi about providing a choice bewteen a URI and a specific context
   identifier.  Chris replied that that is how it works in the TRP
   specification.  In fact the to and from addresses may be specified
   differently (e.g. one with URI and the other with a context identifier).
   He is proposing the same scheme for CPP/CPA.  Different options can be
   used for the two parties in the CPA.

Chris said that in his proposed DTD, provision is made for multiple IDs for
the same party.

Marty announced that we must have a first version of the specification
completed by January 14 in order to complete the approval process in time
for approval in Geneva.  Chris added that in Vienna, we will be processing
the comments from the first pass through Quality Review.

Marty reviewed his proposal for the CPA chapter in the first verion of the
specification.  It will consist of the XML text of the CPA, the DTD of the
CPA, descriptions of the CPA tags which are not in the CPP, and
non-normative text discussing what we know about how to compose or
negotiate a CPA from two CPPs.

Chris said that the BP team is started to look into multi-part messages and
to identify the parts.  For example, a message might consist of an order
plus one or more drawings.  Antoine clarified that this is not batching of
independent messges; the multiple parts are related and form a single
intent.  The result is a single response which may also be multipart.

Marty stated that the DTD that Chris sent out yesterday is considered
frozen (but see below) until the first version of the specification is
completed.  We should keep the discussions going and the outcome of the
discussions will be applied to the next draft of the spec.

Dale asked about the packaging profile reference.  Chris replied that
because BP is making changes in this area, he doesn't know how to specify
this reference (e.g. where in the CPA or CPP it should go).  Therefore he
removed the reference pending clarification by the BP team.  Dale expressed
concern that the BP team may be getting into a lower level issue that isn't
in their domain and that they might be micromanging the lower levels.
Chris assured him that the BP team is aware of this exposure and will take
care not to let it happen.  Marty asked Dale and Chris to track the
multipart message subject.

Chris decided to put the security/packaging link under document exchange
for now and will send out one more distribution of the restructured DTD.
This distribution will also include a link to party details and will be
accomapnied by a discussion of the new elements.  He expects to do this
distribution today.

Ravi asked for an udate on the status of the BP-TP interaction.  Marty
replied that he was very satisfied with the outcome of the discussion in
Boston although the multipart question is new.

Dale said that we need something on the CPP and CPA for the Vancouver he is
writing a joint Security/TP proposal and will discuss with Chris inclusion
of digital signature.  Marty asked him to communicate with Scott Hinkelman
on this.

Marty asked if we will need any other security specification in the CPP/CPA
besides the certificate parameters.  The answer was that this is not yet
clear.  Marty asked about signing of the CPA.  For now this will be a
futures item.

The next conference call will be Wednesday, Jan. 3, 1-2:30 PM US Eastern
time.  The main topics will be the BP specification metamodel and multipart
messages.

Regards,
Marty

*************************************************************************************

Martin W. Sachs
IBM T. J. Watson Research Center
P. O. B. 704
Yorktown Hts, NY 10598
914-784-7287;  IBM tie line 863-7287
Notes address:  Martin W Sachs/Watson/IBM
Internet address:  mwsachs @ us.ibm.com
*************************************************************************************