RE: [ebxml-cppa-negot] Re: Negotiation pattern, transactions, CPP A

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

Brian,

I agree with your colleague that the conversation ID is a runtime concept.
However, it is up to the application to indicate to the middleware (or MSH)
when a conversation begins or ends. So, indicating the boundaries of the
conversation really is a logic or semantic matter.

It is possible, of course, that the application code might not see the
actual conversation ID.  An implementation might, for example, instantiate
a new instance of the collaborative process when new conversation is
started and incoming messages are then simply routed to the proper process
instance based on the conversation ID.

My take on the discussion thus far is that we should not assume that the
value of a conversation ID is available to the application process.
However, we do have to decide on the boundaries of a conversation and
define the autonegotiation choreography accordingly.   We might also have
some application-level identifier that corresponds to one instance of the
conversation, which I think is where this discussion started.

Does BPSS say anything about conversation boundaries?  If not, does it at
least define specific beginning and end points for one pass through the
choreography defined by the BPSS instance (which would implicitly be the
conversation boundaries)?  If there is nothing in the BPSS spec that
indicates (or at least implies) the beginning and end of a conversation,
then the ebXML specifications provide no guidance to the runtime with
regard to starting and ending conversations.  To me that would be a
deficiency in the BPSS.

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


                                                                                                                              
                      "Hayes, Brian"                                                                                          
                      <Brian.Hayes@Comme        To:       "'Arvola Chan'" <arvola@tibco.com>, Martin W                        
                      rceone.com>                Sachs/Watson/IBM@IBMUS, "Hayes, Brian" <Brian.Hayes@Commerceone.com>         
                                                cc:       ebxml-cppa-negot@lists.oasis-open.org                               
                      03/17/2002 01:50          Subject:  RE: [ebxml-cppa-negot] Re: Negotiation pattern, transactions, CPP   
                      PM                                  A                                                                   
                                                                                                                              
                                                                                                                              
                                                                                                                              



A comment from one of my Commerce One colleages regarding conversation Ids:
-----Original Message-----
From: Yuan, George
Sent: Thursday, March 14, 2002 1:26 PM
To: Hayes, Brian
Cc: Chen, Qiming
Subject: RE: [ebxml-cppa-negot] Re: Negotiation pattern, transactions, CPP
A


Brian,   here is my 2-cent:  conversation_id is a runtime concept to ensure
the choreography, not a logic or semantic one.  The issue of "the shipment
notice and/or the invoice will cover multiple purchase orders" seems at the
logic or semantic level.  As long as the collaboration is defined to handle
the shipment notice in such a way, it should be acceptable...

-George
>