Dale:
<JD>
From: Dale Moberg
[mailto:]
Sent: Monday, April 17, 2006 2:09
PM
To:
Subject: [ebxml-msg] Correction
RE: [ebxml-msg] wd-12 section 2.2.3.1 question
(sorry, made a paste error on the section
header value)
Is the third subcase a typo or editorial leftover?
The One-Way Push Message Exchange Pattern
(pasted in wrong section header)
This MEP involves a single ebMS user
message. The message sending is initiated by a Sending MSH
either:
as a SOAP One-way MEP instance,
as a SOAP Request in a SOAP
Request-Response MEP instance, or
as a SOAP Response in a SOAP
Request-response MEP instance.
The One-Way Pull Message Exchange Pattern of section 2.2.3.2
of course has a single ebMS user message that is transferred as a SOAP Response
in a SOAP Request-response MEP instance.
So what is the third case supposed to capture?
<JD> I am not sure why we added that
3rd case… The only reason I can think of is to allow a form of
optimization where an MSH would piggyback [ebMS] one-ways on the SOAP response
of another ebMS one-way. I think it can be removed – causes issues with
reliability.
Also, where is a One-Way Push of a user
message of type Response (so the refToMessageId is used)?
<JD> This would involve two SOAP MEPs,
not just one. So that is no longer a “simple” ebMS MEP, but an
aggregate ebMS MEP, not described in core part. We had this initially described
as the “Two-way Push” aggregate MEP, decided to move it in part 2.
It would be described as two ebMS One-way Push in opposite directions, one
referring to the other (RefToMessageId). Do you see a need to say a word on
this in part 1?
Jacques