[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: i082 - Is there a more detailed proposal?
|
> a proposed direction rather than a formal proposal Right, we don't have a formal
proposal yet. We still need to agree on the real meaning
we want to give to "response message" and "request-response
pattern". Exact places that are of a concern: On CDII (Jan 13): L237: response
message L256: response
message L301: response
message L316: response
message L394: response
message L402: response
message L611:
request-response pattern Two ways to go from there: (1)
decide
that all above expressions must be understood in a SOAP MEP context, where "response
message" actually means "SOAP response message in a SOAP
Request-response MEP". This has some consequences on the use of the underlying
transport. (2)
Decide to
not tie CS/CSR and other request-response ops to any particular SOAP MEP, in
the same way as no assumption is made about which kind of SOAP MEP is used for
carrying messages sent reliably. Then any mention of "the response
message" in the spec is misleading and superfluous. For example, reword L237
as: "...responds with a message containing
either a <wsrm:CreateSequenceResponse> or a CreateSequenceRefused
fault." I initially suggested (1) which was what
the spec appeared to mean. But I now see some good things in (2). My main issue
here is that we need to clearly settle for one way or the other. > I'm concerned with the assumption the proposal
in this issue makes, that req/resp are always used > in the SOAP sense and
not the HTTP sense. The SOAP sense should suffice, given that
this TC is not concerned with binding the spec to underlying transports. The
standard way to bind SOAP to HTTP actually leads to the "HTTP sense"
of a req-resp in (1). But is that always what we want? E.g. in case RMS reliably
sends a sequence of messages over HTTP responses for some reason, shouldn't
we allow the CS to be also transmitted over an HTTP response? Jacques From: Marc Goodner
[mailto:mgoodner@microsoft.com] In reviewing some of the issues up for this weeks call
I'll note that the proposal for i082 seems more like a proposed direction
rather than a formal proposal. Will a more formal proposal be available before
the call? I'm concerned with the assumption the proposal in this issue
makes, that req/resp are always used in the SOAP sense and not the HTTP sense.
Has anyone done a thorough review of the spec to check the uses of these terms
to check this? http://docs.oasis-open.org/ws-rx/issues/ReliableMessagingIssues.xml#i082
Marc Goodner Technical Diplomat Microsoft Corporation Tel: (425) 703-1903 Blog: http://spaces.msn.com/members/mrgoodner/
|
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]