Re: section 2.2.22 Message Exchange Pattern (MEP)

Mark,

I think I agree. I'm certainly more comfortable with this revised 
definition (below) than with "single use of the 
service".

Thanks,

Christopher Ferris
STSM, Emerging e-business Industry Architecture
email: chrisfer@us.ibm.com
phone: +1 508 234 3624

www-ws-arch-request@w3.org wrote on 07/08/2003 11:32:32 AM:

> 
> Chris,
> 
> I see the MEP as a fundamental construct that relates an initial
> message and its possible response message(s).  In the context of SOAP,
> this is a pattern that gets supported by a binding which tends to view
> things from the perspective of interacting nodes.  Protocol binding
> specs "declare their support for one or more named MEPs".  In the
> context of WSDL, an MEP is viewed from the perspective of the nodes
> themselves.
> 
> The SOAP definition occurs in the context of the SOAP spec with its
> binding framework.  Even in that context, it probably should have been
> tightened up a bit, but it certainly needs some qualification in
> our spec since we are in a more general context.
> 
> I think that David Booth was trying to capture a more WSDL-centric
> view of an MEP with the phrase "a single use of the service".  For
> example, the SOAP Req/Resp MEP looks like an in-out pattern at the
> operation level from the perspective of the ultimateReceiver.
> 
> I personally find the SOAP view of an MEP more coherent.  The problem
> with the WSDL operation perspective is that a single, simple SOAP
> request/response MEP ends up having complementary WSDL operation
> patterns for each endpoint in a peer-to-peer environment -- an out-in
> at the initial sender and an in-out at the ultimateReceiver.  But
> these operations are not formally related to each other in the WSDL
> framework.  The SOAP view provides that coherence.
> 
> 
> What about the following definition:
> 
>  A message exchange pattern is a template for the exchange of messages
>  between agents that arise from a message and its responses, if any.
> 
> Is that any better?
> 
> By the way, I don't think we want to say that choreography isn't 
concerned with
> patterns.  They can be MEPPs -- MEP Patterns!
> 
> Mark Jones
> AT&T
> 
>    To: www-ws-arch@w3.org
>    From: Christopher B Ferris <chrisfer@us.ibm.com>
>    Date: Mon, 7 Jul 2003 16:36:25 -0400
>    Subject: Re: section 2.2.22 Message Exchange Pattern (MEP)
> 
>    Mark,
> 
>    I am still uncertain as to what "a single use of the service" means 
in 
>    this, or any
>    context. A service may have many operations, and fulfilment of the 
service 
>    may
>    require more than one of the operations to be invoked by the client. 
Given 
>    this,
>    where does the SOAP Req/Resp MEP come into play?
> 
>    I believe that WSD WG is mapping MEP at the operation level and hence
>    at the very least, we should be aligning our definition with that 
notion.
> 
>    I agree that it is important to distinguish between an MEP and a 
>    choreographed
>    exchange of messages, one is a pattern (the P in MEP) and the other 
is 
>    not.
> 
>    SOAP1.2[1] defines an MEP as:
> 
>            A Message Exchange Pattern (MEP) is a template that 
establishes a 
>    pattern for the exchange of messages between SOAP nodes. 
> 
>    Why isn't that definition good enough for us?
> 
>    [1] http://www.w3.org/TR/soap12-part1/#soapmep
> 
>    Cheers,
> 
>    Christopher Ferris
>    STSM, Emerging e-business Industry Architecture
>    email: chrisfer@us.ibm.com
>    phone: +1 508 234 3624
>

Received on Tuesday, 8 July 2003 11:52:25 UTC