Re: message exchange patterns and # of parties

I won't be able to join the mep discussion :-( I have not yet 
recollected all the bits of the discussion.
Anyway, from last f2f, I had the impression that wsdl meps while 
different share and align a lot to soap meps.
IMHO, the sole point of difference is that in today's wsdl mep 
definition, one node is labelled as the service while in soapworld there 
is no service notion.
It would be cool to split the mep description in two layers:
    - one layer that defines the nodes and messages
    - one layer that tells who is the service
For instance, a first layer description (let's call it M) could define
    - a node "Requester" and a node "Responder"
    -two messages "Request" and "Response".
The second layer will then say:
    - mep M1 uses M as a basis and for M1, node "Requester" is the service
    - mep M2 uses M as a basis and for M2, node "Responder" is the service
The first layer description is the same for M1 and M2. M1 and M2 would 
then be referenced through wsdl operations, like   <operation type="M1"/>

Another approach is to keep the first layer description, identify this 
description with a uri to it and define the second layer information 
directly in a wsdl description, using the operation attribute defining 
the mep used. We could have something as simple as :    <operation 
type="M:Requester"/> 
which is equivalent to <operation type="M1"/>

The bonus of this approach is that:
    - in the in/out, out/in, in and out cases, we will define only two 
mep first layer description, while not loosing any functionnality
     - the first mep layer description could be used as an abstract soap 
mep. A concrete soap mep would then add its own special information

What do you think?
    Youenn

David Booth wrote:

>
> At 10:46 AM 9/19/2003 -0400, Amelia A. Lewis wrote:
>
>> . . .
>> Or, in short: I would prefer to replace "broadcast" with "multicast",
>> and otherwise would wait for others to argue that this wording is not
>> okay before trying to do any further editing.
>
>
> Sounds fine to me.
>
>