- From: <noah_mendelsohn@us.ibm.com>
- Date: Wed, 18 Jan 2006 17:37:23 -0500
- To: Rich Salz <rsalz@datapower.com>
- Cc: "xml-dist-app@w3.org" <xml-dist-app@w3.org>
Rich Salz writes:
> I'm inclined to agree with this,
You are a man of great wisdom and judgement.
> I just don't know how the intermediary knows the MEP. Short of
> making SOAP (or at least non-default MEPs) depend
> on WSDL, then the messages themselves must be tagged, right?
Not necessarily in the envelope or nonstandard headers though. I think I
answered this for HTTP. The HTTP binding makes clear that the WebMethods
used for the SOAP binding are POST and GET, and that the MEP can be
inferred from these. I would expect other bindings to use similar
techniques when applicable.
I would agree that we haven't clearly said whether the MEP must >in all
cases< be determinable from the message, implicit in the endpoint, etc.,
but we've certainly shown that it commonly can be. Furthermore, I'd make
the case that at worst the intermediary is no worse off than the ultimate
receiver. Surely the latter must know what its options are for faulting,
responding etc., and I would start from the position that the intermediary
is in this limited respect acting as a proxy for the ultimate receiver.
So, I agree there are some important edge cases to discuss someday, but I
think the main path cases are already clear. Many bindings will support
only one MEP, in which case you know statically what the MEP is. The next
most common case is the HTTP-like case, where the binding has a natural
way of signalling the MEP on the underlying protocol. In both those
cases, the intermediary knows either statically or from the message. I
think the remaining cases are interesting but lower priority.
--------------------------------------
Noah Mendelsohn
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------
Rich Salz <rsalz@datapower.com>
01/18/2006 04:23 PM
To: "noah_mendelsohn@us.ibm.com" <noah_mendelsohn@us.ibm.com>
cc: "xml-dist-app@w3.org" <xml-dist-app@w3.org>
Subject: Re: Action item - Part 2: SOAP request-response,
response, request-optional-response ...
> Indeed, I'd say that the intermediary is responsible
> for ensuring that the 2nd hop binding can faithfully implment the MEP
used
> by the first hop.
I'm inclined to agree with this, I just don't know how the intermediary
knows the MEP. Short of making SOAP (or at least non-default MEPs) depend
on WSDL, then the messages themselves must be tagged, right?
/r$
--
SOA Appliance Group
IBM Application Integration Middleware
* This address is going away; please use rsalz@us.ibm.com *
Received on Wednesday, 18 January 2006 22:37:32 UTC