RE: Rewrite of SOAP 1.2 Adjuncts

David Orchard writes:

> I don't think it's legit for a soap stack to "fluff up" an 
> envelope unless specifically licensed. 

I agree, at least if we make reasonable assumptions about how MEPs are 
used, etc.

Basically, I see the SOAP model as:  MEPs tell you who can send SOAP 
messages and who they are delivered to.  It's assumed that such deliveries 
are achieved in the manner that the framework gives you for implementing a 
binding:  I.e. the requirement is to faithfully reconstruct an infoset 
supplied by the sender. (See [1], paragraph 6)

I suppose you could imagine a very strange MEP that would call for 
messages to appear that the sender never sent, but I view that as at best 
unusual and almost surely bad practice.  It is reasonable, I think, for a 
multicast MEP to call for the same message or a variant (e.g. with 
modified destination addresses) to be delivered to more than receiver, 
etc.  Indeed, someday it would be nice to flesh out the MEP story for 
multicast (can you fold responses?), for intermediaries etc.

Nonetheless, I agree with David's answer.  For a binding, SOAP Message In 
=> SOAP Message Out;  No SOAP Message In => No SOAP Message Out.

Noah

[1] http://www.w3.org/TR/soap12-part1/#bindfw

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------

Received on Friday, 13 January 2006 03:30:02 UTC