Re: Initial formulation of intermediary semantics for MTOM

Jean-Jacques Moreau writes:

>> This looks good to me. 

Thank you!

>> Maybe one additional point would be to cover active intermediaries, 
>> i.e. can it optimize someelse's header block?

Hmm, good question.  I think bindings can optimize anything the like per 
the SOAP recommendation.  The requirement is to reconstruct the Infoset, 
hop-by-hop.  So, I'm not sure we need to or should say anything, except 
maybe to observe that it is possible.  I don't think the result is an 
active intermediary in the SOAP sense, since the Infoset is not changed by 
the optimization.   MTOM is just one convention for doing some particular 
optimizations, primarily under the control of the sending node.  The 
proposed text merely points out that certain implementations MAY conspire 
to use implementation artifacts from the receiving code to facilitate the 
work of sending code at an intermediary.  Otherwise, I think we have the 
usual SOAP rules.    So, I think the bindings can do this, but not as 
active intermediaries.  I think the permission is implicit in the SOAP 
spec, as opposed to MTOM. 

------------------------------------------------------------------
Noah Mendelsohn                              Voice: 1-617-693-4036
IBM Corporation                                Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------

Received on Tuesday, 16 September 2003 17:33:36 UTC