W3C home > Mailing lists > Public > xml-dist-app@w3.org > March 2002

RE: Draft Resolution for Issue 41

From: <noah_mendelsohn@us.ibm.com>
Date: Tue, 12 Mar 2002 17:48:41 -0500
To: "Appleton, Pete M" <PMAppleton@bemis.com>
Cc: amr.f.yassin@philips.com, "'Jacek Kopecky'" <jacek@systinet.com>, "'Williams, Stuart'" <skw@hplb.hpl.hp.com>, xml-dist-app@w3.org, xml-dist-app-request@w3.org
Message-ID: <OF246F7673.D2B10DE7-ON85256B7A.007DC04D@lotus.com>
I think we should go a bit slowly here, and build on our experience.  The 
one binding we've seen widely deployed does not employ a destination 
address encoded in the envelope.  This has nothing to do with 
intermediaries, SOAP or otherwise.  The natural way to send an XML 
document using HTTP POST is not to put the destination URI in the 
envelope;  you pass the URI along with the envelope.  HTTP is not unusual 
in that regard.  Furthermore, it's quite reasonable to assume that in many 
realistic scenarios, the envelope will be prepared by one level of 
software, but routed by separate pieces of middleware.  For example, it's 
common that high volume implementations have load sharing and failover 
sites spread around the world.  Requiring the envelope to be cracked open 
so that the message can be re-addressed is a nuissance, and doesn't solve 
any problems.

We've already seen proposals like RP which provide very flexible means of 
routing messages using information carried in the envelope.  These suggest 
that we should not standardize any means of addressing until we understand 
routing, and especially dynamic routing.  SOAP 1.2 as proposed allows the 
binding specification to determine the means used to deliver a request or 
response.  Not that, in the case of a response, it is typical that the 
responding application (that builds the application) does not know the URI 
or IP address of the requestor.  In practice, connections (often 
abstracted as sockets) remain open during request processing, and are used 
for response routing.  Surely we don't want to require a destination 
address in the envelope for the response.

The point is that there are all kinds of ways to get this wrong.  SOAP 1.2 
provides flexibility in the right places, while remaining simple and 
extensible.   I think it's fine for this round.

------------------------------------------------------------------
Noah Mendelsohn                              Voice: 1-617-693-4036
IBM Corporation                                Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------
Received on Tuesday, 12 March 2002 18:05:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:09 GMT