Issue #2

On Fri, Sep 27, 2002 at 12:01:16PM -0700, Ugo Corda wrote:
> >"ultimate" is a deceptive word in the gateway case.  I believe it refers
> >to the the ultimate recipient as specified by the initial sender.  This
> >is consistent with SOAP 1.2's use of the word, AFAICT.  It also relates
> >to issue #2;
> But issue #2 is exactly about the fact that SOAP 1.2 does not provide any
> explicit mechanism for specifying the ultimate recipient. (Of course I can
> always define a SOAP extension that provides that, but it does not seem to
> be part of the basic spec).

Right.  Hence my comment about a SOAP envelope not necessarily being
a message (not a SOAP message, just a message in general), since a
message should have meaning, and SOAP envelopes don't by themselves,
necessarily.  They do when you use them to tunnel, or when sent over
transport protocols, but not when *not* tunneling (aka "chameleon")
through an application protocol; in that latter case, the envelope of
the underlying protocol is part of the message, which is why I don't
believe that the SOAP envelope needs the target URI in it.

Back to the original issue;

I think it's quite possible to build such an intermediary without the
URI in the envelope.  The intermediary just needs to look at the whole
message, and then somehow relate the ultimate destination identifier to
the identifying mechanism of the outbound protocol, which is part of
what you have to do when you undertake to build a gateway between two
application protocols with their own identifier system.  You have to
do that with or without SOAP in the picture.

Mark Baker, CTO, Idokorro Mobile (formerly Planetfred)
Ottawa, Ontario, CANADA.     

Received on Friday, 27 September 2002 15:58:49 UTC