ISSUE: Rewriting of destinations

The current draft states (as per the recently approved changes):

    The scope of a one-way MEP is limited to transmission of (nearly)
    identical messages from one sending node to zero or more receiving
    SOAP node(s); typically, in the case of multiple receivers, the
    messages differ only in their destinations.

I'm not convinced the implied rewriting will happen in real bindings. 
For example, in an email binding I would naturally take the
ImmediateDestination from the To: field(s) (and probably CC: as well,
but that's for binding writers to decide).  If I send an email To:
xml-dist-app@w3.org, the To: field will still be xml-dist-app@w3.org
when the message arrives (I just checked :-).  The /email/ message will
have been re-written, with "resent-*" fields and probably a host of
other stuff, but the /SOAP /message -- meaning the properties we list --
remains the same.

Mentally running through the list of likely candidates for bindings, I
can't think of any that would re-write the SOAP properties of a message
sent.  Unless someone can come up with a compelling counterexample, I
propose removing any mention of this possibility, leaving

    The scope of a one-way MEP is limited to transmission of (nearly)
    identical messages from one sending node to zero or more receiving
    SOAP node(s); typically, in the case of multiple receivers, the
    messages differ only in their destinations.

As a comment, I'd cite this as an example of a complication that never
had to be introduced in the first place, despite a strong but vague
conviction to the contrary.  I'm more and more convinced that much if
not all of the concern about "introducing complexity" and "slippery
slopes" stems from just such convictions.  As far as I can tell, the
pertinent change is from "transmission of a single message" to
"transmission of identical messages", and this is driven by the need to
bring the text in line with the definitions in the SOAP model.  Bringing
the text in line with existing definitions is a good thing, but there's
no change in what's required of a binding.

Received on Thursday, 31 August 2006 18:14:25 UTC