Re: Must understand mustUnderstand proposal

	> Date: Wed, 09 May 2001 11:38:13 +0200
	> From: "Jean-Jacques Moreau" <moreau@crf.canon.fr>
	> To: Doug Davis <dug@us.ibm.com>,
	>         Henrik Frystyk Nielsen <frystyk@microsoft.com>
	> Cc: xml-dist-app@w3.org
	> Subject: Re: Must understand mustUnderstand proposal

	> You are right, I had missed that one. From XMLP/SOAP 4.2.2:

	>      "Omitting the XMLP/SOAP actor attribute indicates that the XMLP/SOAP
	>      header block is targeted at the ultimate XMLP/SOAP receiver."

	> But then I think we have an issue with "anonymous" blocks, ie blocks which are
	> not targeted at any specific intermediary, nor the final destination, but which
	> contain information that can be factored out and be referenced by other blocks;
	> examples: a digital signature, credentials, a photograph. If, by default,
	> untargeted (header) blocks are targeted at the ultimate destination,
	> "anonymous" blocks are out.

	> (In a previous thread, "anonymous" blocks used to be called "untargeted"
	> blocks, I believe.)

	> Henrik, "anonymous blocks" sounds like a candidate for the issues list.

Such "anonymous blocks" are what I had in mind with the "None" actor
value in the AM, section 4.2 [1].  The semantics of "None" is simply
that such a block will not match any actor (sort of the opposite of
'Next'); however, the block may be referenced by many other blocks.

The SOAP technique of targeting such shared blocks at the ultimate
destination to ensure their survival is a bit of a hack semantically
since they are not really destined for some module at the ultimate
destination even though they appear to be.

	> Jean-Jacques.


	> Doug Davis wrote:

	> > >> I think this should be an untargeted header and therefore only processed
	> > >> at the ultimate destination.
	> > >You probably mean a body block, since untargeted header blocks cannot be
	> > >processed unless referenced by some other targeted block.
	> >
	> > By "untargeted header" I mean it is a header without an actor attribute.
	> >
	> > -Dug

--mark

[1] http://www.w3.org/2000/xp/Group/1/04/23/XMLProtocolAbstractModel.html#Sec4.2



Mark A. Jones
AT&T Labs - Research
Shannon Laboratory
Room A201
180 Park Ave.
Florham Park, NJ  07932-0971

email: jones@research.att.com
phone: (973) 360-8326
  fax: (973) 360-8970

Received on Wednesday, 9 May 2001 11:06:39 UTC