- From: Chris Lilley <chris@w3.org>
- Date: Tue, 1 Apr 2003 21:01:56 +0200
- To: Noah Mendelsohn/Cambridge/IBM <noah_mendelsohn@us.ibm.com>
- CC: (wrong string) é <fabrice.desre@francetelecom.com>, "Henry S. Thompson" <ht@cogsci.ed.ac.uk>, <www-tag@w3.org>
On Tuesday, April 1, 2003, 8:44:11 PM, Noah wrote: NMCI> I >>think< this means that XMPP has chosen a usage convention NMCI> for XML that's pretty much identical to the one chosen by SOAP. I think so too - no internal or external DTD subsets, no PI, no comments. Interesting that they are both protocol-level usages of XML. NMCI> XMPP may have been influenced by SOAP, but I have no reason to NMCI> believe the reverse is true. Perhaps we came to the same NMCI> conclusion independently? That would be interesting. NMCI> (I still prefer the term "usage convention" to "subset" insofar NMCI> as I am not aware of any intention on the part of Soap designers NMCI> or XMLP to define any sort of general purpose subset of NMCI> XML...the intention was to never send soap messages that use XML NMCI> features inappropriate to our needs, and therefore to allow the NMCI> opportunity for SOAP receivers to optimize accordingly should NMCI> they choose to do so. Whether or not to use a general purpose NMCI> SOAP parser at a SOAP receiver is an implementation choice at NMCI> that receiver. Yes, an implementation choosing to use such a NMCI> general purpose parser must ensure that no proscribed NMCI> constructions have been used.) This sounds fine until the conformant behavior on getting a message with a PI, a comment, a DOCTYPE is defined, and put in the test suite. -- Chris mailto:chris@w3.org
Received on Tuesday, 1 April 2003 14:02:34 UTC