Re: [xmlProfiles-29] xml subsetting in IETF XMPP

Via the archives (sorry, I didn't know this was a public list), I have
been following the conversation regarding the XML restrictions in the 
Extensible Messaging and Presence Protocol (XMPP). As you may know, 
XMPP is the IETF's adaptation of the Jabber protocol, which was 
developed mainly in 1998 and 1999 within the community that formed 
around the open-source Jabber IM server project. As far as I know 
(I became involved in late 1999), the Jabber community did not base 
any of its work on SOAP, so the similarities are a case of independent 
development. BTW, XMPP is perhaps even more similar to BEEP than to 
SOAP -- it is a protocol for streaming XML from point to point and 
therefore is not document-centric in any way.

We definitely do not seek to define an official subset of XML (although
if such a subset is defined by the W3C we would ensure our conformance
with it). Rather, we seek to restrict what XML may be sent over 
XMPP-compliant XML streams. Therefore, from the perspective of the W3C, 
would it be preferable for us to use the phrase "usage convention" rather 
than the term "subset"? (It seems there is not yet consensus on this.) 

Also, I think it would it be appropriate to specify strictly that XMPP
entities MUST NOT generate PIs, comments, DTD subsets, and entity 
references (other than the "big 5" predefined entities), and that this 
restricted data MUST be ignored if received. Therefore I have changed 
the wording in draft-ietf-xmpp-core (at least in CVS) and would be 
happy to post the revised text here for review if desired.

If you have any other questions or concerns about XMPP, let me know.

Thanks.

Peter

--
Peter Saint-Andre
Jabber Software Foundation
http://www.jabber.org/people/stpeter.php

Received on Tuesday, 1 April 2003 18:43:04 UTC