[Standards-JIG] new top level tag?

Chris Mullins cmullins at winfessor.com
Tue Mar 9 18:44:03 UTC 2004


Ulrich Staudinger [us at die-horde.de] Wrote:

>when working on RTP XMPP integration i suddenly had the sudden flash of

>an idea: a new top level tag.
>
><rtp to='rtprelay.com'>
><request>
><session type='private' id='abcd'></session>
></request>
></rtp>
>

I love it. 

The rules for processing the <rtp/> stanza's can be much more
appropriate than the <iq/> processing rules, which may not be at all
appropriate. 

My only concern would be that the processing rules that you come up with
be significantly different than those of IQ - if the processing rules
end up being the same, then it should be an IQ packet. If the rules are
different, and the justifications for the rules are good, then it
belongs in it's own namespace. 

Questions would quickly be: Does every "request" need a "response"? IQ
demands this behavior, but for RTP it may not be appropriate. 

Could RTP stanzas be sent to: 1) "server.com", 2) "user at server.com", 3)
"user at server.com/resource"? IQ demands all 3 be legal, but RTP may not. 

What behavior is expected in the offline cases? 

And so on....


XMPP Core states:
"Three kinds of XML stanza are defined for the 'jabber:client' and
'jabber:server' namespaces: <message/>, <presence/>, and <iq/>." We
might be better off defining a new namespace (ala, SASL, TLS, Dialback),
and then putting the rtp stuff in there. 


-- 
Chris Mullins



More information about the Standards mailing list