[Standards-JIG] Re: Groupchat using original jid's

Peter Saint-Andre stpeter at jabber.org
Mon Oct 25 22:44:27 UTC 2004


In article <36070.3911436305$1098732478 at news.gmane.org>,
 Jens Mikkelsen <gyldenskjold at mail.dk> wrote:

> On Mon, 2004-10-25 at 20:54, Rachel Blackman wrote:
> > > Is there a standard for groupchat, where the group isn't administrated
> > > by the server or where the original jids are used?
> > 
> > MU-C in a non-anonymous room includes the real JIDs in the presence 
> > packets.  It is therefore possible to map the room-resource format 
> > nicknames to the real JID.
> 
> Hmmm... Ok. I haven't heard of the MU-C before. I have looked at
> jep-0045. To me it seems, that the server administrates this group,
> hence it could, if compromised, send the wrong JID's to the users.
> Thereby I had to trust the server, to be sure to get the right JIDs,
> right?

You can always run your own server + groupchat service. Jabber/XMPP is 
not one centralized service like MSN or AIM.

> What I was looking for, was something where the client was an
> administrator and the server, was a dummy forwarding packages. Then I
> could digatally sign the messages and the user JIDs. I hope I am making
> my self clear.

The current model in Jabber/XMPP is for the groupchat service to 
function as a dumb reflector but handle the room administration. This is 
in keeping with the Jabber philosophy of keeping clients simple:

http://www.jabber.org/jeps/jep-0134.html#simple

A better way to handle this would be for clients to sign messages and 
presence sent through the room, as Justin mentioned.

/psa




More information about the Standards mailing list