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

Jens Mikkelsen gyldenskjold at mail.dk
Tue Oct 26 20:13:21 UTC 2004


First of all, sorry if I haven't made my self clear.

On Tue, 2004-10-26 at 21:41, Rachel Blackman wrote:
> > Further:
> > The groupchat is described like an IRC and this is allso what the
> > rfc3921 says.
> >
> > Now I will not implement an IRC groupchat where you can stay anonymous.
> 
> groupchat = old and busted
> MU-C = new hotness
> 
> Seriously, if you want to do anything more than really basic groupchat, 
> work with MU-C for your server 'group chat' component and just don't 
> provide the 'anonymous' feature, and make all your rooms 
> non-anonymous[1].  You will notice that non-anonymous rooms include a 
> 'jid' in the MU-C <item/> portion of the presence.  This jid is the 
> person's real JID (or at least, what JID they're coming from).

After your first reply, I decided to look into MU-C, as it seemed to be
what I needed. I wasn't aware that a new standard were created since the
book was published(2002), hence the question. So I will look into that.

> 
> Given that you're writing a Jabber client specifically to implement 
> some customized encryption stuff, you don't need to worry about older 
> clients not supporting MU-C instead of groupchat; those same clients 
> probably wouldn't support your custom encryption stuff anyway, right?  
> So given that you have the jid in the <item/> already, you should be 
> able to use the digital signatures just fine.

Your right. And that was what I were aiming at. I haven't had the time
to look into the JEP yet, but it seems great! 

> 
> Beyond that, I don't really understand your comment (snipped from what 
> I quoted) that groupchat uses a 'roster.'  It uses a roster in the 
> generic sense ('a list of people') but not a roster in the actual XMPP 
> sense ('jabber:iq:roster' containing JID, nickname and group).
> 

Hmm... I don't know if I have misinterpreted something here.
But that was the problem. The roster needs to be a list of the actual
jids. Which is possible in MU-C, right?

> Maybe if you could explain better to folks what you're trying to do?  
> It seems like you want:
> 
> * A group-chat (i.e. where you can join and send out one message, it 
> goes to many people, you get replies back)
> * With real JIDs (like MU-C in non-anonymous mode)
> * And encryption or digital signing (sort of like the OpenPGP presence 
> and message signing)
> 
> ...yet you seem to feel like the server cannot handle the groupchat.  
> If you want a truly distributed groupchat where no server is involved 
> after the initial exchange, then I'm afraid you're going to need to 
> define your own JEP for 'distributed decentralized groupchat' or 
> something.  But keep in mind that then, instead of sending /one/ 
> message to a room and everyone getting it, the onus of sending a 
> message to everyone in the virtual 'chat room' will be on each person's 
> client, rather than on the server.  (Unless you're supporting extended 
> stanza addressing[2], which I'm not certain any servers yet support.)
> 

Again I'm sorry if I didn't made myself clear. Your right in your three
points. 
What I meant was that I was looking for either a standard that uses the
real jids OR a standard that wasn't administrated by the server.
If the real JIDS are used the server can handle the group and forward
the messages. When I discussed this it was because I understood someones
mail, as if it was not possible to use the real JIDs.

> I hope that helps.
> 

Yes it really does. I will go with the MU-C. Thanks for your kind
assistance.

> --
> 1: http://www.jabber.org/jeps/jep-0045.html#enter-nonanon
> 2: http://www.jabber.org/jeps/jep-0033.html
-- 
Jens Mikkelsen <gyldenskjold at mail.dk>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://mail.jabber.org/pipermail/standards/attachments/20041026/bdc4d838/attachment.sig>


More information about the Standards mailing list