[Standards-JIG] proto-JEP: Smart Presence Distribution

Carlo v. Loesch CvL at mail.symlynX.com
Tue May 30 09:38:24 UTC 2006


Dave Cridland typeth:
| On Wed May 24 08:53:44 2006, Carlo v. Loesch wrote:
| > The new proto-JEP has been updated on
| > 	http://www.jabber.org/jeps/inbox/smartpresence.html
| 
| A few comments. Basically, I think this is improving, and I think 
| that in terms of security/privacy, I think this is fine.
| 
| I'm uncomfortable about the implicit address-list setup. I'm 
| concerned that it may prove quite easy to inadvertantly confuse 
| things, and I suspect that using a different method for handling the 
| setup becomes more reusable.

Reusable: In fact our plan is to re-use the concept & implementation of the
context list and only adapt the messages that modify it, since we don't think
it's nice to turn around the whole protocol.

Our pre-proto-JEP on Smart Unicast for MUC uses the same context
list for MUCs. The server implementation would re-use the techniques
also for MUC, the difference is only that the presence messages used
by MUC to determine membership are used to populate the lists.

I don't know if it makes more sense if we also submitted the 'smartchat'
pre-JEP so you can see how they work together. We just thought having
them in seperate JEPs gives the developers a chance to go step by step
and enjoy the intermediate results, too.

| In particular, I'm thinking about sending presence to a series of 
| people, plus a directed presence, might confuse things, and I'm 
| merely suspicious of multiple resources getting tangled. I just think 
| it might be slightly too tricky to get things right.

We need a way to distinguish directed presence from regular presence,
as we certainly don't want directed presence to mess up our data
structures. One way would be, if regular presence were always aimed
at the bare JIDs whereas directed presence were sent to a specific
resource. Alternatively, the subscription state gives us some
indications (according to XMPP IM 5.1.4) whether it is directed presence,
or not.

| (Oh, I agree that this is a Disco feature, but I also agree that we 
| ought to work on hop-to-hop Disco with a minimum of round-trips - 
| ideally none.)

That's a different mental-pre-proto-JEP, right?  :-)

-- 
» Carlo v. Loesch » http://symlynX.com » psyc://ve.symlynX.com/~lynX
	    xmpp:lynX at ve.symlynX.com » irc://ve.symlynX.com/#symlynX
        CryptoChat » https://ve.symlynX.com:34443/LynX/?room=symlynX



More information about the Standards mailing list