[Standards] pubsub/pep auto-creation

Ralph Meijer jabber.org at ralphm.ik.nu
Thu Mar 22 17:40:14 UTC 2007


On Thu, 2007-03-22 at 11:13 -0600, Peter Saint-Andre wrote:
> Rachel Blackman wrote:
> >>>> The convincing argument [for publish+configure] should be the fact that
> >>>> [without it] you can never assume on a publish that you are 
> >>>> publishing an
> >>>> item the way you want it, unless you check configuration. And even
> >>>> checking doesn't guarantee you that your item will be published 
> >>>> privately
> >>>> if some other resource is posting public events to it.
> >>> Yes, this argument seems very convincing.
> >>> +1 publish+configure for PEP
> >>
> >> But as I just mentioned in the jdev room, publish+configure doesn't 
> >> give you per-item ACLs, it gives you publish plus (perhaps) toggle the 
> >> node configuration, which is quite different.
> > 
> > I don't think we need ACLs, necessarily.  
> 
> OK maybe I misunderstood. It seemed to me that folks wanted to do 
> "publish item #1 and make it private, but publish item #2 and make it 
> public", all to the same node.
> 
> >> I think we need a MUC meeting about this. :)
> > 
> > Honestly, I don't.  If people want ACLs in PEP, that's an entirely 
> > different discussion; it isn't this one. ;)
> 
> We already have ACLs via node configuration. But we don't have per-item 
> control over that. And that seems fine to me!
> 
> > I'll point out that iChat has evidently given up and DOES just stuff 
> > PEP-based XEPs into their presence stanzas (for instance, User Tune is 
> > embedded right there in at least a few iChat presence stanzas I've seen 
> > lately, perhaps from the Leopard iChat).  
> 
> iChat has been doing that for a long time. Part of the problem is that 
> they use one of the open-source servers for iChat server and everything 
> needs to Just Work [tm]. So until the open-source server codebase they 
> use supports PEP, iChat won't support it either. Therefore....
> 
> > If we don't get PEP up and 
> > running after this long, I'm pretty sure that eventually other client 
> > devs will just throw in the towel and follow iChat's lead on this one.
> 
> Yes yes yes. Can we get consensus and start coding? Rough consensus and 
> running code, after all. And yes sometimes the consensus is rougher than 
> others... :)

Right. So as we seem to agree that PEP should cater simple use cases, I
don't see how node configurations should change potentially every time
there is going to be a new item being published.

What I see in PEP, is that you have user tune, user mood, geoloc, blog
nodes tied to your own account. You specify once that all or a
collection of roster groups will be able to subscribe. Or, if you use
the caps stuff, that those people sending that special caps thing will
be temporarily also be subscribed (if they match that config), and then
you will be happily go publish.

Then, once in a while (weeks, months?) you will adjust your config
because maybe you changed around groups or have slightly changed privacy
desires.

As far as I can see, this whole config changing stuff comes from
persistent storage or things like that, where you want more fine grained
control on your stuff. If you want this, you can still implement a more
featureful publish-subscribe system tied to user accounts. Basically PEP
+more. I'm all for that, by the way. Since I do believe people will
eventually want to control more about things they publish. 

-- 
Groetjes,

ralphm




More information about the Standards mailing list