[Standards] disco-publish

Joe Hildebrand hildjj at gmail.com
Fri Apr 18 14:22:29 UTC 2008

On 4/18/08 3:47 AM, "Ralph Meijer" <jabber.org at ralphm.ik.nu> wrote:

> Hmm, considering that I find the node namespace to be shared among disco
> and pubsub, I didn't expect this particular example. For example
> considering the example from XEP-0030:
>         <iq from='kinglear at shakespeare.lit/throne'
>             id='publish1'
>             type='set'>
>           <query xmlns='http://jabber.org/protocol/disco#items'
>                  node='jabber:iq:kids'>
>             <item action='update'
>                   jid='cordelia at shakespeare.lit'
>                   name='Cordelia'/>
>             <item action='update'
>                   jid='goneril at shakespeare.lit'
>                   name='Goneril'/>
>             <item action='update'
>                   jid='regan at shakespeare.lit'
>                   name='Regan'/>
>           </query>
>         </iq>
> This should result in the list of three items to show up when sending a
> disco info query to the node 'jabber:iq:kids'. So I was expecting that
> one would publish those items to the 'jabber:iq:kids' node.

I think you're right here.  The main use case for disco publish was the
now-superceded XEP-119; pointers to pubsub nodes for extended presence.  As
such, it may be that the entire concept of disco publish is no longer
required; it's all just straight PEP.

> I also want to point out again that automatic subscriptions through Caps
> are currently defined such that when you advertise to support a
> particular namespace+notify, you will get notifications of all items
> that carry payload in that namespace, *whatever the NodeID*.

Whoa.  Guess I should have read the XEP-60 changes more carefully.  That was
not the original intent.  It may be that it is better, however.  For
example, it solves the "I publish two blogs" problem, at the expense of two
different clients not necessarily knowing which node has which semantic

More information about the Standards mailing list