[Standards] Publish-subscribe collection node deassociation notification

Ralph Meijer jabber.org at ralphm.ik.nu
Thu Feb 7 15:08:38 UTC 2008


On Mon, 2008-02-04 at 10:26 -0700, Peter Saint-Andre wrote:
> Ralph Meijer wrote:
> > On Wed, 2008-01-30 at 17:13 -0700, Peter Saint-Andre wrote:
> >> Ralph Meijer wrote:
> >>> Hi,
> >>>
> >>> I am working on getting publish-subscribe collection nodes implemented
> >>> in Idavoll. There are two ways to subscribe to a collection nodes using
> >>> the 'pubsub#subscription_type' subscription configuration option: nodes
> >>> and items.
> >>>
> >>> In the 'nodes' case, you get notifications when nodes get associated
> >>> with the collection. However, we didn't define protocol for notifying
> >>> subscribers of deassociation, and I would really like to have that.
> >>>
> >>> To propose new protocol for that, I'd like to know if there is anyone
> >>> actually implementing this bit of the spec.
> >> Ralph, it seems that you're the only one. :)
> > 
> > Heh. That's nice!
> > 
> >> Notification of a new node takes the following form:
> >>
> >> <message from='pubsub.shakespeare.lit'
> >>          to='francisco at denmark.lit'
> >>          id='newnode1'>
> >>   <event xmlns='http://jabber.org/protocol/pubsub#event'>
> >>     <collection>
> >>       <node id='new-node-id'>
> >>     </collection>
> >>   </event>
> >> </message>
> >>
> >> I suppose that notification of node disassociation might look like this:
> >>
> >> <message from='pubsub.shakespeare.lit'
> >>          to='francisco at denmark.lit'
> >>          id='oldnode1'>
> >>   <event xmlns='http://jabber.org/protocol/pubsub#event'>
> >>     <collection>
> >>       <disassociate id='old-node-id'>
> >>     </collection>
> >>   </event>
> >> </message>
> >>
> >> What do you think?
> > 
> > Sure that would work. The only thing that I noticed is that the use of
> > the 'id' attribute is not consistent with the use of the 'node'
> > attribute everywhere else in this spec.
> > 
> > If there are no implementers, I would suggest having
> > <associate node='NodeID'/> and <disassociate node='NodeID'/>.
> 
> Yes I like that better.

Good, we seem to have list consensus :-)

-- 
Groetjes,

ralphm




More information about the Standards mailing list