[standards-jig] UPDATED: Pubsub (JEP-0060)

Peter Millard me at pgmillard.com
Wed Jan 22 16:33:43 UTC 2003


Ralph Meijer wrote:
> The header of Example 13 is wrong.
Fixed, thanx.

> The JEP suggest the possibility to be notified of deleted items. How
> would this take form? Do they get a <message/> with a child in the
> 'http://jabber.org/protocol/pubsub#delete' namespace?
> Of course this is just an example. How would applications have to
> handle extensions like this?

I've added examples for this. I've also added node-deletion-notification. These
changes will show up in the next rev.

> Although it is not explicitely in the JEP, Table 1 suggests that for
> persistent items, the publisher should be remembered to be able to
> verify if that publisher is allowed to delete an item.

Yes. IMO, this is an implementation detail that should be derived if
implementations want to support persistent items... I'll look at the wording and
maybe make a note into the Publish section.

> That said, how would this authorization work when different publishers
> have published to the same node id AND item id, and those items
> are all accumulated in a list (example 3 in 7.3)? Who can delete
> the whole item?

I added a note saying that items should never accumlate in a list.. The current
wording is:
<blockquote>
If item identifiers are used, they MUST be treated as unique within the scope of
the node. Node-id + item-id MUST be unique within a given system and specify a
single published item to a single node. Similar semantics for multiple publishes
to the same item-id MAY have similar semantics to those stated above for nodes
(with the exception of having multiple items in a list with the same item-id as
item-ids MUST be unique within the scope of a node).
</blockquote>

> Does a 'purge' induce item deletion notifications being send?

Yes, I've added a note about this.

> Reminder: subscription configuration.

Subscriber options have been added.

Thanx for the comments.

pgm.




More information about the Standards mailing list