[Standards-JIG] Proposed Changes to JEP-60 - PubSub

Timothy Carpenter timbeau_hk at yahoo.co.uk
Fri Jun 4 09:42:08 UTC 2004

On 4/6/04 9:12 am, "Ralph Meijer" <jabber.org at ralphm.ik.nu> wrote:
>> 1) hierarchy separator
> You say you think nodes ids should always be hierarchical. Why is that?

A hierarchy with a single level is non-hierarchical :-). Hierarchies are
really useful and used in many pub-sub systems. The TIB generally uses a 4
tier hierarchy which allows quick simple subscriptions to broad services
without having to define and subscribe to every single node or define nodes
to represent such sets. If you do not want a hierarchy, leave it at a single

>If you were to store avatars like
> for example 'avatars/ralphm at ik.nu', etc, you have to make sure that only
> ralphm at ik.nu can create that node. Also, when the owner's jid doesn't match
> the jid part of the node id anymore (transfer of control), you have a problem.

This is another topic - permissions.

(as an aside, I would say that it is not "avatars/ralphm at ik.nu", but
"ralphm at ik.nu/avatar"!).

>> 5) specifying itemID & nodeID update behavior
> I think the above at 2), applies here as well.

I am rather concerned to see the ideas on updates - that an update fails on
a republish (urrr - why do you need this?) and the monotonical numbering is
kinda abdication by the publisher, relying on the infrastructure to name and
manage one's entites. The first question to me is how can you then
overwrite/rebuild/update/destroy an appending node iteration? How many have
you spawned into the ether? What happens if you want to delete the nth
iteration - do others shuffle up and down or do we have a gap? This is not
being pedantic, as such lists are used and managing them in a pub/sub
environment is a real issue. If you know how many you have spawned at the
client, then why not just manage the name in the client?

Better to have "generic/pgm-mp3-player-1" created (if "-" is the
hierarchical separator) so at least you can get all of them or delete them
using wildcards...hierarchical again, I know.


More information about the Standards mailing list