[Standards] publish+configure again
remko at el-tramo.be
Sat Mar 31 07:46:03 UTC 2007
> I understand your frustration. We can't go around and around on these
> issues forever. I think we're having a good discussion here and that a
> bit more patience will result in a clear understanding of the costs and
> benefits of publish+configure, which we can document for all time in
> XEP-0163 version 1.1.
Sure, I don't mind people discussing this. I'm just saying that I'm
leaving the rest of the discussion for the council, I don't think
there's much more I can add.
Maybe I just don't like the PubSub model of nodes and items for PEP,
and I would like to have a more natural thing like saying 'publish
this item publicly, publish this item privately'. It would make a lot
more sense from the extended presence perspective. The only problem
with that is that the 'last published item' is hard to track (because
of different rules applying to different contacts), although that's
not impossible to solve (it just introduces extra space requirements).
I guess the major problem is that people like PubSub too much to
create something specialized that does exactly what the PEP-people
want it to do. A specialized PEP protocol would allow us to get what
we want, without conflicting with the existing PubSub standard (which
is destined for greater things than IM), and would also lower the
threshold for a server to implement PEP without having to implement
the full PubSub standard.
But anyway, I'm happy with PEP over PubSub as it is now, just as long
as we finally get out of this handicap of not having extended presence
in Jabber ;-)
More information about the Standards