[Standards] Divergence between XEP-0048 and XEP-0223
dave at cridland.net
Wed Jun 9 16:42:09 UTC 2010
On Wed Jun 9 16:10:38 2010, Peter Saint-Andre wrote:
> On 6/5/10 9:27 AM, Kevin Smith wrote:
> > On Sat, Jun 5, 2010 at 2:28 AM, Peter Saint-Andre
> <stpeter at stpeter.im> wrote:
> > I don't think new profiles of PubSub are particularly nasty -
> > anything more into XEP-0060 is something to be avoided at all
> > but profiles and extensions seem fine.
> Some new profiles might require new features in XEP-0060 (as, for
> instance, PEP required presence-based delivery and notification
> filtering). But we can fight over which features go in which specs.
I'd like to keep a clear distinction between features - which are
low-level, and need actual code added everywhere - and profiles,
which assemble these features into a useful service.
It's entirely logical - and obvious - that profiles get a XEP all of
their own. It's less obvious that we should avoid defining features
in them, but I think it's important for distinguishing between PEP,
222, and 223 - all make use of the "pubsub-onna-jid" feature in
In all cases, features should be either "part of the core", meaning
that they're expected from any implementation of PubSub - ie, any
implementation of XEP-0060, including those for a specific profile,
like PEP - or else the features are truly optional, perhaps in
support of some profile like PEP, or 22. Still, implementing this
feature alone doesn't mean you support PEP, or 222, or 223.
So it'd be interesting to consider removing various optional features
from XEP-0060, such that XEP-0060 became a pure core, then defining a
"Pubsub onna jid" XEP to hold those features, and (perhaps) so on.
Certainly, it feels like an exercise that's worth thinking about,
even if we decide it's not worth doing.
> For the record, I'm quite open to new profiles of pubsub, starting
> a storage profile so that we can clean up XEPs 222 and 223.
Curiously, the profile we've never defined, yet do use and talk about
a lot, is the case of pubsub-onna-dedicated-domain. I'm pretty sure
everyone will disagree, and say "Of course we have! It's just ..." -
but the details may well differ.
I think it'd be very interesting to define what to expect from a
Dave Cridland - mailto:dave at cridland.net - xmpp:dwd at dave.cridland.net
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade
More information about the Standards