On Jun 18, 2007, at 17:54, Peter Saint-Andre wrote:
Now I'm planning on XEP-0154 (User Profile) and XEP-0084 (User
Avatar). Those protocols are still experimental, though. Is it
wise to implement them right now? Why are they still experimental
after 4 and 3 years of development respectively, are major
changes expected?
No major changes are expected. They have not been high priorities.
I agree with Kevin that it would be good to push the avatar spec to
Draft. I'm less sure about profiles yet (see below).
good :)
Avatar would be good to implement, no question. That's not had
implementations (until now) because previously it was pubsub
(rather than pep), which some people had issues with and never
implemented, and then because it was pep and there were no pep
servers. Now there are both pep servers and clients, expect 84 to
pop up places (like Psi).
Feedback would be appreciated from developers who have implemented
this spec.
I've implemented it. It works pretty well, and I haven't had any
issues, except that specifying the width and height isn't that easy
for a client (I actually had write a parser for png headers for
this). Maybe these two could be made optional?
One nice feature I've dreamed up while writing it is the possibility
to have 3D models instead of images as a secondary representation
(that would be especially helpful for MMORPGs). That'd work fine with
the spec, but width and height are undefined for this kind of
representation.
154's trickier. It exists because people have issue with vcard-
temp, but despite this most people are happy with it and continue
to use it.
Agreed. I have not seen a great deal of demand for this since
people seem happy enough with vcard-temp. But vcard-temp has a lot
of problems in theory and has been "temp" since 1999. Maybe it's
time for something better. :)
Usually, clients don't need instant notification for changes to the
profile, that's why it's not really absolutely necessary to use PEP
there. Of course, it'd simplify server implementations, so it's
definitely something to look into.
andy
