temas at box5.net
Tue May 6 16:49:44 UTC 2003
I can appreciate that client authors want to do things for there users,
but I think we need to spell out the "correct" process by which you can
use an experimental JEP. What it will probably boil down to is, using a
namespace significant to your client like: gabber:x:avatar or
http://gabber.jabberstudio.org/protocol/avatar, and a willingness to
move to the standardized versions when they appear.
As to why the JEP was published, that's the correct process for a JEP.
You publish a version you want discussed, it is labelled appropiately
(experimental), and discussion begins. At a certain point it is final
called, and sent to the council. Then it moves on from there. I don't
see anything wrong with how the JEP was treated. Heck, back then it was
even faster to get voted on, and I'm pretty sure the JEP was voted away
rather quickly after introduction. So, it probably boils down to a
labelling issue, and the helpful hints I mentioned above for
About creating a new JEP for the "new" avatar work, I still don't see
why that's necessary. Most of the information is still valid or needs
minor touch ups. Then we add in the pubsub explanations. Plus, the old
revisions are all maintained by the CVS. We aren't losing it, we are
making a new revision of the exact JEP.
On Tue, 2003-05-06 at 11:24, Rachel Blackman wrote:
> > I see the 'but I have implemented it already' argument frequently. On
> > the one hand we have people who say: "if it's experimental and you
> > implement: sorry for you". On the other hand we have client developers
> > who feel ignored because their implementation doesn't count. There are
> > also some political issues which I don't want to get into.
> > If the client developers wouldn't have used the jabber:x:avatar
> > namespace in their implementations, it would be no problem to totally
> > rewrite this JEP. They could have used 'myclientname:x:avatar' for as
> > long as the JEP was not 'draft'. If the JEP would then be totally
> > rewritten, at least their implementation doesn't break down.
> To play devil's advocate for a moment here, look at it from the other side.
> You have client authors who, in many cases, are dealing with people coming
> from the AIM, Yahoo, etc. world. While programs like Exodus are quite
> technically sound, if I tried to get my MSN-using parents onto Jabber
> they'd not be able to wrap their heads around it. They want something that
> looks like MSN or AIM. They'd do much better with, say, Rival Messenger.
> And avatars are a big feature users always ask for; as soon as one client
> implements them, users hear about it and begin asking for it. I know my
> choice not to support sending avatars using JEP-0008 in Trillian is not
> going to be a popular one; I'm already expecting lots of unhappy 'but
> JAJC/Rival/GAIM/whatever does avatars!' comments, and I don't honestly
> think 'it's a deferred and deprecated experimental standard, there'll be a
> new avatar system coming soon' is going to entirely cut it with the users.
> I'll cope, but... :)
> The avatar JEP has been available for close to two years, having been
> initially put up September 14, 2001. That is two years in which time
> client authors looking to get people from the legacy IM networks onto
> Jabber have been trying to implement features; avatars, formatted text and
> file transfer are the three big ones AIM sorts will be asking for.
> To finish the devil's advocate soapbox... if the avatar JEP was never
> intended to last or be implemented in that form, and if it's considered
> 'bad' to implement the experimental protocol extensions... then why was the
> avatar JEP published and left out there for two years in the first place? :)
> Experimental JEPs are useful because they provide discussion on
> implementation. However, leaving a JEP for a widely-desirable feature in
> 'experimental' for two years and then yanking it and saying 'well, they
> should have known better than to implement it'... it's not going to be a
> popular move to begin with.
> Just my $0.02. :)
> > Somehow, we need to make it even more clear to developers to not put code
> > based on experimental JEPs in software meant for widespread use. I can
> > imagine a big red warning at the top of JEPs. Maybe it is a good idea to
> > have early implementers use a different namespace as long as the JEP is
> > not in, say, Call for Experience phase, or even draft.
> The differing namespace almost makes sense. Call it 'jabber:jep:<jep>'
> while it's experimental; when the protocol is finalized, they move it to
> the normal namespace. :)
More information about the Standards