AAAHHHH! Too many JEPS - (was: RE: [standards-jig] JEP-0025)
stpeter at jabber.org
Mon Jan 19 16:09:43 UTC 2004
On Sun, Jan 18, 2004 at 07:31:36PM -0000, Richard Dobson wrote:
> Well then the simple solution to this is to only implement JEP's in draft or
> final, not experimental, it has plain warnings about implementing
> experiemental JEP's on them, I dont see why this is a real problem, if
> things are not being updated and have been in experimental for a long time
> it makes simple sence not to bother implementing them, maybe this needs to
> be explained a bit better in the warnings.
Precisely. No one says you have to implement Experimental JEPs. That's
why we have an approval and standardization process, which results in
protocols being added to the official list:
Most companies I know of don't implement JEPs until they move forward to
Draft or Final, as a matter of company policy.
Now it's true that we're in the long-term process of cleaning up some of
the older protocols and replacing them with modern, improved versions.
The replacement of Agent Information with Service Discovery is a good
example. Is this is a fun process? No. Same goes for the so-far halting
efforts to supersede vCard. I'm slowly working out all the issues there.
There are now 32 JEPs in an Experimental state. As they are no longer
worked on, they will be automatically changed to Deferred. Some JEPs may
re-emerge from Deferred, but I think most will not.
My sense is that the pace of initial JEP publication has slowed quite a
bit. But we're still clearing out the older queue.
Would we use a dependency chart? Sure. I'll add that to my .plan but I
don't see it as critically important for us to make progress.
Jabber Software Foundation
More information about the Standards