[standards-jig] Avatars

Richard Dobson richard at dobson-i.net
Tue May 6 18:11:43 UTC 2003


Hi all,

Ill just chime in my 2 pence on this to say that people who go ahead and
implement experimental JEPs have to expect there to be potensially major
changes to the protocol and that they MUST implement those changes either
when an incompatible change is made or other manditory elements are
introduced and to fully implement the finalised protocol, so on this note I
think its fine to change the current JEP, and sorry to say it but anyone who
complains that they have implemented it and that the new proposal will break
their implementation that it is "hard cheese" since that is the risk they
took implementing an experiemental protocol. Plus whether it is a popular
feature is not really relevant since the cost of implementing an
experimental JEP is to continue to keep it compliant with the current
version of the published JEP, and if you dont want the burden of keeping
your implementation up to date then you must not implement it until it goes
draft or final, it seems simple enough to me.

Richard

----- Original Message ----- 
>From: "David 'TheRaven' Chisnall" <theraven at sucs.org>
To: <standards-jig at jabber.org>
Sent: Tuesday, May 06, 2003 6:52 PM
Subject: Re: [standards-jig] Avatars


> Just another minor point:
> If we do manage discourage client developers from implementing
> experimental features, then you have less (well, actually no) testing
> before something becomes a standard, so problems are less likely to be
> worked out.
> Having an experimental namespace allows this.
> On the other hand, we need to use something like disco to allow clients
> to determine which experimental features other clients support, and
> attempt to use them with other clients (e.g. not send Avatar data to a
> mobile client).
>
> Hmm.  That may have been 2 points.  Oh well, there are only 3 sorts of
> mathematicians...
>
> Justin Kirby wrote:




More information about the Standards mailing list