[standards-jig] JEP-0102

Richard Dobson richard at dobson-i.net
Wed Jul 2 10:01:17 UTC 2003


> > Why don't we tell software authors explicitly: Hey, sorry, we don't have
> > a final JEP yet, but here is an experimental one (or just a good idea).
> > It may change later, but if you want to take the risk, it would be great
> > if you implement it now, so we can see if it works.
>
> That used to be the approach.  Look at JEP-0008, however.  Never advanced
> past 'EXPERIMENTAL' but so many people wanted it that they all implemented
> it.  Now we have many clients that do avatars, using this method...and
then
> the JEP was retracted.  There was rather a bit of kerfluffle on the lists
> about that, because not everyone was happy with the decision to retract
> 0008.
>
> As I recall, it was because of the JEP-0008 mess that the Experimental 'Do
> not implement' warning was added and people were discouraged from
> implementing experimental JEPs.

Yup I think that was the reason, people do need to realise that it is risky
implementing an experimental JEP, it may completely change between versions
in an incompatible way or simply be retracted, and to be blut its tuff to
those people who implement but dont want to update their implementation
because it is lots of extra effort (you shouldnt have implemented it in the
first place if you are not committed to keep your implementation up to
date). If you implement an experimental JEP you have the responsibility to
keep your implementation up do date no matter where the protocol goes (IMO
you enter into a virtual "contract" to keep your implementation up to date
when you implement an experimental JEP).

Richard




More information about the Standards mailing list