Re: Release coordination - a clarification

[Havoc Pennington <hp redhat com>]

Seth Aaron Nickell <snickell Stanford EDU> writes:
> Maintainers cannot be clearly designated for every decision, and I don't 
> think they should be. Decisions that are really about GNOME as a whole 
> can't clearly be passed to two or three maintainers. 

But I think they'd be fine to pass to 5 or 7 maintainers willing to
spend a lot of time on a particular issue. e.g. something like the
list I suggested for gconf/bonobo-config. And of course with public
comment/discussion as those 5 or 7 made up their mind, required by an
RFP process.

> I don't think we ever got to the point in the gconf/bonobo-config dispute, 
> for example, where voting would have been desirable. BUT, I think it would 
> have been worth having a "show of hands" regarding various solutions to 
> the duplication of configuration systems. Show of hands is a very 
> effective technique for trying to reach consensus in a group of live 
> people, and I think it would be worth exploring how we might reach a 
> similar effect using electronic means.

I think you'd just end up with a popularity contest or something at
that point.

I remember strongly the switch from Imlib to gdk-pixbuf - the general
opinion for a year or so was "Imlib rulez," "greatest thing ever,"
etc., then some prominent hackers switched directions and suddenly
everyone's view was "Imlib is crap," etc. But I don't think most
people ever even looked at Imlib or could articulate what was wrong
with it, it was just a follow-the-leader kind of phenomenon.

Plus I think the general tendency in free software is toward severe
featuritis, which usually dooms the long-term maintainability of
software (not to mention usability) - so I think one key role of
maintainers is to keep that under control, and this is a reason you
don't want an opinion poll deciding your direction - no one ever votes
"no" for any feature, ever. It just doesn't happen. 

I can probably count on one hand the number of times someone has
lobbied to have _fewer_ features in anything. ;-) (Yeah lots of people
want "less bloat" or "fewer bugs" or something, but they can never
name the feature they'd remove to get that "less bloat" feature ;-)

Anyhow, I really think a few good people spending lots of thought can
do better on an issue than a vote, especially if there's a public
discussion everyone can participate in before the few people decide.

Havoc