Re: GEPs

[Luis Villa <louie ximian com>]

On Tue, 2003-03-04 at 11:42, Havoc Pennington wrote:
> On Tue, Mar 04, 2003 at 08:35:19PM +1100, Jeff Waugh wrote: 
> > Out of interest: Have we done anything since the GEP process 'disappeared'
> > that should have gone through it?
> 
> There are zillions of things, really. List all the features in 2.2, or
> all the planned ones for 2.4, or for GTK 2.4.

Features for all of GNOME just can't work in a GEP, I don't think, not
unless a handful of people commit to implementing them across all
modules and the GEP serves mainly to codify what those people will be
doing to other people's modules. Using GEP for a module list, or for
specific features of a specific module, would work better. This reflects
the problem with the GEPs- there was no endgame, no plan to figure out
when they were final, and no plan to make them stick. So the ones that
worked were the ones that could be followed through on by just a handful
of people. 

Quick overview, from where I stand:
GEP1: orbit cpp- worked well.
GEP2: theme stuff- was never finished because the principal people
involved could never quite agree on things, IIRC. So the one writing the
code (jrb) basically got to 'win.'
GEP3: remote activation in bonobo-activation: not sure why this was left
open.
GEP4: versioning: officially still pending. Hasn't really happened
because the release team has not been driven to enforce it. Again,
endgame.
GEP5: object adaptor: didn't follow this closely enough to really say
what happened here.
GEP6: toolbar: seems to have been useful, though maybe more comment from
james would be appropriate here. It's still listed as 'pending', though?
GEP7: seems to have been useful, since Alex has implemented this
everywhere. Still listed as pending, though. 
GEP8: more orbit stuff: discussion period ended four months ago, still
pending. I'm not familiar enough with this to suggest why it didn't
work.

So.. dunno. Hard to draw general conclusions from such a small sample
size. But I'm still pretty convinced of my analysis above- the GEPs were
a decent process for figuring out 'what do we do' but when it comes to
'someone needs to do it' GEPs can't be a substitute for someone sitting
down and writing code. If that second part doesn't happen, the first
part will usually end up not meaning much.

Luis

> We need to figure out why it didn't work in the cases where it didn't,
> such as the Theme control panel or icon themes in libgnomeui, and why
> it did with orbitcpp.
> 
> Havoc
> 
> _______________________________________________
> desktop-devel-list mailing list
> desktop-devel-list gnome org
> http://mail.gnome.org/mailman/listinfo/desktop-devel-list
>