Re: Feature Proposal Guidelines? / Schedule

[Matthias Clasen <matthias clasen gmail com>]

On Sun, Apr 17, 2011 at 7:14 PM, Andre Klapper <ak-47 gmx net> wrote:
> On Thu, 2010-10-07 at 16:11 +0200, Andre Klapper wrote:
>> I've dumped a first 3.1 schedule draft at
>> http://live.gnome.org/AndreKlapper/Sched1
>> as distros prefer to have this a bit earlier IIRC.
>
> I've moved the schedule to
> https://live.gnome.org/ThreePointOne#Schedule , however for the "Feature
> Proposal and Discussion Period" we first need some guidelines on a
> wikipage instead of linking directly to the 3.1 Planning page IMO, like
>
> * Who can add features there?
>  => anybody proposing a platform-wide feature (not on an app level)
>     with a (likely) assignee (somebody allocating time to work on it).

'Platform-level' sounds approximately right - if it involves changes
to the shell or new control-center panels, it is likely something that
we want to track as a feature. If it is 'just' a new app or a new
feature in an existing app, it is still a good idea to plan it, get
designers involved, let marketing people know, etc. But that can
happen independently.

> * Does that imply a feature freeze after 3.1.1?

'Freeze' somewhat implies  bureaucracy and having to ask permission to
do things. Not sure we really need that. What we really want is to
have a fairly clear idea at the end of the planning phase what major
features we can expect in the next stable release of our core modules.
I don't think having some 'unannounced' features appear later is a bad
thing, so no to 'freeze' that out.

>  => we refer to planning of platform wide features.
>     This is not about features within a module.
>     We'd like to encourage people to improve planning and
>     elaborate future plans for the GNOME platform by this.

> * How does this play well with the "old" module proposals system?
>  => expecting that adding new modules mostly happens because
>     of adding new features for the GNOME platform, this is now
>     part of it.

A good example here is zeitgeist, which will probably be pulled in by
the 'finding & reminding' feature. And libsocialweb, which will likely
be pulled in by the 'web accounts' feature. The main point is to turn
things around so we design the user experience we want, and then
figure out what technologies we can use to implement it.

>     No changes to the current system of bumping dependencies.
>
> Further comments etc or clarifications appreciated.
>
> After getting a basic version of this out (quickly...) I'd announce the
> draft *and* the feature proposal and discussion period.

Sounds good to me.