Re: Modulesets Reorganization

[Tristan Van Berkom <tvb gnome org>]

On Wed, Jun 2, 2010 at 3:14 PM, Murray Cumming <murrayc murrayc com> wrote:
> On Wed, 2010-06-02 at 07:52 -0400, Matthias Clasen wrote:
>> ...total nightmare...
>> ...crush all these efforts...
>>
>> I understand that change is uncomfortable and incites fear, but can we
>> please lay off the melodramatic voice here ?
>>
>> There's three basic facts that lead us to this proposal:
>>
>> 1. We have an ever-growing set of modules and a constantly expanding
>> set of external dependencies.

Sorry to jump in the middle here, I did read the 60 some emails.

Generally I share the feeling of dismay about losing some important
order that we have in GNOME. Also I feel like this coming year might
really be the year for gnome devtools...

The success of the potential features I have in mind for Glade/GTK+
(and hopefully Anjuta) are going to depend vitally on how tightly we can
integrate our developer tools together as a unified chain.

Regardless of my bias in this matter as one who advocated the devtools
suite in the past, I can definitely feel Matthias's pain expressed here
that we are just running low on resources to manage it and as such
maybe the overly large desktop suite is becoming unmanageable.

So, if there are too many apps in the desktop suite (or in other suites)
could we just try to clean it up somehow ?

The process could be something like:
    - Release team decides what modules to thow away for GNOME 3.0
    - Gruesome exciting technical battles occur on d-d-l as the over-all
      3.0 module inclusion discussion heats up
    - Release team and the community would have to help to defend
      the relevance of older modules inclusion in GNOME 3.0 (for modules
      who's maintainers can not be contacted).
    - Accepted modules would at least have to build against the new
      stack as a minimum bar for passing (i.e. fully GSEALed GTK+ and
      only apis available in the new platform).

I'm sure that 3.0 is a great time to throw away alot of stuff that
may have become irrelevant over the years. This route would also
hopefully offload a significant part of the work to the community
(each maintainer would have to re-defend their module's significance
in the new GNOME)... allowing the release-team to focus on other
things while the community dukes it out...

This could even be a process with an undefined timeline, 3.0
could just start with an empty applications suite and we could
go on in the usual way [re]adding new applications every 6 months.

Thoughts ?

Cheers,
           -Tristan