Re: Official GNOME module set

[Frederic Peters <fpeters gnome org>]

Ihar Hrachyshka wrote:

> I'd like to ask if there is any formal process to accept new modules to
> official GNOME release sets (and to delete obsolete modules from there).

The release team defines the "GNOME Core" module set, it has
dependencies on modules not hosted on the GNOME infrastructure (what
was known as "external dependencies", see [1] for a tentative of
classification of those).

  [1] http://mail.gnome.org/archives/desktop-devel-list/2011-May/msg00428.html

On top of that the marketing team defines a set of "featured apps".

There are also modules that are following the GNOME schedule, by
tradition, love of coloured tables, or any other reason. Those also
includes modules that live in a grey area, between core and apps,
there are notable modules in this situation, such as Evolution.


> I have concerns about lack of an officially approved list of release
> modules. As a coordinator of one of local GNOME l10n teams with little
> manpower, I need such list to better prioritize translation efforts.  Of
> course, our main priority should be an official release set coverage.

I believe your main priority should be to provide a translated
environment to your users, rather than strict adherence to a set of
modules, especially as we do not distribute GNOME by ourselves and
rely on distributors, they could, for example, replace Evolution by
Thunderbird.


> Recently, Sep 18, 2011, 'aisleriot' module containing 527 messages was
> (re-)added to official GNOME release set. Now my translation team is
> forced to pay additional attention to these new 527 messages and revisit
> the whole translation to be in line with requirements pushed for
> official modules.

aisleriot was split off gnome-games on April 10th, I don't know how it
was handled in damned lies, I don't know if it got many new strings
but certainly you can have a good chunk of them from gnome-games.


> Are there any plans to make official set population process more formal
> and beforehand?

We could certainly work on making the jhbuild modulesets as the source
of module lists exposed on damned lies, as well as marking modules
deemed "important" by the marketing team, to facilitate their
identification.


Cheers,

        Fred