Re: GNOME Moduleset Reorganization vs. L10N

[Emmanuele Bassi <ebassi gmail com>]

On Sat, 2010-10-16 at 12:32 +0200, daniel g. siegel wrote:

> > > right, but in that case there is no problem to have it hosted on
> > > freedesktop.org, like many other dependencies of gnome.
> > 
> > it's inconsequential: the whole problem is having it, or not having it,
> > on the gnome.org server - because of authentication and access problem.
> > 
> > it doesn't matter where it is: freedesktop.org, clutter-project.org,
> > launchpad.net - they all are *not* gnome.org.
> 
> actually there is a subtle difference here, quote:

I know perfectly well what fd.o is, thank you very much.

the problem at hand is access - and hosting on fd.o has the same issues
as hosting on launchpad or clutter-project.org, or github: translators
want direct access to bypass the maintainers because they feel
(actually: because they made abundantly clear) that the maintainers are
a bottleneck[0].

again, if direct access to the repository is the goal then any hosting
*except* git.gnome.org is going to be a problem.

> however, if we are talking about applications and end user software your
> notion is of course totally right.

I'm talking about *any* project.

ciao,
 Emmanuele.

[0] personally, I don't believe for a minute that this is true. yes,
i18n is hard and we should all go shopping instead - *but* my past
experience with direct access from translator coordinators has been
spotty at best: wrong or partial commits, build breakage, delayed
releases because of commits slap bang in the middle of a distcheck, you
name it. after all these years, I came to the conclusion that I'd rather
have all i18n effort go through a branch and never to master; or, better
yet, go through email and automated notifications; or even a tool like
transifex's command line client. in these cases I can at least retain a
modicum of control on projects that I (and not the i18n teams) maintain.

-- 
W: http://www.emmanuelebassi.name
B: http://blogs.gnome.org/ebassi