Re: Using two translation workflows for one module
- From: Og Maciel <ogmaciel gnome org>
- To: Matej Urban <matej urban gmail com>
- Cc: gnome-i18n <gnome-i18n gnome org>
- Subject: Re: Using two translation workflows for one module
- Date: Tue, 23 Feb 2010 12:04:57 -0500
Cross posting as I feel this is relevant here as well:
On Tue, Feb 23, 2010 at 2:31 AM, Claude Paroz <claude 2xlibre net> wrote:
> So the question is, should we make it an explicit requirement to use
> l10n.gnome.org as the main translation platform for a module to be
> hosted in GNOME Git?
Short answer: absolutely not.
Long answer: As a fellow translator and coordinator for non-GNOME
projects, I understand and share the concern over having multiple
different "entry points" for translators to contribute to a project.
Having said that, however, there is absolutely no difference between
Transifex, Pootle, or DL as far as adding complexity to the process of
Bear with me here, as I try to explain my logic using a couple of scenarios:
Damned Lies: contributors have to register with DL and then ask to be
added to a "team" in order to use the web interface to reserve a
package for translation. All the work is done offline and then
uploaded back to a "holder area" where committers (i.e. those who have
a git account on git.gnome.org) have to manually do their review,
clone, merge, and push back to the repository. However, nothing stops
anyone from cloning the git repository, doing their work offline and
submitting it back via bugzilla.
Transifex: contributors have to register with Tx.net and depending on
the policy for the package mantainer, either request to join/create a
team or just starts translating. The translation process itself can be
done in many different ways, such as: doing using their inline online
web editor, downloading the .po file and doing their work offline,
completely ignoring the web ui and cloning the git repository, doing
their work offline and submitting it via the web ui, etc. It is up to
the project maintainer to decide whether the "submit" process
automatically commits the work straight to the upstream repository or
receive it as a patch via email for later manual submission.
Pootle would sort of have the same model of allowing users to do their
work using a web based editor and submitted back.
Having explained all that, I am not sure what the real problem is. As
you said it on your email, "freedom *is* the key" (my emphasis). There
isn't an overhead for anyone involved, really! To make things easier
for DL maintainers, you could remove the specific project so that
people don't access it via DL or just add a friendly message stating
that said project can be translated by accessing this or that url. The
project maintainer could also make this very clear from the project's
home page and this way avoid getting asked the same question: "how
come I cannot translate this via DL?"
The most important thing is, and I'm sure you will all agree with me,
to get good quality translations added and to lower the entry point
for possible contributors to our projects.
Disclaimer: I have contributed to both DL (minor) and Transifex and
also help administer a Pootle server.
Og B. Maciel
omaciel foresightlinux org
ogmaciel gnome org
ogmaciel ubuntu com
GPG Keys: D5CFC202
] [Thread Prev