Re: Debian Debconf Translation proposal ( again )
On Wed, Jan 07, 2004 at 08:35:14PM +0100, Dominique Devriese wrote:
> Wouter Verhelst writes:
> > Well, I am: it's wrong. If it happens a few times during a single
> > release cycle per package, that's a lot. Heck, we don't even have
> > packages that reach that number in bare uploads, let alone uploads
> > that include a string change in debconf templates.
>
> And here I should have been more clear too. I meant 10000 times in
> total, I didn't mean 10000 times per package. You're right though
> that 10000 packages should be multiplied by the amount of changes in
> templates, so this would amount to, say, 20000 bugs that every
> translation team would have to file for every Debian release.
$ grep-available -nsPackage -FPre-Depends,Depends debconf | wc -l
828
(main and non-US/main only; and a number of those may well not change
over the course of a release. In general the *fewer* packages that use
debconf the better, probably, as every one represents user interaction
in the installation process.)
> Multiply that by the amount of languages in the world ( or, let's say,
> e.g. those for which it would be more or less useful to provide a
> translation, say, 50 ), and we reach a number of one million bugs that
> would have to be opened per Debian release, for translation only.
>
> You can argue about the numbers, of course, there are a lot of
> packages without templates etc., but I don't think you'll be able to
> get lower than, say 2000 bugs per language per release. I don't think
> it is realistic to expect such amounts of work from translators.
Do you seriously claim that it's more effort, or even comparable effort,
to deal with a small amount of administrivia and file a bug with your
translation as it is to translate the template in the first place? I'm
sorry, but this claim seems completely laughable to me. The translation
is real manual effort (even updating a translation), while the
administrivia is trivial and for the most part scriptable.
In addition, it is increasingly the case (what with the introduction of
alioth, etc.) that maintainers may be willing to give you access to
their CVS/Subversion/whatever repository to commit your changes directly
(but still make the actual release themselves), and then you may not
even have to bother with a bug report. See the debian-installer
translation effort for an example.
Cheers,
--
Colin Watson [cjwatson@flatline.org.uk]
Reply to: