[vlc-devel] commit: l10n: Initial Walloon translation ( Ga??tan Rousseaux )

[jpd at videolan.org]

On Sun, Nov 01, 2009 at 12:07:50AM +0200, R?mi Denis-Courmont wrote:
> Le samedi 31 octobre 2009 23:07:24 jpd at videolan.org, vous avez ?crit :
> > I was wondering: Wouldn't it be a good idea to use the french
> > translation as a basis for this one, updating it to more specific
> > walloon as you go?
> 
> Sounds like a bad idea. First, that scheme would only work the first
> time, and fail miserably if the French localization is updated.

It was only ment as a starter; further updates would be up to the
maintainer, so his call whether to track the big'un or not.


> Second, it would obfuscate which strings was translated and which
> wasn't.

Agreed. Could be worked around with the maintainer informing us what
part is actually translated and what part is still on the startup
booster language. Of course, keeping track increases the maintenance
burden and thus requires an active translator.


> Of course, you can question the usefulness of dialectal localizations.

I for me am not terribly enamoured with them, not so much because of
disk space[1], but because of the other reasons: Likelyhood of use,
likelyhood of being incomplete/inaccurate/otherwise undesirable to use.


> Of course, I don't have any statistics, let alone accurate ones,
> except for that one: ISO 639-3 has 7704 languages. So I won't propose
> a policy on this. But that's what it comes down to... language
> politics.

Most are as close to dead as to make no difference for our purposes, but
we can't really decide which. Even if we limit ourselves to `mainstream'
(and how many speakers is that?) we'd end up with a few dozen languages
as merely the top almost-a-dozen of languages with >100M speakers covers
maybe half the world's population[2]. So I'd prefer not to play that
particular game, but I don't want unlimited language burden either.
This brings us back to [1].


[1] That should be handled by the installer giving you a choice what
    languages to install. Even in an international setting there would
    likely be no more than half a dozen or so languages with any chance
    to be used in a particular installation, with one designated default
    for individual users on that machine who haven't made a choice yet.
    This is why debian came up with `localepurge' and macosx gives you
    exactly that choice in its installer. We do want to have support
    for many more, but users really shouldn't be burdened by masses of
    language material they'll never use because nobody in the vicinity
    even understands it.
[2] Depending on whether `secondary speakers' count.