From: Dom Lachowicz (doml@appligent.com)
Date: Tue Sep 10 2002 - 17:48:49 EDT
On Tuesday, September 10, 2002, at 05:30 PM, Alan Horkan wrote:
>> 1) Have translators use tools that they like and are familiar with,
>> including all of their benefits
>
> yeah, maybe
This is not a maybe. This is a definite.
>> 2) Drop our .po files inside of GNOME's CVS repository a few days
>> before a release. Because of #1, translators will just translate all
>> of
>
> maybe i am wrong but i would imagine there is not much of a crossover
> (at
> least at the moment) between the people who translate for Gnome and the
> people who translate Abiword.
>
> Of course it would be good to get some help from the Gnome translator
> too
> of course.
That's the benefit of this approach. We get more translators to:
1) Update existing translations
2) Improve and correct existing translations
3) Add new translations
This is a *huge* benefit, and works _because_ our sets of translators
aren't 100% overlapping - but they would become so.
>> our strings for us, and then we re-merge them. Completely effortless
>> on
>> our part. Relatively easier on our translators' parts.
>
> hmm, i thought there were quirks in the strings -> po -> strings
> but maybe they have been ironed out ...
Dude, the idea is that strings files will die. No more strings. Ever.
> there is also something at the back of my mind i remeber mentioning to
> dom
> about locales should include information about the next best fallback
> for
> missing strings and if i remember correctly dom told me to file an
> enhancement request against gettext but i got distracted.
>
> So does gettext handle dialects well? it might well be won over if
> there
> is a much easier way to make the same change to en-IE, en-AU, en-GB,
> en-NZ
> all in one fell swoop.
It handles it better than we could ever hope to using our strings files.
>> Still don't see any benefit? I see 2 *HUGE* benefits.
>
> Okay so the add on tools that work with gettext are the whole point,
> not
> so much gettext itself.
gettext like software is the most boring stuff on the planet. It's
nearly functionally identical to our strings file or the hash map you
wrote in "Intro to Computer Science"
Define Translator as Map <string, string>
The fact that:
1) gettext is fast and bug-free
2) handles fallback locales reasonably well
3) is the de-facto standard for all GNU and other free tools
4) translators like it
5) tools exist for manipulating the map files
is why we want to do this.
> Just to point out that winamp3 uses a system almost identical to the
> strings files abiword uses
Wow. Why aren't more Winamp translators translating AbiWord then?
</sarcasm>
Just to point out that:
ftp://ftp.gnu.org/pub/gnu/*
ftp://ftp.gnome.org/pub/*
ftp://ftp.kde.org/pub/*
ftp;//ftp.redhat.com/pub/*
...
all use gettext...
If we can make translator's lives easier. potentially attract more
translators, resulting in more people using our product *without any
negative impact on our product*, what else can I do other than
whole-heartedly endorse that plan?
Dom
This archive was generated by hypermail 2.1.4 : Tue Sep 10 2002 - 17:52:18 EDT