[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-i18n-doc
Subject:    Re: Wotm: translations on kde 3.3
From:       Eduard Werner <edi.werner () gmx ! de>
Date:       2004-07-17 19:22:27
Message-ID: 200407172122.27619.edi.werner () gmx ! de
[Download RAW message or body]

Njedźelu 18 julija 200401:05, Diego Iastrubni pisaše:
> בSaturday 17 July 2004 19:14, × ×›×ª×‘ על ידי Heiko Evermann:
> > >Which point am I missing here?
> >
> > That is an important point. And that is exactly the way the language
> > fallback was designed and it works fine.
>
> For mixing strings in an application? I dont mind having applications in
> different languages (kmail in Hebrew, konqui in Spanish, and konsole in
> German), but I demand that  each application will be in one language alone
> (or at least the vast majority of it).

But that is obviously impossible when you have an incomplete translation.
And if the user defines A as his preferred language and B as his second-best, 
he definitely won't be amused when he gets a mixture of A and English which 
he might not understand at all. Why do you demand this?

> Look at this from the translator point of view:
> I released an application, since I have tested it, and I have control of
> how does it look.
> With the current method, someone introduces a new app, and it will be
> "translated" even without me checking it's quality. I cannot garantte how
> good will it be.

Without checking the quality of what? I don't understand you here. Do you mean 
"without you checking the quality of the translation"? I would understand 
that, but will you be able to check all translations of all programs anyway?
Even for our small team, I already can't guarantee that ...

> now kstars has no translation. I did not check it, and the user should get
> it as in the first language that has a po-file.

And why don't you want to give the user already translated strings in his 
preferred language? 

> I understand that the "language fallback issue", but it should happen with
> 10% of the strings, not 90%. If you have 90% of untranslated strings, you
> are in the wrong language.

I'm sorry, but this is completely absurd. It would mean that you can't release 
incomplete translations, and furthermore, this would be *your* decision, not 
the user's.

> I do no expect KDE to test the percentage of translated/untranslated
> strings, but if a po-file for that language does not exist, them you KDE
> should try the next one as the "base", and then fill it with the ones
> bellow it.

This is not the point, at least not the main point. The main point was that a 
user would generally prefer a language he has defined as a fallback to a 
language he as not listed as a fallback.

> My main issue, is that I decided not to translate some applications, and
> still they appear in 10% Hebrew. This can be also Japanese, Chinese,
> Arabic, Russian or Yiddish, this is all the same:
> the application officially does not have a translation, and yet it displays
> a very minimal translation which is inherited from other modules.

A user with no command of English will like that. You could still suppress it 
by giving a translation file with msgstrs identical to msgids. This would 
also make clear that you *wanted* the English strings in a Hebrew environment
(as opposed to not yet finding the time for translation).

But generally, it might be nice to know where strings came from in an 
application ...

Cheers

Edi


[prev in list] [next in list] [prev in thread] [next in thread]