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

List:       kde-i18n-doc
Subject:    Re: Improving KDE with translators
From:       Rinse de Vries <rinsedevries () kde ! nl>
Date:       2004-08-27 22:04:28
Message-ID: 200408280004.28891.rinsedevries () kde ! nl
[Download RAW message or body]

Op vrijdag 27 augustus 2004 22:51, schreef Gérard Delafond:
>
> What is the problem ? Nobody is supposed to translate everything. People do
> only what they want and what they can.
> Simply, it would be more responsible if the users would know what quality
> they can expect.


The problem is, how do you measure the quality of translations?
Based on the stats you mentioned, it is inpossible to rate the translations.

the only way to rate translations is to actually use it.
>
>
> And what will you do if anybody does this independantly ? Imagine a user,
> or a company, or somebody who does not like KDE, takes our stats and
> publishes a sort of classment ?
How can one make a classment based on stats?
The fact that KDE is only for 40% available in a certain language does not 
mean that the translation is bad. It could also mean that (to ensure quality) 
the team had decided to translate only the desktop and certain main 
applications.



> It would be more secure if we would rate 
> ourselves. There is no shame to be alone to translate KDE. I even think it
> is glorious. But don't tell people they can trust your work.
Well, the licenrs explicity tells users not to trust our work :o)
>
> I have been very surprised to learn many french people don't use french
> translation when using Linux-based systems (not only KDE). The reason is
> they don't trust the translation (I must acknowledge they are often right,
> some manpages have been translated in 1997 !)

Would they trust the French translation more if it got a 2 star rating?

If I look at the Dutch situation, a lot of Dutch kde users don't like our 
translation. Not that its bad, or that they don't trust it, they just don't 
want to use Dutch software. This doesn't say anything about Dutch 
translations in general, but more about the users. Most of them ar Tech 
Users, and are used to use English software, reading English UNIX-books, 
howto's, tech websites, etc. etc.. They won't use the Dutch language, no 
matter how many stars it is given, simply because they want to read the 
commands in their native language, and not what the translator has made of 
it.

On the other hand, a translated desktop lowers the tresshold for novice, 
inexperiend users.
Those advanced users, that don't use the French or Dutch translations (because 
they don't like, trust the translation) often do install it for their less 
experienced household members that want a computer in their language.

I also know some Dutch companies that switched from another system to KDE 2.0, 
using a Dutch translation that was less than 6 months old, translated by 3 
inexperienced translators (1 cook, with just 1 year computer experience, a 
14-year old boy and a ict-guy that just started using kde..), no translated 
documentation available, about 75% of the desktop translated, most of the 
apps 'blind' translated (i often had no idea what an application did while 
translating it)
It had all the ingredients to be a complete disaster (based on you quality 
checks), but it wan't. None of the companies (or users for that manner) 
complained about the quality of the translations (did get compliments 
though :). I don't think that they would have used the translation if they 
would judge it based on ratings (don't expect that the dutch kde 2.0 would 
get high ratings based on the stats), Even more, they probably would not have 
used KDE at all since they were looking Dutch environment.
In stead, they just installed SusE 7.x on their test machines, selected the 
Dutch language, tested it, liked it, and now uses it.


> The problem of the quality in the translation of computer programs is a
> complex program. Perhaps, there is no perfect solution.
> It would be sad if the KDE team would not treat it itself.

Threat the problem by solving it, not by rating it :)

Rinse

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