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

List:       kde-i18n-doc
Subject:    Re: amarok 1.4.7 - string freeze
From:       Harald Sitter <sitter.harald () gmail ! com>
Date:       2007-08-06 16:17:20
Message-ID: 200708061817.23096.harald () getamarok ! com
[Download RAW message or body]


On Monday 06 August 2007 17:35:14 Mashrab Kuvatov wrote:
> Hi,
>
> On Saturday 04 August 2007 14:42, Harald Sitter wrote:
> > We also tought about making the UI more consistent in matters of
> > translation. In fact we are considering to remove translations which are
> > lower than average.
>
> I hope you reconsider it and continue to package all available translations
> of the Amarok. Hopefully, at least, you provide those translations which
> are officially shipped with the KDE (the others will not work without
> kdelibs and Co. translations anyway).

I brought this up for discussion, just wanted to get some thougths from you 
guys, because it recently came up in a amarok internal discussion.

> It is impossible to estimate and judge if the particular translation is
> usable, unless one speaks that language. For example, the Uzbek translation
> of the Amarok is at ~50%. The Amarok is perfectly usable in Uzbek, I know
> many Uzbek speaking users have no problem with it and like it very much.
> According your definitions 50% of translation is not enough and the Uzbek
> translation needs to be removed. However, in the limit you set you do not
> take into account how many messages user most probable will see are
> translated. For example, there are quite a lot of messages of the helix,
> yauap, etc. engines which are barely used by our taget users (Uzbek users).
> We do not translate those messages, since first we do not have enough
> manpower, and second it is very unlikely that our users will see those
> messages. There are quite a lot of error messages which are very technical
> and there is no understandable Uzbek translation for them. Therefore, we
> leave them in English to so at least the users with enough technical skills
> could understand what is going on. I could continue with other such
> examples. The point is that although for you it seems that the translation
> is not good enough, it can be so for the native speakers.

You're totally right in matters of unjudgability of the translation, the 
problem in general is that we can only rate a translation's quality by it's 
completness. And it's probably out of doubt, that shipping translations like 
az, be or rw is more harmful than useful, since they don't even reach 1% 
completed strings. But from _just looking at the numbers_, I'd say that uz 
can't be in shipable condition either, since even en_GB (for which I'd say it 
is reasonable to have less than 50% "translated") is almost 10% more 
translated. To be continued in the next paragraph... ^_^

> You (please do not take it personal) approached the problem from the point
> of view of the person to whom having computer, Internet, Linux, compiling
> from sources, etc. are normal. Unfortunately, for us people from the
> third/developing/you name it world it is not so. We have to struggle much
> more to get our work (localization) to the user. At the moment, our work
> reaches the users only through the Linux distros. So, if you remove the
> Uzbek translations, distros will not package it, and our users will not get
> it.

Well, at least we didn't notice then.
Anyway, your objection that upstream shipping is the only way to reach the 
users at all (in various regions) is actually totally reasonable and I have 
to admit that I didn't think about this at all. I have to say, that actually 
changes almost everything. Good job there, now we have to rediscuss 
everything within the team :-P

> Moreover, if the application is not fully translated, it can be a starting
> point for someone who wants to contribute. So, like I said, please
> reconsider your decision, and continue distributing all existing
> translations.

Again wasn't a decision at all ;-). But actually we also discussed the point 
that actually not complete translated apps might attract people to finish the 
translation and maybe take maintainership. Though isn't it the same with no 
translation? ... For starting the translation one has to get the SVN version 
anyway, so I'm not sure whether this is something to worry about at all.

Special exceptions might be an option? Or maye set the limit so low that no 
reasonable translation would be kicked?

Btw, I don't think having a general level for extragear is the way to go. As 
Mashrab pointed out there are a lot of strings in Amarok, for example, which 
aren't considered vital for the application usage, and the amount of such 
strings varies from application to application. So, if at all, then I'd think 
of setting a very low default limit which has to be reached to be shipped 
with extragear apps at all, and special conditions if the app want to (yeah 
this is might be a problem of tracking the actual conditions as long as they 
are not collected on a very central and very visible place, but it's the best 
thing to do, I can think of right now).

-- 
Harald Sitter
Amarok team        Rokymotion division     Mail: harald@getamarok.com
amarok.kde.org     Good news everyone!     Jabber: apachelogger@kdetalk.net

["signature.asc" (application/pgp-signature)]

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