What happens to translations when upstream is brought in

[Matthew East]

* Krzysztof Lichota:
> Matthew East napisał(a):
>>
>> * Krzysztof Lichota:
>>>> Christian Robottom Reis napisa?(a):
>>>>> On Wed, Aug 16, 2006 at 06:28:20PM +0200, Krzysztof Lichota wrote:
>>>>>>> Well, you made your point very clear, and believe me when I say that I
>>>>>>> share it as well.  However, you have to give us, Rosetta translator
>>>>>>> and teams a bit more of credit, for we too take a lot of pride in what
>>>>>>> we do, and I for one thing don't consider the many hours I spend
>>>>>>> translating as "mocking around".  Once again, I strongly believe that
>>>>>>> we should all get together and piece together a plan for incorporating
>>>>>>> more disciplined and well-tested procedures and protocols from now on.
>>>>>> Well, from my experience (in Polish translations) the results are not
>>>>>> very good.
>>>>> Thanks for the comment. Can you give me a better rationale as to why you
>>>>> found the results not to be good? I'd like to hear about your top 3
>>>>> issues so we can actually go ahead and address them.
>>>> I meant that translations were not very good. Mainly because translators
>>>> in Rosetta are not aware that we have our vocabulary, style guides, etc.
>>>> So they start translating ad hoc.
>>>>
>>>> There is also no way to control who translates what in Rosetta, so
>>>> everyone can translate everything (if this person is a member of
>>>> translation team). This is straight way to disaster as there is no
>>>> quality control.
>> You are missing the point, and saying it at the same time.
>>
>> The way to control who translates what in Rosetta is to control
>> membership of the translation team. This is what is done in every single
>> project where quality control is important - development, upstream
>> translation, documentation. The way to fix this situation is to make
>> sure that the translation teams are aware of the styleguides *before*
>> they are given the right to commit directly to the Ubuntu distribution.
>>
>> It's similar to giving translators the right to commit directly to Gnome
>> CVS, or elsewhere - there is nothing in the design of Rosetta that
>> prevents teams from doing proper quality control.
> 
> No, you are missing the point. Ubuntu translators team has no connection
> to KDE translation team. The person approving members of translation
> team does not have to even use KDE, not to mention to know style guide
> of KDE team. And there is no way to limit which translation can be done
> by whom, so person who is great translator of, lets say, XFCE tools
> might go and screw up KDE translation by using wrong vocabulary,
> incorrect plural forms, etc.
> The rule "everyone can translate anything" is wrong.
> Inside KDE translation team not everyone has right to commit
> translations to repository, exactly because not everyone is trusted to
> create correct (and good) translation. Control is necessary to provide
> quality.

Ok, I can see where you have misunderstood the situation.

There is no rule that everyone can translate anything.

Rosetta does not allow Ubuntu translator teams to disturb KDE
translations, or any other translations except for Ubuntu translations.
Equally, an Ubuntu developer cannot commit a patch to KDE code upstream
without also being a KDE developer.

At the same time, an individual cannot commit changes to Ubuntu
translations unless he is a member of the Ubuntu translator team. This
membership is the key stage in the quality assurance of Ubuntu
translation, and in the same way that you control who has commit rights
to the KDE repository, only trusted people should be added to the Ubuntu
translator team.

Do you see how it works now? Essentially, Rosetta (for Ubuntu) is a bit
similar to the KDE repository.

Matt

P.S. Please don't send messages to me as well as the mailing lists, I
read them both.