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

List:       kde-i18n-doc
Subject:    Re: Rosetta for Edgy and KDE [Was: Rosetta for Kubuntu Dapper and KDE]
From:       "Donatas G." <dgvirtual () akl ! lt>
Date:       2006-07-12 9:59:46
Message-ID: 44B4C812.80106 () akl ! lt
[Download RAW message or body]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Vedran Ljubovic wrote:
>> You are free and infact encouraged to take translations from Rosetta
>> and import them into KDE.
> 
> You make it sound like something trivial. Unless there are some shortcuts that 
> I couldn't see, this involves:
> - visit launchpad every day to see what's new, swallow some Ubuntu propaganda 
> in the process
> - wade through the list of translated files that doesn't appear to be sorted 
> in any meaningful way, mixing files from KDE, GNOME and Ubuntu (latter two 
> obviously don't interest me)
> - export each file so that it gets sent to my email

Up to this point it sounds so painfully true...

> - make a diff to my local repository and investigate every single string to 
> see what exactly was done, because there is no way to be sure that Rosetta 
> translations are more complete from what I have
> - for every useful string, make a copypaste to KBabel window
> 
> Now imagine doing that for 5-6 major distributions. Also add that I personally 
> maintain translations for a few projects except KDE.
> 
> Jonathan, you seem to see only the best case where a translator works on 
> previously untranslated file on Rosetta and immediately pings the maintainer 
> to import it into svn.kde.org. How about the worst case scenario? When a very 
> large file, like kstars.po, gets *partially* translated on Rosetta and 
> *partially* translated on svn.kde.org at the same time only in different 
> parts? You honestly expect me to Kompare the 700+ kB files with 5000+ strings 
> and copypaste relevant strings into KBabel? It's easier to just translate the 
> file again.
> 
> And yes, that kind of thing happened to us.

You don't have to do it in a complicated way like this. You can take the
upstream file as a base and then apply a gettext command  to update it
using the rosetta version as a compendium:

msgmerge upstream.po upstream.pot --compendium=rosetta.po
- --output-file=upstream_updated.po

Then only the strings untranslated in the upstream version will be
carried over to the upstream version, while the ones translated on both
sides will be ignored. Then you can do

diff upstream.po upstream_updated.po

to get the strings that were translated (you might have to unwrap the
text first) and get the idea of what you want to change in the translation.

Once an acquired skill, this method IS MUCH FASTER than translating anew.

Instead of the single program file you can also use a compendium of all
rosetta translations that are extra to what is in the upstream, though
first rosetta has to provide a convenient interface to get those updated
strings. Instead of a single file you can apply this command then to the
whole of the translation tree....

I guess a Kommander script could be easily built to do that. Anyone
interested in having it? I could make a script like that.


> You can see in Launchpad logs, if there is such a thing. Some 3-4 young 
> translators decided to work using Rosetta and translated a bunch of files. I 
> asked them several times not to do that but they wouldn't listen. Then I 
> showed them files that were already translated some months ago... needless to 
> say they were furious. Now I'm trying to get them into translation the proper 
> way but they're just not interested anymore.

you could have shown them the location of the previously done
translations earlier, couldn't you, since you monitored what they were
doing?

> Now I might not have the best people skills, but it's no wonder that people 
> get disenchanted with the whole FOSS translation process when things like 
> that happen! Rosetta harmed bs team directly by creating discord among team 
> members and driving them away from translation as a whole.

I don't see that "Rosetta harmed bs team directly" - it is rather
indirect, and related too much to other team dynamics factors.

However, much of the criticism towards Rosetta and upstream here seems
to be valid - let's try to change it in the way the OpenSource works
rather than denounce Rosetta as a whole.

Donatas
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iQCVAwUBRLTIEvDtAb8AVdgrAQJslwP/auZj+l21+z5V/L5i4bFyJpGQm9diSAKN
ba58FjrgmG2YkG3PN36TcJT03vndrSSYr7eXJyTlaF+J+KepJyDqePo+2dUioORv
c468Yw/c4IIc6+dUqBB8LTlKCTlbBfi9dPH6s/Cb9EuFNoY6CIu+qq6QA4mC7U7F
slRwdeuvW44=
=XMer
-----END PGP SIGNATURE-----
[prev in list] [next in list] [prev in thread] [next in thread]