Semi-mechanizing the DTTP translations

l10n
#1

HI :slight_smile:
Ouch!! Sorry!  Please ignore my previous email and if you get a response from the LibreOffice subscription place then please just delete that too (unless you want to join in the fun there too) and please accept my apologies.  That was really embarrasing, sorry.

Apols and regards from
Tom :slight_smile:

________________________________
From: Tom Davies <tomdavies04@yahoo.co.uk>
To: Pierre Slamich <pierre.slamich@gmail.com>; "l10n+subscribe@global.libreoffice.org" <l10n+subscribe@global.libreoffice.org>
Cc: Hannie Dumoleyn <lafeber-dumoleyn2@zonnet.nl>; Ubuntu Translators <ubuntu-translators@lists.ubuntu.com>
Sent: Thursday, 17 January 2013, 16:30
Subject: Re: Semi-mechanizing the DTTP translations

Hi :slight_smile:
I finally tracked down the LibreOffice Internationalisation Mailing List!  Does the idea of semi-mechanising translations seem like a good idea for LibreOffice?  I don't know what systems are currently being used so please forgive me if you are already doing this or if it's a terrible idea.

Regards from

Tom :slight_smile:

________________________________
From: Pierre Slamich <pierre.slamich@gmail.com>
To: Tom Davies <tomdavies04@yahoo.co.uk>
Cc: Hannie Dumoleyn <lafeber-dumoleyn2@zonnet.nl>; Ubuntu Translators <ubuntu-translators@lists.ubuntu.com>
Sent: Sunday, 6 January 2013, 16:45
Subject: Re: Semi-mechanizing the DTTP translations

Hi Tom,
The approach works best for large files where the scale effect works best vs manual translations. We have tested it on documentation and related stuff so far. It works on virtually any po file, but you need to check whether it outputs translations good enough to actually reduce translation load.
Feel free to forward the original mail.

Pierre

Hi :slight_smile:

Would this semi-mechanising tool be good for other projects to use?  Is it good for translating websites, wiki's or printed documentation or all 3?

If it's good for other projects is anyone here on the main LibreOffice LoCos mailing list?  Could one of you approach them to suggest it?  If not please let me know. 
Regards from
Tom :slight_smile:

________________________________
From: Pierre Slamich <pierre.slamich@gmail.com>
To: Hannie Dumoleyn <lafeber-dumoleyn2@zonnet.nl>
Cc: Ubuntu Translators <ubuntu-translators@lists.ubuntu.com>
Sent: Friday, 4 January 2013, 20:03
Subject: Re: Semi-mechanizing the DTTP translations

We keep making incredible progress thanks to the process: we validated on average 400 strings a day going from 49289 untranslated strings on Dec 16th to 42746 today.

I've updated the structure and the instructions on the Pad to be more detailed and more linear. I've added a link to Redmar's script, instructions on validating the files and mass-correcting translations errors before upload.
Feel free to ask if you're stuck at any point.

http://lite.framapad.org/p/ddtpUbuntu

Pierre

Viva Low-Tech :wink:

When you come at the point of importing them back, let me know so that I can grant you upload rights to the mock project.

Sincerely,

Pierre
pierre.slamich@gmail.com

Hello Hendrik, Redmar, Pierre,

Redmar, thanks for writing the script.
The way I did the splitting so far is: open the sorted ddtp file

      in gedit, select lines 1 - 30.000 (which is about 940 Kb), copy
      these in a new document and save it. It only takes a few minutes.
      Then you can select the next 30.000 lines, and the next. Done!

Of course, using a script to split the whole file in one go is

      also very useful.

Hannie
Ubuntu Dutch Translators

Op 23-12-12 11:39, Hendrik Knackstedt schreef:

Hendrik Knackstedt schreef op do 20-12-2012 om 17:39 [+0100]:

I don't have a clean way to split them right now. I split them by

size to keep below 900ko (I took 800 for safety), but I then had to
adjust manually because the strings were split right in the middle.

Ok, I'll take a look at it and see if I can come up with something

useful.

I've been working with python-polib for a bit, so I think I'd be able to

create a script to split up a po file into multiple parts pretty
quickly. I haven't started yet, since I don't want to do duplicate work,
but please let me know if you want me to make a script or if you need
help with python-polib.