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

List:       kde-i18n-doc
Subject:    KDE translation history for various teams
From:       Karl Ove Hufthammer <karl () huftis ! org>
Date:       2009-08-14 22:12:56
Message-ID: 200908150012.56797.karl () huftis ! org
[Download RAW message or body]

KDE translators,

Are you interested in the history of KDE translation for your language, which 
people have been working on the translation, and when they have done so?
This information is actually stored in the PO files, and I've created a simple 
script for displaying it graphically. Here's two examples, for Norwegian 
Nynorsk and Norwegian Bokmål:

http://huftis.org/kritikk/translation-statistics-nn.png
http://huftis.org/kritikk/translation-statistics-nb.png

The blocks show when each translator has been working, and the number inside 
each block shows how many files (not times) s/he edited that year. The names 
are sorted by last year worked, and then by first year worked.

How to generate this statistics? First you'll need to normalize the headers in 
the PO files. You can easily do this with the pology tools:

posieve normalize-header folder

Then you have to download the scripts:

svn checkout --depth empty \
svn://svn.skolelinux.org/skolelinux/trunk/i18n/skript/
cd skript
svn up translation-statistics.sh
svn up translation-graph.R

The system consists of a simple bash script and a R script.
You'll find R at http://www.r-project.org/ and it's also packaged
in all major distros. You'll also need to ensure that the ggplot2
package is installed in R. (Running R, preferably as root, and typing
install.packages("ggplot2") should do the trick.)

Then you only need to run

./translation-statistics.sh folder-with-po-files

and the image will be created automatically, and even optimised for putting on 
the Web if you have optipng installed.

Some caveats: The statistics may be unreliable for several reasons. For 
instance, if you use a text editor to edit the files, your name will not 
automatically be added to the file. Also, in older versions of KBabel, it was 
possible to turn off the automatic credits insertion. And if you used the auto-
translation feature of KBabel, your name will be listed in many files, even if 
you did not actually edit them. (There are also various other reasons that the 
statistics may be wrong or misleading.)

If the statistics looks messed up because the same translator has used 
slightly different names, or different e-mail addresses, this is an excellent 
time for cleaning up the PO file headers … This command should give you 
a nice overview:

find folder-with-po-files -name '*.po' | xargs grep -h '^# .*<.*@' | \
sed -r 's/([^,]+),.*/\1/' | sort | uniq -c | sort -nr

Feel free to suggest any improvements to the scripts.

-- 
Karl Ove Hufthammer
http://huftis.org/
Jabber: karl@huftis.org

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