Sunday, January 25, 2009

Comparing former Yugoslav localisations

I was asked if I could write with some regularity something about localisation for a Serbian public. So I am looking more with more interest at the Serbian effort in Betawiki.  I find that there is enough to write about..

When you compare the localisation of these closely related languages, I find that the Serbian language is not keeping up with the others. In a way it is not that easy because there is both the Latin and the Cyrillic script that is to be maintained. The number of articles would have it that the amount of attention the Serbian Wikipedia is getting is much bigger.


bs Bosanski 100.00% 99.91% 100.00% 39.12% 12.07% 26,057
hr Hrvatski 100.00% 99.45% 94.58% 35.74% 41.72% 54,003
mk Македонски 100.00% 98.67% 25.72% 18.78% 74.83% 23,626
sr-ec ћирилица 100.00% 89.44% 65.61% 23.88% 63.79% 70,702
sr-el latinica 79.34% 57.72% 7.85% 3.78% 0.34% 70,702

Localisation is one of the most important usability factors. It is one of the few things that our communities can do to improve the user experience; localisation and writing the best articles. The rest is in the hands of the developers.
Thanks,
      GerardM

2 comments:

Anonymous said...

It's strange. My impression was that there is a strictly bijective correspondence between cyrillic and latin scripts in Serbian.

You could generate automaticaly the missing latin strings using available cyrillic strings. And why do you support 2 different localisation sets at all?

Anonymous said...

@Спас Колев: MediaWiki does not allow for real-time UI script conversion. It is currently thought of as too expensive (in computer power). It would be possible to write a convertor to convert sr-ec into sr-el, but even sr-el could not be complete because sr-ec is not complete. 2 different localisation have to be kept because a user must be able to choose his script preference for 'sr' (default is Cyrillic).