Thursday, December 18, 2008

Supporting Lingala in MediaWiki

Lingala is a language spoken by 200.000 native speakers and 7 million people speak it as a second language according to Ethnologue. It has its own Wikipedia, with a little over a thousand articles. The Lingala language makes use of other characters of the Latin script like the ɔ́. The accent has to be right on top in order to display well. In order to make it so, the CSS on the Lingala Wikipedia has been changed and a specific set of fonts have been ordered and identified.

The problem is that other MediaWiki installations do not provide this change. Consequently Lingala people are unlikely to support their language in Betawiki and, they are unlikely to upload their images to Commons.

In a discussion on Betawiki, browsers have been tested for this problem.
  • Dell WinXP Explorer: squares instead of open e/o (with or with no accents).
  • Dell WinXP Firefox 2.0.0.14: open o/e well; accents well.
  • Mac 10.2.8 Safari 1.0.3: open o/e well; accents not.
  • Mac 10.2.8 Firefox 2.0.9: open o/e well; accents not.
  • Mac 10.2.8 Explorer 5.2: open o/e as "?"; accents not.
  • Mac 10.4.11 Safari 3.1.1: open o/e well; accents not.
  • Mac 10.4.11 Firefox 2.0.0: open o/e well; accents not.
As I understand it, it is a matter for people to use a modern browser. If this is all there is to this, there are two parts to upgrading to a modern browser. How do we get this message out and, how do people get their upgrade?
Thanks,
GerardM

**Update**
It is not only browser related, the ɔ́ shows properly while I am editing this blog entry, but it does not in final form..

8 comments:

GerardM said...

My blog entry DOES show up properly in the Google Reader.
Thanks,
GerardM

Adam said...

That is interesting that it reads properly in my blog reader, too. I had to go to your actual blog to see what you were saying about it not showing properly. Interesting.

Adam

Anonymous said...

I bet the issue is the specific font being used. There is discussion of this very issue, with further links in a recent language log post: http://languagelog.ldc.upenn.edu/nll/?p=915

Anonymous said...

Really strange. The blog posting looks fine on en.planet.wikimedia.org, and not OK on http://ultimategerardm.blogspot.com/2008/12/supporting-lingala-in-mediawiki.html.

IT still has a long way to go when it comes to supporting languages...

Siebrand

Tgr said...

Your stylesheet requires a font which is not Unicode-compatible (Verdana in this case). Modern browsers (as in: not IE) are clever enough to disregard font settings when a certain character cannot be displayed in that font, but ususally not clever enough to the same when a character-accent combination cannot be displayed. (By the way, it doesn't work in FF 3.0.4/XP either.) Try setting the font to Arial Unicode MS, and it will look perfect. Of course, Arial Unicode MS is Windows-only, and even there not installed by default (nor is any other font with decent Unicode support), and somewhat esoteric to get working. (It is hidden between the Windows Office nonstandard install options, under some nondescript name I cannot remember... you can look it up in http://www.alanwood.net/unicode/, which is one of the best Unicode info pages (the other being fileformat.info).)

Tgr said...

In short, Unicode support is a pain. The upcoming server side font proposals might make thigs easier (though Unicode font files are large, and might take long to download).

GerardM said...

I entered a bug in bugzilla.
Thanks,
GerardM

Anonymous said...
This comment has been removed by a blog administrator.