Danny's Weblog

Progress with using Cambodian fonts

I had been trying for a long time to find someone who spoke good English and could explain to me how Cambodian people use computers. This week it occurred to me to try the opposite: to find a Cambodian person who knows computers and speaks some English. I immediately made progress.

1. Some months ago I happened to be using a recently reimaged computer with just the right security settings, and was able to view the procedure of using Cambodian fonts with Internet Explorer. Unfortunately I don't have my notes in front of me presently, but it was something like this:

I went to a Cambodian website which has both English and Cambodian text versions. (I wish I could remember what it was; many sites installed by Westerners have Cambodian present only as GIF images, not text – and Cambodian-only sites, while they may offer to download fonts, may well not offer English-language help.)

2. Immediately IE offered a prompt to install fonts. I forget the exact term it used. I suspect that with the default security settings this prompt is never seen.

3. I clicked OK and the fonts downloaded and installed without needing a reboot or anything. (I wish I knew exactly what you have to do on a website to provide this feature.)

4. When I checked, a couple of fonts had been installed in the standard fonts system. These were the oldfashioned (1995) Limon fonts, and only two of them.

5. The webpages do not call for the Cambodian language because "kh" does not seem to be supported by the fonts system – that is, there is a "kh" option in IE but it seems to be connected to nothing. Instead, the HTML simply calls for a certain font (I guess this can be done with CSS also) which happens to be a Cambodian font.

6. But how do you *enter* text in Cambodian? The answer that I got this week is that you go to Control panel – keyboard and select the "US International" keyboard layout. (These options vary somewhat between Windows versions. In particular W98 seems to allow only one US English keyboard to be in the dropdown list at a time, so switching between "US/101" and "US International" is a pain.)

7. With "US International", the keyboard connects certain ctrl-alt and ctrl-alt-shift combinations to extra characters, so the lush profusion of characters in Cambodian can be accommodated. Many computer and CD stores in PP have a copy of this sheet which they will give you for free. (However, I have actually not seen stick-on key labels.) The standard "US 101" layout seems to trap or throw away most ctrl-alt combinations.

8. This keyboard layout has a number of drawbacks. In particular, the font system seems to have no real intelligence. It doesn't know there is a relationship between standard and subscript consonants, and because it doesn't look at the height of the preceding character, you have to pick the correct height for an upper diacritic vowel yourself! Likewise there is only one character to represent subscript d and dt: although that is in correspondence with the *graphical image* of text, it causes irritating issues in spellcheckers, indexing etc. However, as far as I can see, nobody on the ground is currently using Unicode fonts, despite NGO recommendations.

9. I am currently preparing a table showing the correspondance between alt-codes 0 through 255 and the characters in two common Uksor Dai and Uksor Mool fonts. As I get more into it I keep thinking of new features to add, so don't hold your breath.

10. It's interesting to see how the designers of these fonts allocated Khmer characters to alt-codes. For instance, the alt-codes which normally produce Arabic numerals produce the Khmer numerals, but the Arabic numerals can also be created if you know how. Likewise, in many cases, the position of a Khmer character corresponds to the position of an English character with the same *sound*. I want to find some sort of "US-International" *font* to add to my table of Khmer characters: I suspect that characters like u-umlaut will correspond to some sort of u sound in Khmer.

11. It's surprisingly difficult for me to create the table, because I don't *already* know Cambodian very well. For instance, many of the diacritics, and even the subscript characters, look like undefinable squiggles even at 36 points. (Likewise I'm sure there must be websites *in* Cambodian which describe most of this – but if I could read them I wouldn't need to go to all this trouble!)

12. Even using the "US-international" layout a few key combinations are still inaccessible. In my case, a laptop captures one combination to adjust its soundcard balance between channels. In addition, you probably want to turn off the standard keystroke combinations which Windows offers for switching between input locales. If you have set up one of those keyboard accelerator things, you will probably have to disable that too.

Debug: hittotal: 35 startban: 0 dancookie: endbandate: banned: 0 tempdate: tert: jse: jsno jsh: 35