Incidentally, while I don't know much about this area, I have read that the Unicode characters allocated to Burmese, while they are sufficient to represent the language using an advanced Unicode implementation, are insufficient - one needs more glyphs with distinct codes - for conventional software to handle it.

But because there weren't a lot of computers in Burma back when, there weren't pre-existing standard codes with those glyphs defined, so unlike more economically important languages, Burmese got the full purist treatment from Unicode.

So anyone aspiring to design a Unicode font for Burmese will need to know a lot about the writing system to properly implement support for the language - and the list of characters won't show all the glyphs you will need to define.

There's a Burmese font and shaping support included in the Windows 8 Developer Preview. It's not fully baked, but might be worth a look. Can also share some contacts. Send me an email "simonda@..."

Cheers, Si

You should definitely be in touch with John Okell via the School of Oriental and African Studies (SOAS) in London. He literally wrote the book on Burmese script, and is an invaluable resource.

I am in touch with a typographer in Burma too, and I will direct him to this thread.

I strongly recommend checking out the Win8 developer preview to which Si alludes. This is the first official implementation of the script in OpenType (there have been some hacked attempts), and is a useful guide.

John, what you say is true of all the South and Southeast Asian scripts in Unicode: they all require complex character ordering, glyph shaping and dynamic mark positioning. Burmese is no different in this respect than Thai, Khmer, Lao, or any of the many Indian scripts or Tibetan.

Interesting. I gather John Okell created the first Burmese font, so I'll definitely be making a trip to SOAS.

Do I need to install the Win8 Developer OS to get at the font stuff? Si, I'll e-mail, thank you.

>Do I need to install the Win8 Developer OS to get at the font stuff?

Essentially yes, as the font won't work without the shaping support.

>Si, I'll e-mail, thank you.

Actually Okell and the designer John mentions are probably the best people to talk to, but get in touch and we'll try to give you some details of known limitations with the Win8 DP support.

@John Hudson:
Burmese is no different in this respect than Thai, Khmer, Lao, or any of the many Indian scripts or Tibetan.

It is true that Thai has similar complexities - I remember seeing specific complaints about the treatment Burmese had gotten from the Unicode consortium, and I thought that Thai was more economically important and had a longer history of use with computers (i.e. there's a six-level TTY code for Thai that's been in use a long time) and so might have had a slightly easier time of it.

But I also had read other comments that Burmese was just one example of the short shrift the less major languages had received - unlike even, say, Arabic - so I'm not surprised to hear it's a general complaint.

Right. I've installed Windows 8 on my desktop machine. Not enjoying the experience unfortunately, but I'll have a good explore with Myanmar Text that comes with it. At first glance, I'm seeing a lot of OpenType features I've never heard of before (rphr, psts for example) but after a couple of days' learning about how complicated the writing system is, I'm not surprised there's going to be a lot to learn.

It seems I need to get really clear on the distinction between 'letter', 'character' and 'glyph'. There are some odd behaviours I hadn't counted on (unlike in Thai which I'm very familiar with). In Burmese, the 'a' vowel should have a tall variant when positioned next to certain initial consonants, but the tall variant also needs to be keyable directly and has a codepoint — no problem there (at least until you start on Karen which I understand uses the two characters to represent different letters!). By contrast, the 'u' under-vowels are sometimes contextually switched to tall variants (where there's already a descender in the way?) and in other cases it'll be interesting to figure out how to allow the user to choose them (perhaps even without a glyph palette) even when not actually necessary, when there isn't a codepoint. Challenges ahead. I'm really starting to see why proper Burmese fonts have taken so long to appear.

I'm looking at the Microsoft spec for Burmese", does anyone know if this is still accurate (it's dated 2002).

Well, this has certainly been a learning experience. My OpenType features are getting very lengthy!

Currently attempting to script the OpenType language tags, and seem to have found the current script tag mym2 is preferable to the outdated mymr tag, though there's nothing I can find anywhere to confirm this.