[Ietf-languages] Between language and script in Burmese

[Simon Cozens <[e-mail]>]

Hello!

I've been working on a system font which covers a number of minority 
languages and scripts of Burma, some of which are not currently 
addressable because they lack IETF (and OpenType) script/language tags, 
or where the correct tag combination is not obvious.

The Burmese script has many language-specific and context-specific 
variant forms (see UTN11 - 
https://www.unicode.org/notes/tn11/UTN11_4.pdf - for examples), and the 
boundary between script and language is not always obvious. Some of 
these differences in letterforms are encoded separately in Unicode and 
some of them as allographs. It's all a bit of mess.

The easy problem we have is the Thai Mon language. This is a variant of 
the Mon language used by Mon people in Thailand. It has its own distinct 
script tradition. 
(https://www.unicode.org/L2/L2020/20163-arakanese-mon.pdf) There's no 
distinct language subtag but I believe mnw-TH is enough to distinguish 
this language - although we may have to pull some OpenType strings to 
enable that distinction to select Thai Mon specific orthographic forms.

The hard problem we have is that some of these language-specific variant 
orthographies are used to write text in *other* languages. In that 
sense, they are essentially functioning as *different scripts* to 
standard Myanmar.

For example: a document written in the Shan language using the Shan 
variant orthography of Burmese is clearly shn-Mymr, and setting the Shan 
language in a document should be enough to activate the Shan variant 
forms. No problems here. And a document written in Pali using the 
standard Burmese orthography is obviously pl-Mymr, and because it's 
standard Burmese, a Burmese font doesn't need to do any magic to get the 
right glyphs.

But what is a document written in the Pali language using the Shan (or 
Khamti, or Mon) orthography? Do we need variant tags to distinguish the 
flavour of Burmese script used in these cases? Shouldn't Shan, Khamti 
and Mon actually be separate scripts? And if not, how on earth are we 
going to get browsers to choose the Shan forms for this document, 
without pretending that it's actually written in the Shan language?

Any advice would be helpful!

Thanks,
Simon