From: Richard Wordingham (richard.wordingham@ntlworld.com)
Date: Sat Nov 19 2005 - 09:50:00 CST
Philippe Verdy wrote:
> What I mean there: does Burmese need ZWNJ in the *middle* of a word or
> only between words to avoid collisions with the next word? If this occurs
> in the middle of a word, does it create a sort of compound word which
> would be interpreted differently if they word was broken into two tokens
> separated by a space?
The virama sign occurs at the end of every syllable written as a closed
syllable or, by an ancient (?) misinterpretation, ending in the 'au'
diphthong. It is only suppressed (by stacking) in words of Indian origin
(chiefly Pali). For example, the word 'Myanmar' is written <m, subjoined r,
n, virama, m, a>. I am not aware that this is a compound - it has a
phonetically reduced form <b, m, a> from which the word 'Burma' is derived.
Burmese is primarily monosyllabic, but there's no shortage of longer words,
e.g. loans from English. In Burmese, space functions as a minor punctuation
mark.
When writing Burmese in Unicode, one should write the virama as ZWNJ even if
it is not followed by a consonant. Burmese conjuncts are pretty regular and
only a few ligatures need to be explicitly stored in a font. Unicode
'virama' (U+1039) followed by consonant should not manifest itself as a
virama. However, if ZWNJ is banned, some other method must be found to
ensure it is not followed by a consonant.
> If this does not change the semantic, then even that ZWNJ can be excluded
> from IDN: you can use the syntaxic ASCII hyphen to separate the two
> tokens, instead of using ZWNJ.
It chan_ges the se_man_tics no more than wri_ting like this.
Richard.
This archive was generated by hypermail 2.1.5 : Sat Nov 19 2005 - 09:51:06 CST