From: Richard Wordingham (richard.wordingham@ntlworld.com)
Date: Thu Apr 06 2006 - 18:53:09 CST
Kenneth Whistler wrote:
>> Well, the implementations is (as far as I can tell), not quite
>> automatic...
>
> Correct.
So where is the non-automatic part? I haven't noticed any mirroring
features in OpenType lay-out.
>> Anyhow, the number of bidi mirrored characters
>> is to be greatly reduced (IIUC).
>
> Actually, no. See:
>
> http://www.unicode.org/Public/5.0.0/extracted/DerivedBinaryProperties-5.0.0d11.txt
Correction:
http://www.unicode.org/Public/5.0.0/ucd/extracted/DerivedBinaryProperties-5.0.0d11.txt
> The total number is up slightly (537) from Unicode 4.1 (507).
Moreover, while I haven't bothered to found out which are new, they don't
all have mirror glyph characters, e.g. U+2246.
>> Also, when Greek (for instance) is written in the ancient
>> boustrophedon way, the glyphs are mirrored on every second
>> line of a paragraph (if not the whole text, I'm not sure). Yet,
>> Greek letters are not given the bidi-mirrored property.
>
> Correct. And likewise there is no need to give Phaistos Disc
> symbols Bidi_Mirrored=Y.
But the situations are not the same. Modern Greek has inherent
directionality - is it used in modern Arabic mathematics? If the Phaistos
disk characters had inherent directionality, it would be right-to-left.
Would it be better to encode two sets - left-to-right and right-to-left?
(This would make a mess of searching for substrings, and, far more
importantly, would be a very bad precedent for hieroglyphics.)
As for boustrophedon, Unicode support is very poor. You would also have to
do your own line-breaking.
Richard.
This archive was generated by hypermail 2.1.5 : Thu Apr 06 2006 - 18:57:17 CST