NSM character type in Bidi algorithm

From: Chookij Vanatham (chookij.vanatham@eng.sun.com)
Date: Thu Dec 09 1999 - 13:55:55 EST


In the unicode bidi algorithm, I have come across the case where NSM characters
from the script written from right-to-left (Thai,...) following the consonants
which have the strong type and are written from left-to-right (Arabic,hebrew),
then, the NSM character will be rendered to the left of the strong right-to-left
consonant. I think this is because Unicode bidi declares that NSM will have
the direction according to the previous character.

        A,B -> arabic consonant (AR)
        x -> Thai vowel (NSM)
        y -> Thai consonant (L)

        Logical inp: A B x
        Visual out: x B A
        I think when comparing to this following case, not too sure if that
        what the intension of the bidi algorithm is.
        Logical inp: A B y x
        Visual out: B A y x
Please let me know if I missed anythings or any suggestion.

Chookij V.

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:56 EDT