Re: Arial Unicode MS and Code2000

From: James Kass (jameskass@worldnet.att.net)
Date: Sun Jul 08 2001 - 09:27:06 EDT


Rajesh Chandrakar wrote about a display problem.

I wrote:

> This could be a font problem, an encoding problem on the
> page in question, or a problem with the operating system
> support.
>
> Here is the line in question reproduced in Unicode (UTF-8):
>
> वेंकटाचलपते (निन्‍नु नम्‍मिति वेगमे नन्‍नु)

...and here it is without the special character which is supposed
to force selection of the half-letter form instead of the
conjunct. On this system, the matras appear as expected:

वेंकटाचलपते (निन्नु नम्मिति वेगमे नन्नु)

According to The Unicode Standard 3.0, Figure 9-7 on page 216,
the half-letter forms are supposed to be generated with the
consonant + virama + ZWJ (Zero-Width Joiner). ZWJ is U+200D,
or #8205.

If my understanding is correct,
consonant + virama = consonant + virama
consonant + virama + ZWJ = half-form
consonant + virama + consonant = conjunct ligature form

(There are additional rules and variations for "ra" forms
and so forth.)

Even though the OpenType tables in the font aren't yet
complete, they should be far enough along to make the
half-form substitutions. Since conjunct ligature forms
are appearing in the second example above on this system,
I've probably made a mistake in the format of the OpenType
table for the "half" feature or something.

Any clues would be appreciated. Does anyone's system and
fonts display the first example above with half-forms? Is
my understanding of the encoding for Devanagari flawed?

Meanwhile, guess it's time to review some OpenType specs...

Best regards,

James Kass.



This archive was generated by hypermail 2.1.2 : Sun Jul 08 2001 - 10:23:06 EDT