[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Freebangfont-devel] Re: +AFs-Freebangfont-devel+AF0- RE: FREE Bangla So
From: |
Deepayan Sarkar |
Subject: |
[Freebangfont-devel] Re: +AFs-Freebangfont-devel+AF0- RE: FREE Bangla Software Avro Keyboard - NewVersion Available |
Date: |
Mon, 22 Sep 2003 09:21:06 -0500 |
User-agent: |
KMail/1.5.3 |
On Friday 28 February 2003 19:10, address@hidden wrote:
> Dear Deepayan,
>
> In using Likhan, when I type ta+halant, it immediately create the
> khandatta. However, after that, typing ZWJ doesn't cause any problem.
> But the developer didn't use exactly ta+halant+ZWJ to create khandatta.
It's not exactly that simple. I don't know how much you know about opentype,
but there's a concept of half forms and halanta forms of all consonants. When
unicode was designed, they basically didn't think about khanda ta, and later
when people asked they just said to use the half form of ta as khanda ta
(which means that the font should contain the khanda ta glyph where it should
have contained the half form of ta). So in a 'correct' font, you should see
see khanda ta wherever the half form was supposed to be used.
This by itself is not enough for proper rendering. Unicode makes a
recommendation of using ta + halanta + ZWJ, which is also not enough (try
writing the the word 'utkoch').
Currently the best you can do is ta + halanta + ZWJ + ZWNJ (this is also
microsoft approved). Andy has just submitted a very well thought out proposal
to the unicode list, and there's always the faint chance that it would be
acted upon, in which case this would change drastically again.
The point is, development on this is far from over, and you should not make it
completely hardcoded in your software. In particular, you should have a way
to change from one encoding to another, that may be useful in the long run.
> Sagar is only using ta+halant to make khandatta.
>
> This issue is important because I have to make it hardcoded in my
> software. Here, some are working with ta+halant, rest are ta+halant+ZWJ.
> This is really causing problem in typing and searching also. Just think in
> this version, as most fonts are using ta+halant, I have hardcoded it to
> make khandatta. In the next version, if I make ta+halant+ZWJ to create
> khandatta, users will face problem in the "search and replace" type
> operations. Again, if the fonts are updated, they will see that in their
> documents, all khandatta has been changed to foolish ta+halant.
> This is demolishing the Goodwill of both FBF and Avro Keyboard.
Well, you will have to accept the fact that khanda ta hasn't been standardized
yet. I have a script that converts any input file from one khanda ta encoding
to another. Maybe you should have something similar.
> Sources are available with the fonts. If the original developers are
> inactive, why others are not taking care of this issue. This should be a
> ten minutes work for you.
Actually, it isn't a 10 minute job. The original developer did say at some
point that he was looking at it. We did try to make an interim release, but
we don't have access to any proprietary font editor (and they don't run on
Linux anyway), and there were some odd problems with Pfaedit. I'll take a
look again when I have time.