[ubuntu-in] SCIM keymap for Hindi Baraha system

[Gora Mohanty]

On Mon, 2007-08-13 at 20:43 -0400, Dinbandhu wrote:
[...]

An updated file has been uploaded to 
http://oriya.sarovar.org/download/hi-baraha.mim
Please see comments below.

> Bugs:
> 
> 1. "M" is not working as bindi. When "M" is typed, the character "M"
> simply appears instead of bindi.
> 
> 2. "~M" is not working as candrabindu. When "~M" is typed, the
> characters "~M" appear instead of candrabindu.

I had assumed the anusvara and candrabindu to be dependent symbols,
i.e., that they would appear properly only when typed after a
consonant. However, in retrospect, it makes more sense for these
to also at least enter the expected character when typed out of
proper linguistic sequence. I have fixed this, as also a similar
problem with other dependent characters like udatta, anudatta,
and visarga.

> 3. "e" should appear as "ए " but instead it appears as "इ " . 

Fixed. Was an oversight on my part.

> 4. When the nukta is used, a conjunct vowel cannot be added. 
> For example, when one tries to write "khush" (happy) using a nukta
> under the "kh", then a "u" cannot be added as a harsh u-kaar. Instead
> the "u" appears as a separate, full vowel. See: ख़उश | One cannot get
> the "u" to appear in conjuct form so long as the nukta is there. 

Fixed. I had omitted to add the nukta-consonants from Urdu as proper
consonants in their own right. 

> 5. "H" should appear as visarg  (ः) but instead it just appears as
> "H".

Done. See above. visarga was being treated as a dependent sign.

> 6. "z" should appear as ज़ | But instead it just appears as "z". This
> hindi character has two ways to be made: "jx", and "z". The "jx"
> works, but the "z" does not.

Oops, had missed that. Done.

> 7. When one writes "shakti", it appears this way: शक्ति | There ought
> to be a special conjuct for "kt". Perhaps that is just a limitation of
> the particular font, I do not know. 

It is a font issue. with OpenType, and Unicode fonts, the keymap is
relevant only for the sequence in which characters are stored. Rendering
issues, such as choosing between alternative forms for a conjunct are
up to the font. Which font are you using? If you have not specially
chosen a font in the editor (you don't need to do this with Unicode),
the font used is the first one found for the language in question. You
can find the list of available Unicode fonts for Hindi by typing in a
terminal
  fc-list :lang=hi
Replace "hi" in the above command with "bn" for Bengali. In any case,
it was my understanding that the above form (i.e., a half-ka followed
by a ta) was the preferred form nowadays, as opposed to the other form
which is sort of ta combined into the ka, which was what you wanted if
I understood you right.

Regards,
Gora