Re: Letter frequency in unicode (Was Re: [Indic-computing-devel] Free UCS outline font)

[Arun S. <ar...@sh...>]

[ Snip lin...@li... - some of this may
  be off topic there ]

On Mon, Mar 11, 2002 at 02:13:32PM +0530, Guntupalli Karunakar wrote:

[...]

>  To support the Inscript keyb layout.
>  on Inscript keyb layout they are under keys
>       ra - 'j'
>       ka - 'k'
>       nA - 'v'
>       ta - 'l'
>       pa - 'h'
>  similarly 
>     halant on  'd'
>     VS I   on  'f'
> 
>  And these keys are the under  normal finger positions  'a s d f'  'h
> j k l'  on a typewriter :)

Yes, somebody must've run a similar analysis before designing the
inscript keyboard. But I suspect that the analysis was done on
devanagari. I'm not sure I could reproduce the same numbers on kannada
for example. So does Kannada-inscript make sense ? May be, if the
characteristics are mostly the same - the cost of deviating from
devangari-inscript may not be justifiable. 

I found that I was typing the syllable "lli" (U+cb2 U+ccd U+cb2 U+cbf)
very often in kannada and it's pretty inconvenient with
kannada-inscript. I don't think this occurs often enough in devanagari
based languages.  I'm sure you'll appreciate this if you have typed your
name with inscript :)

	-Arun

PS: Should this discussion be on -standards ?