Re: Response to Arasan's Comments

[Leong Kok Yong <kokyong@irdu.nus.sg>]

I think the issue does not lies with the sequence of keys input.
Because there are keyboard mapping software (free!) that allows you to
do quite complex keys mapping to achieve what you just describe. But
this is provided the 8-bit font does not requires glyph substitution.
Only kerning and straight one-to-one char->glyph correspondence within
the font.

For example, www.sil.org has a KeyMan software that can allows you to do
"k" "e" and get "ke" right..

The missing link here is a display manager that can take care of
glyph-subsitution made possible with true-type open or other technology.
The keyboard manager that does keys-substitution is actually freely
available. One can also see the key manager and the display manager as
one entity in some cases.

 keys -> [key manager] -> chars/encoding -> [display manager] -> glyphs
on your screen


George Hart wrote:
> 
> Obviously, in an ideal world (soon?), we will encode Tamil based on its
> character set.  That will, among other things, allow us to do proper
> sorting.  Unfortunately, there's no way to do that at present, short of a
> specialized program (such as Muthu's anjal system, and my own now defunct
> Indian Text Editor).  Meanwhile, we need a standard, which, given the state
> of things, must be one-level, i.e. one-to-one correspondence between what
> you type and what appears.  When the two-stage system is generally
> available, I'll be the first to adopt it!  Note that in a one-stage system,
> you to write "ke," you write the sign for "e" and then for "k", as that is
> the order in which the signs glyphs in Tamil.  When we get a 2-stage
> system, you will write "k" and then "e", and it will come out right (as in
> anjal).  But until that time, we need a standard.  G. Hart