Re: [Indic-computing-standards] Keyboard Issues

[Guntupalli K. <kar...@fr...>]

Reply to a old mail which I had composed but never sent.

On Sun, 23 Dec 2001 16:02:04 -0500
"Tapan S. Parikh" <ta...@ya...> wrote:

> 
> Im not sure if I havent read whats been posted so far closely
> enough, but I think one issue omitted so far has to do with keyboard
> issues.  This work would include at least the following.  (Im no
> expert on keyboard / encoding issues, Im just speaking from my
> experience as a developer and computer scientist, so please excuse
> any oversights / errors.)
> 

 I was going to give an account of Keyboard mapping/layouts also, but
since you have raised them appropriately, I will just add to it.

> 1) Document Existing Keyboard Mappings - This would include some of
> the phonetic approaches taken as well as the old typewriter /
> Enscript mapping.  Im sure others exist and that there are also
> probably some special mappings for certain languages.  These should
> be documented and generalized across languages if they can.  I
> believe this information would include the following:    - Keyboard
> Type(s) supported

         By this we probably mean the physical aspect, the key layout
(XT/AT etc) , no of keys (84/101/104), labels printed on the keys.

>     - Script(s) supported
      Since keyboard is not tied to any script (well, I may be wrong,
its tied to latin script), excluding the few special keys, we can have
any key to mean anything. Since all indian languages have basic
character set  Consonants ~= 35,  Vowels ~=14  Matras ~=14, other
symbols ~=10      total ~= 73, this can be represented on keyboard
using (10+26+10) keys along with shift key.

>     - Keypress-Character mapping
     This has been standardized in Inscript. See 
http://indlinux.sourceforge.net/keymap/keymaps.php (this is not the
official page, Official document is the ISCII document  form BIS (
www.bis.org.in )

> 
> 2) Drivers - For each OS, there needs to be an appropriate keyboard
> driver for each mapping -> encoding conversion.  One could imagine
> this could be handled in a modular way.  For example for each
> keyboard type there could be a standard module that does keypress ->
> code mapping, to some standardized encoding scheme across Indian
> languages.  Then, for each encoding, there could be a module that
> does the mapping from the standardized form to the specialized
> encoding.  Maybe too much overhead, but then you would be assured
> that you could use any keyboard type with any encoding / font.  

This would be a good solution , but wont help much in cleaning up the
mess that has been created by use of adhoc encodings. More proper
would be to use standards based approach & support keyboard layouts
which are commonly in use (like Inscript ,  typewriter etc)

> 
> 3) Create Labels / Overlays for these mappings - For each mapping,
> we should create appropriate button stickers or overlays.  In the
> long run we can also think of new types of keyboard hardware for use
> with Indian languages.
> 
  Sticker labels for some mappings (inscript, typewriter) are
available in market costing about Rs 100 . Local manufacturers can
make them available , when there will be good demand for them & by
keeping the layouts open we can ensure that there be many players &
cost remains low.

> Anyway, nothing new, but as long as we are trying to be complete I
> dont think we should leave this out.  Other tools such as
> text2speech, voice recognition, character recognition would be
> useful from an application (if not OS) perspective.  Another very
> important area is the documentation of which encodings are supported
> by which Database systems, and development of appropriate APIs for
> database interaction where encodings are not well supported.
> 
 Hmm, hav to work on this part.


Regards,
Karunakar