Tamil Discussion archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [WMASTERS] Some thing different!




________________________________________________

This week's sponsors -The Asia Pacific Internet Company (APIC)
  @  Nothing Less Than A Tamil Digital Renaissance Now   @
<http://www.apic.net> Click now<mailto:info@apic.net> for instant info
________________________________________________


Dear Arunmozhi;
Welcome to the current discussions. 

> * Ideal character encoding is (12 place for vowels)+(18 (or22) for
>  consonents)+(1 for     Ayutham)+numerals and that is all! This will come to
>  a total of about 31 + other assigned codes.
---(snipped)----
>* a mapping standard that maps the combination of 12+18(or22)+1+numeral
> codes to a combination of these glyphs.

As far as I know, the idea of using simply the uyir  and mei to
generate the corresponding uyirmei using softwares is not new. 
In our earlier discussions, we labelled them as "phonetic input".  
Using dedicated softwares, it is possible to "interpret" such 
sequential occurrence of meis followed by uyirs and replace 
them with corresponding uyirmeis. In Malaysia and Singapore, 
for nearly two decades tamils have been using DTP packages 
(of Naa.Govindasamy, Ravindran Paul, Muthu Nedumaran 
and others) based on this concept.  Anu mentioned about 
having thoughts of writing softwares of this kind for a long time. 
Recently the Tamilnadu Advisory Committee on Tamil 
computing has chosen this form of keyboard input as one 
of the three standards for keyboard layouts. 
For a review of various approaches to tamil computing, consult
the URL
http://www.geocities.com/Athens/5180/sintalk1.html

In the present discussions of 8-bit schemes, we are focussing
on a minimal number of glyphs that can go to define a simple
truetype/PS1 font that can be used without any hitch even on primitive
machines (such as 386 PCs, MacPlus,...). Tamilnadu Minister
for Tamilculture Prof. Thamizhkudimagan has already expressed
his accord to get TN Govt to distribute FREE worldwide 
one such font. It will be like the Times and other Roman fonts
we get free to keep on english communication going. No one
needs to go and buy a software to do simple tasks related to Tamil.

Of course, the propositions of the kind you invoke
>In PCs, Just the way we have a different softwares for accessing UTF8 texts,
>we will have to have different softwares for editing, storing, printing and
>display tamil characters. ...
is possible only on high-end/recent generation of computers.
To my knowledge, Unicode fonts for example cannot be used 
on primitive machines (because of the involvement of recent advances
in font handling technology).

Kalyan

-----
Subject: [WMASTERS] Some thing different!
Date:  Thu, 18 Sep 1997 11:28:50 -0500
 From:    "B. Arunmozhi" <arun@techapp.com>
Reply-To:  "B. Arunmozhi" <arun@techapp.com>, webmasters@tamil.net
 To:   webmasters@tamil.net

All,

Sorry for jumping into to the discussion all of a sudden ( just finished
a
project and hence have some free time).  I have good experience in
design
and developing software systems based for UNICODE(aca VC++).  But I do
not
have any idea about how font programming is done etc., So, I have
imagined
how font could be looked up and rendered on screen.  These imaginations
could be totally wrong.  As I do not know how the OSes like Mac and
Win95
are rendering the fonts, I have left my imagination run loose :-)!

Please forgive me if what I have putforth below is totally wrong and as
a
result, wasting all your time.  But please explain why it is wrong and
which
assumption is not correct.

An IDEAL chracter encoding system
---------------------------------

Out of enthusiasm to make the tamil character encoding without any
exceptions from logical sequencing, I started thinking on an Ideal
chracter
encoding system and tried to apply to an application.

        The result was to "Replace the 52 upper and lower english
chracters
in ASCII with Tamil
        character encoding."

( Now, instead of trying to fitting into 128, we will have to fit into
52
slots:-)!

* Ideal character encoding is (12 place for vowels)+(18 (or22) for
consonents)+(1 for     Ayutham)+numerals and that is all! This will come
to
a total of about 31 + other assigned codes.
  The idea is to replace the 52 slots allocated for english with tamil
character codes in ASCII.
  We will still have about 23 free chracter codes.  We can add the 4
grantha
chracters also and
  also we can add the numerals as part of the basic standards if
necessary.
Even after these,
  we will have at least 15 codes unassigned.
* There will also be a 512 glyph encoding standard and
* a mapping standard that maps the combination of 12+18(or22)+1+numeral
codes to a combination
  of these glyphs.

Both the character encoding and glyph encoding tables will have many
slots
free for future use.

Private or public font producers are free to add symbols and map them to
the
free character codes in what ever way they want.

It is possible to find a mathematical formula or a lookup table(2
dimensional to lookup for a given consonent and a vowel) for getting a
glyph
combination from character code combinations.

All POS systems and Terminals will have a lookup table to get a glyphs
given
a code or a combination of code.  The software in these systems will
then
receive character codes from computer and display them using the glyphs.
The software knows the mapping logic defined as part of our standards.

In PCs, Just the way we have a different softwares for accessing UTF8
texts,
we will have to have different softwares for editing, storing, printing
and
display tamil characters.  Once this method of spliting characters and
fonts
encoding is built into the operating systems, we can get rid of these
new
softwares and start using the native softwares. (again, please someone
should correct me on this).

This plan of encoding is not taking into considerations of limitations
of
existing operating systems.  It is purely based on the theory of
UNICODE.

Font developers and hobbyists
-----------------------------

I found the following remark written by Muthu in mail answering Nagu.

Muthu Wrote:
> 1. Ecoding pure characters may *cut-off* hobbyists who develop fonts for
>    Tamil without any inclination to writing codes (manipulation routines)

Dont you think, this should not be a consideration at all, as we can
provide
a simple font developer toolkit for the new character encoding. This
will
hide the gory details from the font developers.

POS systems and Character Terminals
-----------------------------------

It has been talked again and again about having to take into
consideration
of old systems while we do the standardisation.

IMHO, we do not have to be unduly worried about POS systems and
character
terminals.  They anyway cannot support the new fonts without
reprogramming
their hardware.  My assumption is, When we do the reprogramming, we can
introduce kerning to a limited extent. Please correct me if I am wrong.

The point is, we should try to encode only characters as far as possible
with STORAGE as main concern instead of DISPLAY as main concern.  Here,
I am
requesting you all to look into this problem with a different
perspective
from what we had been discussing all along.


Conclusion
----------

The benefits of encoding the characters alone in a logical sequence, are
many fold. The sorting will be simple and no extra programming work
involved.  You can just use the same Oracle, Sybase and other databases
with
out any modifications and sort and store the tamils names and search for
"Ǝ" or "ƵƎ" :-).  The comercial applications can be written
very easily and with less errors.  This will hyelp comercial use of
Tamil
applications (Ex: on Hotel Saravanabavan ґ!) grow very fast. 
Otherwis,
we will have to wait till we get a Tamil version of Sybase or Oracle!

I request your valued comments.

-arun
PS: 
-------------------------------------------------------------
B. Arunmozhi
Off : 314-916-7057              Technology Applications Inc.,
Home: 314-469-6459              #17 Research Park Drive,
http://www.techapp.com  St. Charles, MO 63304-5601
~~~~ People solve problems.  Technology only ~~~~~
~~~~ helps to make the solutions faster      ~~~~~
**Visit Larsen & Toubro at http://www.dbsindia.com/ltitl**
-------------------------------------------------------------


________________________________________________

Sponsors/Advertisers  needed -  please email bala@tamil.net
Check out the tamil.net web site on <http://tamil.net>
Postings to <webmasters@tamil.net>. To unsubscribe send
the text - unsubscribe webmasters - to majordomo@tamil.net
________________________________________________



Home | Main Index | Thread Index