Re: On the proposal for a font encoding scheme

["Muthu Nedumaran (Ezil)" <muthu@murasu.com>]

Some of my thoughts :

1. Direct sorting (i.e. using standard sorting functions) wil not be 
   possible with this as what is being encoded are glyphs more than
   characters.  In this case, we'll have to map all characters in the
   text to an intermediate table before passing it to a std. sorting 
   function.  The sorted text will then have to be reconverted back.

2. I see some problems with having a character code for "vaathu". 
   Search engines will have to search for (as an eg.) "talaivan"
   twice : one with the old-style lai and one with the new.  Am I
   right ? 

~ MUTHU

At 11:52 AM 9/2/97 +0000, Dr.K. Kalyanasundaram wrote:
>Dear Sujatha:
>Thanks for your prompt response. 
>
>Here are some quick responses to the points you raised:
>
>> Is sorting taken care of or is it dependant on software tricks?
>Only and if and only if we have all the 250+ tamil alphabets
>taken as such in some ordered way in the font encoding scheme,
>simple sorting by character number in the font encoding scheme
>possible. Else  (as in the present case of bilingual 8-bit schemes)
>one needs to resort to softwares to do the sorting. But this should
>not be a complicated one to write. All that one needs is to write
>up an equivalent mapping table for all 256 tamil characters where
>the equivalents of character positions are shown. 
>
>>How exactly are old style letters retrived ?
>Old style Raa, Naa and naa can be included as such as a glyph
>in the font encoding scheme. The modifiers of the old style letters
> (forward hook/kokki for lai/Nai/nai) allow getting the 
>corresponding old style letters -in two keystrokes. Mylai
>tamil font for example  allow typing these old style letters.
>
>> Are diacritical marks enough to accommodate the romanised scheme?
>The following markers are what we need:
>horizontal bar that goes on the top of roman letter ( a, i, u, e, o, n,
>)
>horizontal bar that goes at the bottom of the roman letter (k, l, n )
>virama type dot that goes at the bottom of the roman letter (t, n, l )
>and virama type dot that goes on the top of roman letter (n ) 
>In fact we do not need the last one, since one can use the same virama 
>dot that are used with tamil consonants to get ik, ith, ip etc.
>
>Is lower case way of writing adapted by research scholars?
>As a response to Vasu's posting, in a seperate posting I raised this 
>topic of what should be a standard for a transliteration scheme. 
>Currently there are TWO types of scheme (each with own variations)
>for writing transliterated tamil - one that uses diacritical markers
>on roman characters and one that is based on plain ASCII.
>Both schemes have their own merits and demerits. 
>
>With best regards,
>Kalyan
>
>