Re: tamil character set choice for a font encoding scheme

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

Dear Vasu,

Thanks for pointing out the issues with transliteration schemes.  Some
points I'd like to raise from a 'usability' perspective - with a lesser
inclination to computer based linguistic research.

I guess Kalyan's use of n^a came from the fact that it's alot easier
to read n^adan^tu (or even n^adan^thu) in plain text versus ndadandtu.
I was massaging this point when I was working on MURASU Anjal's romanised
keyboard (though I used '-' instead of '^' to avoid a shift).

Also the use of caps in the middle of a word (I can understand this is
a matter of getting used to) seem to get lesser acceptance.

I can live with oo for O, uu for U, ee for E etc.

For computer based linguistics research, I can understand your points.
Do you have any thoughts on how we could make this "human acceptable" ?
Esp. so with the growing interest to input Tamil through roman text ?

Regards,

~ MUTHU

At 02:27 PM 9/1/97 -0400, Vasu Renganathan wrote:
>
>Dear Kalyan and Sujatha,
>
>I like to point out a few problems, at least from the point of view of
>Tamil linguistics researchers, in the following transliteration chart
>Kalyan has suggested. 
>
>> 
>> CHARACTER CHOICES
>> vowels: 12   
>>        (a, aa/A, i, ii/I, u, oo/U, e, ee/E, ai, o, O, au, ak)
>
>Use of oo for U may not be appropriate because it is followed based on
>English sound system.  I don't understand why Tamil transliteration scheme
>should follow English sound system.  For example, English words with oo
>are pronounced U as in cool, tool and so on.  Why should we use oo for U? 
>Some other existing transliteration schemes also wrongly use ee for I
>based on English words like: reel, wheel etc. 
>
>I think Tamil lexicon has a standardized transliteration scheme that is
>accepted by many researchers, and it can be adopted for our purpose as
>well with suitable modifications.   
>
>> consonants:  18 
>>       (ka, nga, ca, nya, da/ta, Na, tha, n^a, pa, ma, ya, 
>>      ra, la,  va, zha, Ra, La, na )
>
>1) You are using ^ for dental na.  Why do we need a non-alphabetic
>character in the transliteration scheme?  We use combinations like nga,
>nya and so on to represent other characters.  Tamil University is using nd
>for dental n, which makes sense, in the sense that both are dental
>consonants. n is dental and d (dh) is dental too (spread the tip of tounge
>on the front teeth), and thus we make a combination of dental consonants
>to represent dental n. So, nda may be used instead of n^a following the
>scheme adopted by Tamil University. 
>
>2) You seem to mix voiced consonant in the chart da/ta. Again, based on
>the scheme that Tamil lexicon adopts, Ta may be used for (eTTu, paTTu
>etc.,), and t may be used for (pattu 'ten', vaattu 'duck' and so on.) 
>Use of da/ta here reflects English sound system, like in the words time,
>tree etc. 
>
>Like Sujatha points out, we always have problem with three n's, l's and
>r's, when we attempt to transliterate them. (nd,n,N / l, L/ r, R).  Where
>does zh belong to?  It is considered an r(hotic) sound rather than an
>l(ateral) sound.
>
>The transliteration scheme adopted by Tamil Lexicon conforms to the
>standards of IPA (International Phonetic Alphabet) system, and I don't
>think there could be any problem in adopting it for the purposes of
>computer processing.  I think the use of English sound system for Tamil
>transliteration scheme should be avoided. 
>
>With kind regards,
>Vasu. 
>
>
>
>