From: Michael Everson (everson@evertype.com)
Date: Sun Aug 19 2007 - 16:28:17 CDT
At 22:19 +0100 2007-08-19, Sinnathurai Srivas wrote:
>Tamil writing does not directly represent phonemes, rather
>it represents places of articulation, and this writing generates
>scalable phonemes for use.
Tamil writing (like any writing) is a set of structured marks, such
as ink on paper. The marks to not "generate" phonemes. Human beings
give certain values to the marks. We don't encode what happens in
people's heads. We encode structured marks in a structured way.
>Following the explanation given below, I'm not satisfied that even
>current Unicode standards are followed. In that, Tamil does not have
>conjunct. How can Unicode encode conjuncts of Graphical form for
>Tamil.
KSSA is historically a conjunct. It is analyzed for other scripts as
a conjunct, and it has been for Tamil as well. The fact that it is
encoded as a string of three characters IS NOT A PROBLEM.
>Additionally, the conjunct processing is a very daunting programming task.
This isn't true. It is extremely easy to deal with this string of
three characters.
>What justification Unicode has, even if it decides x is a conjunct,
>when only one character count of conjunct type is idetified?
It is a historic conjunct of KA with a killed vowel plus SSA.
>Sorting does not work, voice recognition breaks, developers postpone
>their work siting daunting tasks, etc.. are some problems introduced
>by this wrong assumption.that Tamil has one conjunct.
NONE of these problems are caused by KSSA being represented as a
string of three characters. English language text-to-speech
recognizes the string of two charcaters "s" and "h" easily, and the
sound they make is neither /s/ nor /h/.
>Simply said, there is no conjunct in Tamil. Why Unicode says there
>is conjunct in Tamil?
Simply said, what you are saying is simply not correct.
-- Michael Everson * http://www.evertype.com
This archive was generated by hypermail 2.1.5 : Sun Aug 19 2007 - 16:33:04 CDT