> suzmccarth wrote:

> > I think the important factor was the way the users thought of the
> > characters -- as syllabograms.
>
> Well, my Tamil student brought his pen and paper matrix of
> syllabagrams to school with him clutched in his hot little
> hand and it certainly can be found on the internet. This is
> one that I have copied onto my site. It wasn't hard to find.


> suzmccarth also wrote:

> John. I appreciate your detailed answers but wish to delve
> deeper if I may. Tamil also has an array, a matrix of
> syllables, a syllabary, so they call it. Tamil also has
> fewer phonemes than other Indic languages, both Indic and
> Dravidian. It has fewer syllables than Korean. I realize
> that the large number of Indic scripts is probably the
> biggest argument against precomposed units.
>
> The old typewriter input method, which we also used, had
> visual sequence input and while useless on the internet and
> no good for ligatures, shaping and rendering has a certain
> logic to it - it imitated handwriting.

Hello suzmcarth,

We have an *encoding* for Tamil in Unicode, largely influenced, rightly or
wrongly, by previous approaches to encoding Northern Indian scripts. I
don't think we don't want a proliferation of new encodings taking us back to
the old problems of lack of interoperability.

A key point in my mind is that what you are concerned about is more related
to how the users (such as the children you have come across) *input* Tamil.
That is quite a separate question from the encoding model used.

In CJK you can use input methods that are sound oriented (eg pinyin), shape
oriented (eg. Changjie), derived from roman transcription (romaji) or native
alphabetic transcription (eg. Bopomofo) or native syllabic (eg. hiragana)
but they all produce the same codes. Latin keyboards come in various
different layouts, but are capable of producing the same codes. So there is
nothing, technically, that prevents the development of a keyboard or even
input method that provides an approach based on alternative models.

I must say, though, that I'm not clear how a matrix-based keyboard would
much be different, in practice, from the current default input method that
is closely related to the encoding. I assume that you don't mean that there
should be separate keys for each syllable?

RI


============
Richard Ishida
W3C

contact info:
http://www.w3.org/People/Ishida/

W3C Internationalization:
http://www.w3.org/International/