Solution for Extended Tamil

[James Kass]

On 2024-01-25 1:02 PM, Richard Wordingham via Unicode wrote:
>> downloadable font for 'correct' display.  Are there any such notices
>> or disclaimers?
> That's probably a question for those who read Tamil - there may be
> little point in putting up such notices in English.
Exactly.

The print era exhibits in the proposal documents all clearly show that 
the digit is placed next to the consonant, which seems to be the 
classical convention.  It should be considered important to find out 
what the actual users think about this.  If the digital era practice of 
placing the superscripts at the syllable final position is considered a 
temporary work-around, that's one thing.  But if the users consider this 
to be a new, digital era convention which supersedes the classical 
convention, that's something else.  In which case the question becomes 
should both conventions be representable in computer plain-text?  If 
yes, can this be accomplished without changing the /de facto/ standard 
encoding order?  I think it *should* be possible, but that doesn't mean 
it is.  And perhaps the users (and the UTC) would prefer to represent 
both conventions at the encoding level instead of handling it at the 
display level.  I don't know and do not claim to speak for either the 
user community or the UTC.  Just asking questions in an effort to 
understand the issues involved.

> Additionally, I’m not sure that a font that works on Internet Explorer
> was possible.  I’m not even sure that an OpenType font can be made for
> Chromium or Firefox.  Can one even kern Tamil syllables and superscript
> digits using OpenType?
It should work, but I'm not set up to test it at the moment.  If it 
didn't work, I might try "rlig" and make precomposed glyphs 
accordingly.  Even if that meant scads of precomposed glyphs.

-----

(some background)
In the 2010 proposal, 
https://www.unicode.org/L2/L2010/10256r-extended-tamil.pdf , it was 
explained that the proposed characters would have considerable glyphic 
variation.  The code charts used the Tamil forms with western 
superscript digits for reasons explained within the document.  The 
objections to the proposal were essentially that the glyphs being 
proposed as characters could already be represented as sequences and 
they worked just fine.  (As long as one either liked dotted circles or 
was comfortable with the digits appearing syllable final, but this part 
was not mentioned in the objections.)  Even though the proposal 
explained why the proposed characters should not have decompositions.  
If the proposal had used the Grantha style glyphs in the charts for the 
proposed characters, the objections probably would have been that the 
proposed characters were already encoded in the Grantha range.

For example, IIUC, the Tamil form looks like "த³" and the Grantha form 
looks like "𑌦" for the character proposed for U+xx10, TAMIL LETTER DA.

If Shriramana Sharma's 2010 proposal (and revisions) had been accepted, 
we would not be having this discussion.  But here we are.