Re: [Indic-computing-devel] [review-request] Indic computing handbook: devanagari

[Dr. U.B. P. <pav...@vi...>]

> About 'phonetic' representation of letters of the
> script : though there is no 'standard' to express
> Indian languages/scripts in Roman phonetic script,
> several developers/apps have been following some
> 'unwritten' standards (itrans, varamozhi, baraha etc
> etc). Just check out those and attempt to synch up
> with sync up with the current 'unwritten' stds (like
> giving alternate representations).  

There is a representation available at TDIL web-site. I am not 
sure whether they have notified it and made it a national 
standard (BIS) like ISCII.
              
> What is vowel #8 (ee) ? It's just after 'e', so I am
> assuming it's the elongated version of 'e', but I am
> not sure if it exists in the languages that use
> Devanagari (AFAIK, only South Indian languages have
> it). Same Q with vowel #10 ('oo'). Actually 'oo' is a
> nice representation for the elongated version of 'u'
> (like in moong dal). Can't understand vowels 11 and
> 12.    

Yes. These are mostly used by South-Indian languages.
 
> Several consonants (mostly represented by the original
> consonant with a 'dot' below) have been added for
> Urdu. Pls mention them. 

They are called Nukta variants. Unicode chart for Devanagari 
lists them as separate characters. In the latest Unicode for 
Kannada, Nukta has been addded. In case of Kannada Nukta is a 
separate character and not every Nukta variant of consonants 
have been mentioned as done in Devanagari (this is appropriate). 
Probably their idea is to give them separate position 
(Deavanagari) in the collation chart, which makes sense. 
 
> Consonant list - 'ksha', 'tra' and 'Gna' are missing
> (actually I prefer the Roman phonetic equivalent
> 'thra' instead of 'tra').

They are actually combinations. You can get their independent 
display forms by using the GSUB rule of opentype font (the 
lookup Akhand).
 
> Pls mention that 'mathras' - that is half-consonants -
> are graphically 'partial' consonants and they add on
> before/below the 'base' consonant to generate
> 'samyukthaaksharas'.  

"matras" actually refer to "vowel signs" and not "half-
consonants".

There is no place for half-consonants in Unicode chart as it is 
for the opentype font to take care of these extra glyphs needed. 
 
> Conjuncts with multiple combinations (C*V) - you have
> shown theoritical possibilities, but just about a few
> thousands are practically legal/valid. Would be better
> to mention it.

The total number of possible combinations of V,C,CV,CCV, etc. 
exceeds 15,000! 
 
> Specials - in addition to anuswar, I think visarg
> (like the ':' symbol) needs to be mentioned.

Also chandrabindu (0901).
 
> Double vertical bar - one of the vedic specials (e.g
> used at the end of a shlok)

Its Unicode value is 0965.

In general, better to look at Unicode chart and see whether all 
of them  are covered.

Rgds,
Pavanaja
-----------------------------------------------------
Dr. U.B. Pavanaja
Editor, Vishva Kannada
World's first Internet magazine in Kannada
http://www.vishvakannada.com/

Note: I don't worry about pselling mixtakes