From: N. Ganesan (naa.ganesan@gmail.com)
Date: Wed Apr 19 2006 - 09:42:07 CST
Richard W. wrote:
>There's a complication in using nakarapollu for final /n/ in
>the word 'pooyen', now written పోయెన్. Apart from
>the issue of how 0C2A 0C4B 0C28 0C4D 0C2F 0C46
>(PA + OO + NA + virama + YA + E) should be rendered with a
>font etc. supporting nakarapollu, there is an argument that,
>as the former could be read as 'poonye', that one should
>encode the nakarapollu-using 'pooyen' as
>0C2A 0C4B 0C2F 0C46 0C4D 0C28 (PA + OO + YA + E + virama + NA).
>The argument is the precedent of Khmer (TUS Figure 10.3),
>where [ha@i] 'already' is encoded as HA +OE + coeng + YA
>rather than HA+coeng+YA+OE. (There is a difference in this
>case *unless* the rendering is wrong - the vowel appears above
>the HA, whereas for a stack with a spacing subscript
>consonant superscript vowels normally align with the superscripts
>ascender on the hanging baseline.) A related argument is the
>dictionary order - PA + OO + YA + E + virama + NA would
>sort close to PA + OO + YA + E + NA + virama, the modern spelling.
Richard, I don't think the tradition of spelling in Telugu
words need to be changed, code point sequence should
reflect Telugu dictionaries, is it not? Take Sanskrit
words, brahma, brAhmaNa 'brahmin', subrahmaNya 'Lord Murukan
of Dravidian culture', vahni 'fire' - they are spoken out
as bramha, brAmhaNa, vanhi. But still the spellings do not
change in Indian scripts.
----------------------
Let us say we have in (future) Unicode
MALAYALAM SIGN CILLU with a dotted circle in front
(cf. malayalam sign virama)
which can be annotated as
" . works upon ka, na, nna, la, lla, llla, ra, ta and ma
. can occur word-medially or word-finally"
Because Telugu nakarapollu (I think Kent said
he does not know the meaning of n., it's 'pulli on letter NA' )
is Telugu equivalent of cillakSaram,
a Unicode code-point in Telugu codechart will be
TELUGU SIGN CILLU at a codespot parallel with
MALAYALAM SIGN CILLU.
(cf. tamil sign virama, using a sanskrit name.
cillu 'small/fragment etc.' is common all
4 Dravidian languages and has cognates).
The future TELUGU SIGN CILLU codepoint
can be annotated (something like) as
" . works upon letter NA
. called nagarapollu "
have given some 25+ cillaksarams attested
in Dravidian language scripts, and the cillu
sign will come in other Dravidian scripts also.
An appropriate and parallel naming convention
is after danda sign disunification in Unicode:
DEVANAGARI DANDA
DEVANAGARI DOUBLE DANDA
BENGALI DANDA
BENGALI DOUBLE DANDA
TAMIL DANDA
TAMIL DOUBLE DANDA
KANNADA DANDA
KANNADA DOUBLE DANDA
ORIYA DANDA
ORIYA DOUBLE DANDA
etc.,
-----------------------------
A.Leca>> How does it deal with valapalagilaka
>>(the old-fashioned repha-like)? Another
>symbol? ZWJ?
NV> Possibly. If you are going to design a
>special symbol, this is a very good scenario -
>to signal something that is rarely seen.
I agree. Overburdening joiners is not the solution,
one problem is they get ignored in collation
(Eg. K.SSA conjunct or non-conjunct with zwnj
in Tamil words are the same words, hence can
be treated equivalent.) But not the valapalagilaka,
nakarapollu - their occurence in a Telugu text,
tho' less frequent, is important and needs to
be diffrentiated.
A codepoint in Telugu codechart with a
valapalagilaka mark is recommended.
N. Ganesan
This archive was generated by hypermail 2.1.5 : Wed Apr 19 2006 - 09:49:51 CST