> >Pali & Tamil. What SLS 1134 did was they defined a standard (I would
rather
> >call it guidelines) for Sinhalese language font designers.
> >
> I would say, we defined how Sinhala should be encoded in Unicode. This
> would be used by font designers, display driver designers, etc.
mmmmm... SLS 1134 Sinhala encoding standard defines how Unicode Sinhala
encoding standard should be used for encoding Sinhala? :-)
> >They used ZWJ as an control character to enforce certain rendering
> rules, which totally
> >violates the purpose of a phonetic system.
> >
> No. it uses ZWJ to indicate when and how adjacent characters should be
> joined.
take 'ksha' -> ka halant sha, is a ligature. The lookup for this ligature
should come under Opentype feature 'akhand', which means unbreakable. So
why would you want to place a ZWJ to join ka & sha? I have never come
across a unicode script which uses ZWJ to indicate how *every* adjacent
characters should be joined.
Dont you think bandi akuru is a style of writing? which should be handled
by the opentype font rather than at encoding level? If I can remember,
there were two different ways of writing 'dya', will you use two ZWJs to
differentiate between to two styles or let the opentype font have two sets
of glyphs?
Quoted from opentype spec by adobe :
http://www.adobe.co.uk/type/opentype/main.html
"OpenType feature support Central to a discussion of OpenType feature
support lies the distinction between characters and glyphs. Characters are
the code points assigned by the Unicode standard, which represent the
*smallest semantic units of language*, such as letters. Glyphs are the
specific forms that those characters can take. One character may correspond
to several glyphs; the lowercase "a," a small cap "a" and an alternate
swash lowercase "a" are all the same character, but they are three separate
glyphs. One glyph can also represent multiple characters, as in the case of
the "ffi" ligature, which corresponds to a sequence of three characters: f,
f and i. OpenType layout features can be used to position or substitute
glyphs"
> Whether you want a pure consonant (without a vowel) to be written
> touching the next letter, joined to the next letter, or with an explicit
> al-lakuna, is your writing style, and independent of the language used.
The whole point is you should not encode style.
> SLS1134 supports all three writing styles, and you can mix them in the
> same document so an author can get the effect he wants.
As you correctly said, Pali & Sanscrit are writing *styles*, therefore IMHO
should be handled by a layer on top of encoding. I admit, there are certain
drawbacks in what I am suggesting but SLS 1134 solution is not the most
elegant either. I wont rant on this, since the majority of the community
agrees with SLS 1134 :-) .. but I thought its better I share my concerns
now rather than later :-)
> >Once again I would like to emphasize that the contents in SLS 1134
should
> >go into the unicode book since they have similar guidelines for each
script
> >
> >
> It will be forwarded as a tech note.
that means we dont have to buy a copy of SLS 1134? Super!!!
regards,
Harsha.
This Mail Has Been Scanned For Virus By Scanmail For Lotus Notes
_______________________________________________
Sinhala mailing list
Sinhala@???
https://secure.linux.lk/mailman/listinfo/sinhala
(text/plain)