RE: [idn] I-D ACTION:draft-ietf-idn-cjk-00.txt

[Karlsson Kent - keka <keka@im.se>]

> -----Original Message-----
> From: James Seng [mailto:James@Seng.cc]
...
> If han folding is difficult, then zVariant folding is 10x 
> that and bring it to
> infinity. (No, I am not saying it cannot be done...we done it 
> but it is incredible time consuming...)

If I understand correctly, z-variants is the font variation
for a SINGLE UCS CJK ideographic character (and still listed
separately in 10646-1:2000; though never listed separately
in the Unicode standard).

That is, z-variants are the characters *already unified* by
the IRG.  There is a small number of z-variant *duplicates*
in Unicode/10646 that have resulted from the (now dropped)
'source separation rule', and, as far as I understand, they
are listed in 10646-1:2000 annex S.3 (the heading is "Source
code separation examples"; but I think that is a typo,
and should say "Source code separation exceptions").

I'm not saying that such folding on the Z-variant
*duplicates* should be done; I'm just questioning your
"10x" and "to infinity".

Or do you mean something else with "z-variants"?
Z-variants are font variations of the same character.
The Y-variants are different-looking characters
with the "same" meaning; I don't know how language 
dependent that is.

> % In alphabetic scripts, there is also requirement to fold Latin, Greek,
> % Hebrew, Cyrillic, Hebrew and Arabic together. 

Well, high quality transliteration/transcription may
sometimes be desirable. But *not* for IDN, I would think,
in particular not automated. Note that such transliterations
or transcriptions depend on the target language, and
possibly other factors as well.

		/Kent Karlsson