> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]]On Behalf Of Markus Kuhn
...
> Jungshik Shin wrote on 2002-02-23 22:30 UTC:
> > In addition, due to its another
> > not-so-insightful decision, whatever NF we use, we still are left with
> > multiple representation of Hangul syllables as Kent noted.
>
> Well, you can lobby with the Unicode consortium to at least formally
> define a Normalization Form J that uses only Jamos. Costs a factor three
> memory, not that it really matters in practice though.
I have just submitted a paper on this to the Unicode consortium.
Copy available upon request.
The precomposed syllables are unneeded, but not the major problem. The
major problems are
1) the lack of canonical decompositions of letter cluster jamos
(these are also unneeded), and
2) the 'not-always-proper' compatibility decompositions of
Hangul compatibility letters and compatibility cluster letters.
Note that just applying NFKC or NFKD on Hangul compatibility letters
wreaks havoc on strings with such characters. Instead most processes
(extended normalisations) should use the "free-standing" mappings
(see my paper) instead of the Unicode compatibility mappings for
Hangul compatibility letters and cluster letters.
Kind regards
/kent k
--
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/linux-utf8/
