[sword-devel] Re: Westcott-Hort
Costas I. Stergiou
Tue, 6 Apr 2004 09:51:22 +0300 (EET DST)
> It sounds like you know what you're doing. My only concern was that the
> Greek Extended area was categorized as compatability or presentation,
> in which case they might not be canonically equivalent to decomposed
> codepoint sequences. But if you're doing NFC normalization already,
> then obviously they are.
I did not know that the greek extended area is categorized as
compatibility (I should take a look at what exactly this means). I
can positively affirm though that taking a fully decomposed text (only
standard greek & combining diacriticals) and using the icu4j NFC turns it
all in all-precomposed chars. I can alse affirm that no app (either IE,
MS Word, Sword, etc) can correctly (or at least nicely) present greek with
combining diacritics. That's why many many greek texts that
do not use
pre-combosed characters use some diacritics that are not combining (just
to get a decent visual output) which I think is worse.
Actually, the test accented WH module uses mostly precomposed characters
(but it does resort to some non-combining diacritical in places to get
nice visual output).
> sword-devel mailing list