On Fri, 7 Jan 2005, Yukihiro Matsumoto wrote:
> |Do I need to submit a RCR for those, or should I go for two of them ?
> I think a fixnum is not enough since a character may not be
> represented by a single codepoint, e.g. character composition, or
> surrogation. Besides that, a character is represented by combination
> of codepoint(s) and

questions:

1. are those features of unicode ?

2. are those required for a support of Japanese that make sense ?

Those are not rhetorical questions. I don't have the answers nor the
background to attempt tackling them. I didn't know that those features
were important to you, and I barely ever hear about them.

Character composition, as I know it, is a hack to make accented letters,
underlining, bold, and such, by interleaving the special code 8
(backspace) between characters to be composed, destined to be used with
daisy printers and TTY's, and already became dated when TTY's got replaced
by CRT monitors during the 1970's. It stemmed from previous usual practice
in trying to type French on an American typewriter (and other similar
situations in other countries)

ISO-Latin-1 (that I normally use) contains all the self-contained
codepoints necessary to write in many European languages, and as such, you
never need composition in those languages with that charset, and as far as
I can tell, all other charsets are designed with that in mind, including
unicode. So it is difficult for me to believe composition is required.

Is this because of the way Japanese works? (then if it's so specific to
Japanese, why hasn't Ruby gotten that feature a decade ago ?? I'm _really_
puzzled)

_____________________________________________________________________
Mathieu Bouchard -=- Montr٬l QC Canada -=- http://artengine.ca/matju