Markus,
We actually distribute freely such an API that can provide this
information, however it is a binary, and currently only compiled for
Windows systems. It is easily compilable on Unix systems. The
advantages of a binary are, of course:
1) Size
2) Speed
3) Ease of use
And users/developers who want such a table can always generate one from
the API.
Keep in mind that the current database skips over tens of thousands of
CJK characters that have names, but are not listed because they can be
algorithmically generated. So the current 1/4 megabyte of character
names would increase to maybe over a megabyte if listed in a plain
ASCII text file. We've managed to compress all this into 80K of data
by putting it into an exe. It's much quicker to retrieve this way as
well....much, much quicker!
Additionally, conversion routines (efficient ones that is) can be a pain
in the rear (or at least a nuisance) to implement, and the burden would
be left up to the developer.
Dean
> What would be extremely important is to define a file format for a
> standard /etc/unicode database that we hopefully can soon expect to be
> installed on every Unix/POSIX workstation. Many applications (editors,
> debuggers, terminal emulators) will need a table that maps Unicode codes
> into names. If I shift-click on a character A in xterm/emacs/etc., I
> want to see a tiny window pop up that tells me whether A is "cyrillic
> capital letter a" or "greek capital letter alpha". Such a table should
> not be compiled into binaries, but it should be globally provided as a
> database file by standard distributions, and ftp.unicode.org should
> provide the latest update version whenever the standard is extended.
--
______
O O
----------------------------oOOo-(__)-oOOo----------------------------
Dean Abramson E-mail: dean@bjondi.com
Bjondi International Tel: (310) 395-3820
1234 Third Street Promenade #220 Fax: (310) 395-3402
Santa Monica, CA 90401 http://www.bjondi.com
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:43 EDT