🚀 go-pugleaf

In article <1i13rrINNars@rodan.UU.NET> avg@rodan.UU.NET (Vadim Antonov) writes:

>Somehow i had introduced my "own" cyrillic encoding and it lived (and
>still lives) and was used by a lot of people (it became obsolete nearly
>at the time it was approved as a national standard). The company i worked
>before managed to create a totally bilingual family of Unix-like systems --
>something unheard of in the West.

You're uninformed.  Try looking at Xerox Viewpoint (West European,
East European, Greek, Arabic, Hebrew, Chinese, Japanese, Korean, et al);
try looking at BB&N Slate (West & East European, Greek, Hebrew, Arabic,
Korean, Thai).  There are others: Nota Bene, Multilingual Scholar,
AlKatiib, etc.

Tell me how to sort Tibetan, Burmese, Khmer, and Japanese, providing
all established culturally expected orderings.  After you've done
this, let me know whether you think a character set can handle the
burden of correct sorting.  Its not that this isn't an admirable goal;
its just that it is untenable given the overall design goals
of Unicode -- or any other character set that could stand a chance of
becoming a universal character set standard, and not being horribly
inefficient.

You might want to read a paper entitled "Collating Unicode Data"
by Alain LaBonte, Ministry of Communications, Quebec, before
arguing further.  It is contained in the proceedings of the Unicode
Implementors Workshop.  It will verily destroy all the arguments
you've offered so far.

If you are interested in a paper discussing the architecture of a
multilingual toolkit in which sorting *is* handled, and which *does*
support Unicode, and which *does* facilitate language tagging, then
see the file /pub/mlt2-overview.ps.Z (a compressed postscript file) on
METIS.COM [140.186.33.40] by anonymous FTP.  [If you have no FTP access,
let me know and I can uuencode and email it.]

Glenn Adams