[ On Friday, October 15, 1999 at 18:22:08 (+0900), Noriyuki Soda wrote: ]
> Subject: Re: Unicode support in kernel
>
> Universal charset concept just doesn't work. :-<

I spent a lot of time a few years ago (about 1994) trying to understand
all of this stuff and all of the related issues.  I've yet to see
concrete technical reasons why Unicode (or indeed any similar
all-encompassing universal charset concept), together with UTF of
course, cannot be made to work.  Certainly it doesn't come for free, and
indeed some locales end up spending a lot more effort and paying more
for ongoing resource needs than others.  Remember that Unicode is still
evolving to meet previously unforseen needs too!  Everything I've seen
about it since 1994 suggests that it just keeps getting better.  In fact
so far as I can tell Unicode (or something almost identical to it) is
the only way we will ever enjoy the ability to store phrases from all
languages in one single document that can be shared universally amongst
homogenous systems everywhere.  Generally the arguments I've seen
against Unicode over the years have demonstrated that their proponents
are at least somewhat ignorant of the actual specifications, etc.,
practical arguments aside of course (such as lack of fonts and/or
current difficulties with easy translation back to other charsets).

> We have a lot of trouble with this, and I strongly believe that
> we have to make kernel policy-free.

I certainly agree with keeping the kernel policy-free in this respect
though!   :-)  [Which is why I'd rather see either only strict Unicode
support in the kernel, *or* no special charset support at all.]

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods@acm.org>      <robohack!woods>
Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com>