Re: [tlug] Re: Piping stderr?

["Stephen J. Turnbull" <stephen@example.com>]

>>>>> "Jiro" == Jiro SEKIBA <jir@example.com> writes:

    Jiro> please forget about UTF-8 apps as well, such are like
    Jiro> uxterm.

I disagree.  These have their place.  They _can_ handle all text in
all languages.  (What with the planes 2 & 3 damage, you can even
distinguish between kanji and hanzi without plane 14 tags or higher
level markup mechanisms.)

But it's crucial that they _not_ break when handling text/plain.  One
way to make that happen is to say "you can have any character set you
like as long as it's encoded in UTF-8."  I think that shells, consoles
(and *terms intended to basically be like consoles), and scripting
languages SHOULD (in the sense of RFC 2119) behave that way.  Yes,
scripting languages should provide IO filters which are convenient for
people who only ever use Shift JIS or KOI-8 or ISO-8859-15.  But
internally, nothing but a UCS (and experience with Mule says it best
be Unicode).

Wordprocessors, DTP, browsers, etc, where having content is a severely
deprecated practice and Maya is all, that's different.  But I don't
use those.  I follow the Tao.<wink>


-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN

       Shift JIS in text processing internals is like boots on tatami.