Re: [tlug] Re: Piping stderr?

[Jiro SEKIBA <jir@example.com>]

At 24 Jun 2002 14:32:56 +0900,
Stephen J. Turnbull <stephen@example.com> wrote:

>>  So I believe I18N software can satisfy both people who
>> don't want I18N and people who want.  Isn't it nice?
> 
> But we're not talking about I18N software.  We're talking about
> "multibyte patches."

 ??? What do you mean it, you meant localozed hard coded patch?
Then you are absolutely right.  If not, I don't understand what you
mean exactly.   If you think of I18N, it includes multibyte capability.

 I'd like to ask you that if I18N software is acceptable for you? 
Why/Why not?

> Software should be written to be general, with as little dependency on
> external text representation as possible, and the I18N facilities
> should be provided by robust, specialized libraries.[1]  The usual
> "multibyte patches" are exactly the opposite; they provide specific
> facilities in idiosyncratic ways in each application.

 Yes, you are right, and that IS I18N CSI(Code Set Independent) software.
Strip Encode/Locale dependent stuff from software, and use proper
library functionalities.

> This has cost the Emacs, Ghostscript, Pine, ... communities in Japan
> dearly.

 I have never seen GhostScript/Pine I18N version, but only Japanized
patch.  It's not good, I have to say.  But in some cases, it's acceptable.

 About emacs, bit difficult, coz emacs itself is lisp environment.
So it's proper away from CSI I18N design, which makes very portable
as a result.  It works fine even system doesn't have I18N functionality.

-- 
Jiro SEKIBA | Web tools & AP Linux Competency Center, YSL, IBM Japan
            | email: jir@example.com, jir@example.com