RE: Linux & Unicode

From: Addison Phillips (
Date: Mon Dec 07 1998 - 13:05:32 EST

I agree with Glen.

The UNIX development tools *are* important to deal with. There is a vast
amount of plumbing work in Linix/UNIX that needs work to be truly
Unicode ready (see the recent OS 99 thread on this list). All those
configuration files in /etc, all those basic utilities (sed, awk, lex,
...): UNIX is rife with multi-byte and Unicode processing problems
simply because of its heritage. One of the founding design priciples,
after all, is that *everything* is a text file!

X-Windows, however, is a client/server application similar in many ways
to a Web browser. All of the major Web browsers are architected to use
Unicode internally because they have to deal with input streams from
disparate servers running in locales that they cannot control and output
streams to a display device that is usually tieds to a code page or
character set *other than* Unicode. They need, therefore, to be able to
store and even potentially to render the data without help from the
underlying operating system.

If Qt and KDE become entirely Unicode internally then the developers
will reap benefits even in a single-locale environment, since they will
only have to worry about character set conversions "at the borders" and
because their code will no longer be code page dependent (just how many
multibyte processing algorithms do YOU want to code?). Just because the
operating system is not entirely Unicode-ready doesn't mean that the
windowing system should wait.

Perhaps it will stimulate developers to write those other bits we've
been waiting *forever* for...



Addison Phillips
Director, Globalization Consulting
SimulTrans, LLC

"22 languages. One release date."

-----Original Message-----
From: Glen []
Sent: Friday, December 04, 1998 1:22 PM
To: Unicode List
Subject: Re: Linux & Unicode

-----Original Message-----
From: Markus Kuhn <>
To: Unicode List <>
Date: Friday, December 04, 1998 4:34 AM
Subject: Re: Linux & Unicode

>It is most important that the most basic Unix development tools such as
>xterm, emacs, vi, grep, bash, etc. become fully UTF-8 compliant first.
>It makes little sense to focus on new fancy widget sets like Qt and
>first before we haven't introduced UTF-8 in the classic basic tools.

Actually, I think it makes little sense to wait on pressuring Qt and
GTK+ to
go Unicode until all of the legacy command line tools have been
first. That's like trying to lift a car using the wrong end of a lever.

GNOME and/or KDE may do for Linux what the Web did for the Internet:
make it
popular among average consumers, who aren't generally fond of
interfaces. If all the GUI apps on Linux were Unicode-based by default,
they were the apps that really drove the widespread adoption of Linux,
there would be tremendous pressure on the developers and maintainers of
legacy command line tools to upgrade them to Unicode.

I'm hardly advocating that you stop doing the wonderful work you're
doing. I
just think that getting the Linux GUIs to be based on Unicode as early
possible in their lifecycles, before all the (future) popular GUI apps
saddled with single-byte legacy baggage, is probably a useful goal.

__Glen Perkins__

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:43 EDT