Re: Unicode support

[Eli Zaretskii]

On Wed, 18 Jul 2001, Werner LEMBERG wrote:

> The reason for not working on Unicode support is that we all wait that
> Emacs 21 is out.  It has the necessary capabilities for real Unicode
> output -- there were so many internal changes to almost everything
> that it would be hard to develop something in parallel.

I think you are wrong: Emacs 21 does not have any capabilities Emacs 20 
didn't have, and the massive internal changes are mostly related to the 
display engine, much less to Mule.

As far as Mule is concerned, the goal of Emacs 21 development is 
stabilizing it and making it more user-friendly and less prone to 
problems, not introduction of nifty new features.  (That wasn't exactly 
planned that way, but that's how it turned out eventually.)

The Unicode support in Emacs 21 is based on 3 additional charsets (which 
cover part of the BMP), but it doesn't change anything in the internal 
design of how non-ASCII characters are treated and represented.  As a 
consequence, if you use Unicode, you are trapped: you can only read and 
write UTF-8 or Latin-1 text, but you cannot encode in anything else; and 
if you mix Unicode with non-Unicode text in the same buffer, you are in 
for a bumpy ride.  (Add-on packages such as Mule/UCS and local hacks can 
alleviate some of these problems, but I'm talking about stock Emacs 21.1 
and an average user that isn't privy to the charset intricacies.)

So there's no reason to wait for Emacs 21 or hope that it will solve any 
substantial problems on the way to switching the internal representation 
to a Unicode-based one.