[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] LAMBDA and Unicode (Was: [r6rs-discuss] Scheme pattern matching & R*RS)

On Wed, Dec 15, 2010 at 11:14 AM, Shiro Kawai <shiro@x> wrote:
> Two points regarding using lambda and characters beyond unicode.
> I don't object using extended character sets within the source
> code, but is it completely safe to move on to Unicode-topia?
> I mean, don't you still have text files in Latin-1?  I do have
> plenty of legacy EUC-JP text files around.

Americans use US-ASCII almost exclusively.  One
exception is foreign contributor's names in comments.

I actually skimmed through my code and found out that
my match.scm has "Ludovic Courtès" in a comment and
my default environment saved the file in EUC-JP.

But since this is the uncommon case we should be more
concerned with what non-English speaking programmers are

> For example, there are some languages (e.g. Python),
> Scheme implementations (e.g. Gauche) and editors (e.g. Emacs)
> that recognizes a line like "-*-coding: utf-8-*-" near the
> beginning of the plain text file.

The standard currently says nothing about the CES of
source files (though we should, I'll add a ticket for this).
I think it's implied you should just convert them all to UTF-8
for portability.

The advantage of the Emacs convention is it allows other
encodings (Scheme may outlive Unicode), and may coincidentally
work on some existing files since Emacs is so popular
among Schemers.  But it seems a little hacky to put into
the standard.


Scheme-reports mailing list