[plt-scheme] Why do MzScheme ports not respect the locale's encoding by default?

From: Alex Shinn (foof at synthcode.com)
Date: Sat Feb 26 15:17:00 EST 2005

At 25 Feb 2005 14:51:58 -0500, Jim Blandy wrote:
> Alex Shinn <foof at synthcode.com> writes:
> > At 18 Feb 2005 20:34:37 -0500, Jim Blandy wrote:
> > > 
> > > [...] do you think most users want applications to respect locale
> > > settings --- and specifically, the character encoding --- or do you
> > > think most users would like applications to ignore them?
> > 
> > Most users don't even know what character encodings are and could care
> > less.  They do, however, want software to interact together, which
> > causes complications.
> Okay, maybe my question is a little like asking whether users prefer
> little-endian or big-endian byte orderings.  To rephrase it more
> meaningfully: are most users better served by using the current
> locale's character encoding, or by ignoring it and always using UTF-8?

This is indeed the real problem.  You have to choose between backwards
compatibility (with no means of breaking away) and consistency and
forwards compatibility.  Since Scheme places less emphasis on
compatibility and more emphasis on doing things right, I think the
latter is preferable.

Personally, I just use automatic encoding detection these days when
the format doesn't specify the encoding.


Posted on the users mailing list.