Character and output-width proposals

[Scott L. Burson <eros!gyro@drums.reasoning.com>]

   From: Richard Mlynarik <mly@adoc.xerox.com>
   Date: Wed, 3 Jun 1992 15:56:45 PDT

   I see no reason why CHAR->INTEGER and INTEGER->CHAR should exist at
   all.  They are and can be of no portable use.

Not so.  One could, for instance, create a vector mapping characters to other
characters, without requiring any property of the mapping between characters
and integers other than that it be one-to-one.

Beyond that, character arithmetic and comparisons are sometimes quite handy.
I think it would be worthwhile to specify that the mapping between characters
and integers should always be that of ASCII, even when the host system uses a
different mapping; so it's up to the Scheme implementor to perform any
necessary translations.

(More precisely: any character that is a printable ASCII character, plus the
usual handful of control characters, should map into its ASCII equivalent.
This is not intended to prevent implementations from adding non-ASCII
characters.)

-- Scott