Re: Character and output-width proposals

[david carlton <carlton@husc.harvard.edu>]

On Thu, 4 Jun 92 11:24:40 -0400, "Aubrey Jaffer" <jaffer@martigny.ai.mit.edu> said:
>    From: tmb@idiap.ch  (Thomas M. Breuel)

>    In general, however, I'd prefer if Scheme supported efficient binary
>    I/O via some essential procedures. The current set of standard I/O
>    procedures, both text based and for data structures, is at best
>    adequate for writing toy programs.

> Yes.  And this is also tied in with CHAR-CODE-LIMIT.

Why should binary I/O have anything inherent to do with characters?
It would seem to me that, to do this right, you'd want to have some
sort of machine-word object, or something like Ken Dickey's abstract
stores (proposed for this purpose), or maybe bitvectors.

> I have written routines which convert integers to 4 byte strings.
> It is not portable because it depends on 8 bit characters.  But if
> CHAR-CODE-LIMIT were defined, it could be portable.

What would the portable version do?  It can't convert integers to 4
byte strings of characters if 2^32 isn't a power of the number of
characters.  I suppose it could convert them to strings of 4
characters, but I'm not sure why I would want to do that.

david carlton
carlton@husc.harvard.edu

       My polyvinyl cowboy wallet was made in Hong Kong by Montgomery
       Clift!