Re: reinventing ASCII?

James Harris <james.harris.1@[EMAIL PROTECTED]
> writes:


> In the context of this newsgroup - i.e. language design - my belief is
> that a language would benefit from independence from the character
> representation. In other words I'd say ASCII is not pragmatically
> better or worse than any other representation EXCEPT I would say that
> the non-printable characters are a pain.
>
> Which non-printable characters are still needed? How about
>
>    0  NUL
>    4  EOT (end of transmission)
>    9  TAB (or field separator)
>    10 LF (or end of line)

Why would we need a null character?  I know that it in C is used as
end-of-string marker, but that is a horrible design decision.  If any
single character should mark the end of a string, it should be EOT.

As for TAB, I would standardise the tab positions to be 8 characters
wide.  And I would, as you imply, keep a single newline character
instead of the mess of incompatible combinations of CR and LF.

> Any more?

Page feed might be useful.  At the moment, it also doubles as clear
screen, which is a bit dubious.  ASCII was clearly designed for paper
terminals, so it is not really prepared for erasable displays.  Hence,
some control codes have been modified to mean different things on
printers and screens.  This is a bit of a mess, so this should be
cleaned up.

> Of course, only one or two control chars are really needed,
> perhaps one to say that the next byte is to be interpreted as a non-
> printable (perhaps a control code) and one to introduce a composite
> character.

Yes, that would work.  Backspace used to be used for creating
composite characters by moving the carriage back and striking on top
of the previous character, but these days it is used to delete the
previous character.  So a control code that explicitly merges the
following two characters would be better.

	Torben