# Re: Character and output-width proposals

```On Tue, 2 Jun 92 11:33:48 -0400, "Aubrey Jaffer" <jaffer@martigny.ai.mit.edu> said:
>From: david carlton <carlton@husc.harvard.edu>
>>From: Aubrey Jaffer

>>>   Proposal:

>>>   char-code-limit					constant

>>> Is an integer 1 larger that the upper bound of the values which can be
>>> returned by char->integer.

>>>   char-code-lower-limit					constant

>>> Is the lower bound of the values which can be returned by
>>> char->integer.

>> I see no reason why these numbers should exist at all.

> Suppose you are writing a string search program and you want to
> dispatch on a character through a table (vector).  You need to make
> the vector the right size or you may get (or worse, not get) runtime
> errors.

That's not my point.  I don't understand why there should be such an
arbitrary restriction on characters.  I can imagine that it is useful,
but I can also imagine that lots of arbitrary restrictions would be
useful; I just think that they typically raise more problems than they
solve.  And I certainly wouldn't write programs that do things like
the above and expect them to work - what if your implementation does
have a finite number of characters, but they are 32-bit or 64-bit
characters, so such a vector won't fit in your finite address space?

> My example of Grobner base ordering is to assign numbers to objects in
> such a way that the ordering has certain properties.  One of these
> properties is that the first item in a list has more priority than all
> others.  In order to do this I have to have a bound on objects based
> on characters (like strings and symbols).  In order to do this I need
> to know the bounds for characters.

Would a lower bound be sufficient for this?  Having a lower bound (of
0, presumably) would make more sense to me than having an upper bound.
That would at least give you a natural well-ordering on the set of
strings.

david carlton
carlton@husc.harvard.edu

I know things about TROY DONAHUE that can't even be PRINTED!!

```