Case-sensitivity goals, policy and implementation



On Wed, 2004-10-13 at 08:59, Stavros Macrakis wrote:
> I don't understand the proposal. 

And I don't understand your difficulty with the proposal. One of us must
be missing something. (It still could be me!)

>  Why do we need to futz with the
> internal representation of Lisp symbols at all?  All we care about is
> the correspondence of *Maxima* symbols and Lisp symbols.

Because Lisp would do everything for us.

> And the reading of Maxima symbols is 100% under our control.   Let
> Lisp keep its default behavior, which I guess is that "foo" reads as
> 'FOO, and map the Maxima symbol "foo" to "FOO", and the Maxima symbol
> "FOO" to "foo".
> 
> One way of doing that is using the Alias mechanism as I suggested
> before, another is to change the read/print code.  Changing the code
> is the cleaner way to do it, but requires some careful code reading in
> suprv1, commac, nparse, and displa (I think those are the only places
> that need to be checked).  Alias is yucky, because you are
> special-casing built-in symbols, but might be more expedient in the
> short run.

Because Lisp would do all that for us automatically. What you are
proposing is to manually re-implement the effects of readtable-case
:invert in the maxima code. Why bother?

--Jim

P.S. Per