isn't computation-rules redundant? Alex Shinn 23 Mar 2004 03:21 UTC

I have the same pet-peeve with {define,let,letrec}-syntax and
syntax-rules, but in

  (define-syntax-computation foo
    (computation-rules () ...))

computation-rules is a noiseword.  You could just as easily write

  (define-syntax-computation foo ()

which is shorter and more to the point.  Likewise for the let- and
letrec- versions.  Of course, you do still want computation-rules by
itself for anonymous computations, much as people have made the argument
for syntax-rules by itself.

A possible reason for requiring the explicit computation-rules in the
define- and let- forms is if you want to allow people to use alternate
forms for computation-rules.  For instance, you could "alias" syntax

  (define-syntax-computation foo bar)

or possibly generate something dynamically with

  (define-syntax-computation foo implicit-syntax-inspect-form)

but currently the reference implementation doesn't allow this and these
extensions could be added by pattern matching regardless of whether or
not you require the computation-rules.

Am I missing something?