Fri Aug 14 19:47:53 EDT 2015


If it's so easy to translate between (r (a, b)) and (r a, r b), why
should I have to?

Somehow it seems that the plumbing necessary to pack/unpack between
these two can be avoided.


The difference is visible in Eval quite clearly:
[(a,b)] is not the same as ([a],[b]).

The former definitely seems more appropriate.

What about this:
- internally, use currying and tupling
- externally: curry and pair input/output

This is about "canonical forms".  Do this automatically.