Mon Nov 15 18:55:07 EST 2010

Partial evaluation

Is the current greedy approach actually smart, or should reduction be
defined in a different way?  Find a good explanation for the two

The good part is that partial evaluation is expressed as simple and
straightforward evaluation on "code stacks".  This is good, as long as
the language comprising the code stacks is simple.  Currently it also
includes machine asm, so it's screaming for a semantics that can be
used to verify partial evaluation rules.