Comment by pfdietz

Hmm, this sounds like exactly the opposite of what I was talking about. It delays execution rather than promoting execution to compile time.

What I had expected you to talk about was some way of getting the compile time execution of macros by a sufficiently smart compiler that could do extensive partial evaluation at compile time, including crossing procedure boundaries. Of course that's antithetical to the Lisp philosophy of allowing dynamic redefinition of functions and such.