Tue Oct 18 21:29:47 EDT 2011
Applicative vs. dataflow
Still, really, why is this such a big deal.. Why can't it be
applicative until the very end, when the last return value is
"allocated". I don't know why I keep feeling resistance to the idea
of just store it and be done..
Since the conversion from applicative to dataflow (output binding by
unification of caller-provided bindings) has to happen at some point,
the real question is: what is the most natural place for this? Should
we do any operations on code that is in this form, or should it be
just the final step, i.e. part of the "printing".
Maybe that's the real problem though. Unification.. Is there a
simple way to express this in haskell? 
Anyway, sleeping about it a bit, it seems best to separate memory
references from local variables, but reuse the "Bind" object to print