This directory is not related to the soundness and completeness proofs
developed for the three strategies.  Rather, we exhibit trivial
machines, in which we can satisfy Manolios and Moore's
characterization of partial correctness.  The key is that the machine
essentially ignores the program, but sets its state to a certain
constant value and then halts.  Suppose we now want to prove a routine
correct where the routine does not end in a halting state and makes
some nontrivial modifications defined by a function modify.  Manolios
and Moore's notion says that we need to prove the following theorem.

    (implies (pre s) (= (stepw s) (stepw (modify s))))

But the condition does not guarantee that the machine ever reaches a
state (modify s) from s, merely that from s and (modify s) we can
reach the same halting state (if one exists, which it does in every
case here).  So if the machine essentially ignores the program and
memory and everything else, and merely halts with a constant value for
the state then that condition is satisfied in this case.
