Hi Paul:

while browsing your PODC '99 paper: Liveness-Preserving Simulation
Relations, I noticed what seems an interesting
connection between your work and our work
on Adaptive Programming (AP). 
A recent paper with Boaz Patt-Shamir:
http://www.ccs.neu.edu/research/demeter/biblio/strategies.html
explains our recent theoretical work on AP.
http://www.ccs.neu.edu/research/demeter/ is the AP project home page.

You work with abstract automata and concrete automata as executable notations.
We work with abstract graphs and concrete graphs as executable notations.

You consider a relation between the states/transitions of the abstract
and concrete automaton.
We consider a relation between the paths of the abstract
and concrete graphs.

You consider "trace inclusion" as the meaning of concrete automaton A
being a correct implementation of abstract automaton B.
We consider "path expansion" as the meaning of concrete graph A
being a correct implementation of abstract graph B.

You have various concepts of A simulating B.
We have various concepts of A being a refinement of B.

There are differences: your automata seem to be
more complicated than our graphs; our executions involve object graphs.

I look forward to discussing some of those connections with
you personally.

-- Karl