Hi Luke:

I suggest that you use the Mona tool:

[Google: Mona Schwartzbach]
http://citeseer.nj.nec.com/327523.html

http://www.brics.dk/mona/manual.html

to test whether two strategy/class graph pairs are identical on common objects.
More below.

For further motivation see:

http://citeseer.nj.nec.com/cache/papers/cs/280/ftp:zSzzSzftp.ccs.neu.eduzSzpubzSzpeoplezSzlieberzSztaposzSzrev2.pdf/klarlund97formal.pdf

Also search for "Demeter" in this document.

The Mona tool is likely useful for AOP. It kind of describes a powerful, decidable
pointcut language.

Your host is the PhD student Theo Skotiniotis. He is working on a more expressive pointcut 
language.


-- Karl


Adapt known automata algorithms to solve the following problem:

TRAVERSAL EQUIVALENCE PROBLEM:
Given two strategy/class graph pairs (S1,G1) and (S2,G2) such that
Common = Objects(G1) intersect Objects(G2) is not empty,
determine whether they define the same traversals for all objects
in Common. (Also treat the case where the question is whether one
is always a subtraversal of the other.) If the traversals are different
for some objects, give an object where the traversals are different.

Motivation: during the evolution of adaptive programs and class graphs,
we need to know how traversals change.

Ideas: Construct the traversal graph as described in
http://www.ccs.neu.edu/research/demeter/biblio/strategies.html
using Doug's implementation in Demeter/Java.
View the traversal graph as a generalized finite state machine
(instead of processing strings, we process objects. The output
is in both cases a string.) Generalize finite automata eqivalence
algorithms.

Use the Mona tool to check traversal strategies for equivalence.