Closed rene-thiemann closed 10 years ago
yes, the usability is guessed (it is the UsableSymbol
map here
https://github.com/jwaldmann/co4/blob/usable-rules/tc/CO4/Test/TermComp2014/Standalone.hs#L97 )
and then closure is verified in the constraint (https://github.com/jwaldmann/co4/blob/usable-rules/tc/CO4/Test/TermComp2014/Standalone.hs#L187 )
ok, then that's fine.
Looking at the proof https://github.com/jwaldmann/co4/blob/usable-rules/JFP_Ex31.plain.2 again, although the issue with unlabProc seem to be fixed, there might be a problem in the computation of usable rules.
For example, in https://github.com/jwaldmann/co4/blob/usable-rules/JFP_Ex31.plain.2#L134-135 the two
top
-rules are mentioned in the set of usable rules, although they are obviously not usable (neither the pairs nor any of the other rules contain the symboltop
). Is this intended, or may the SAT-solver over approximate the usable rules by just guessing a set of suitable usable rules which are closed under right-hand sides?