
Roberto Bagnara wrote:
Enea Zaffanella wrote:
[...]
This is why we have a Status class in the first place: to make it simple to add new introspective capabilities to the polyhedra classes. The idea of distinguishing between minimization and NNC-minimization looks promising (even though the name "NNC-minimization" is horrible ;-)
Well, choosing the right name was not the top item in my priority list. Any better idea is welcome ...
So, a first question for Elisa: as things are now, do we ever "assume" that whenever both con_sys and gen_sys are minimized, then the saturation matrices are up-to-date ?
We never assume that consciously (i.e., if the actual code happens to rely on that assumptiom, then it is a bug).
Good. However, in the method update_sat_X, we DO assume that if both con_sys and gen_sys are minimized, then they are the dual of each other. Such an assumption, which was completely legal before, will have to be relaxed if we adopt the above scenario.
ciao, Enea.