I've only removed the calls to the visualization tool in the GRS files and replaced it with a listing of Place nodes at the end of a transformation. But yeah, a fresh git clone is probably better. While we are at it: yesterday you said that the performance models would be completely reducible to one single Place. That's indeed the case for the models up to sp20000, but sp40000 and the even larger ones terminate with several places left (and thus no top-AND-state in the statechart) also with your GrGen solution.

Hello Pieter, done! cheers

OK Tassilo, I have split the question in two parts now: the first question is specific to language experience in a transformation context. The second one is on PL experience in the broad sense. Best, Pieter

It does :) At least for the Flowgraph and the Petri Net case, I did not check the Class Diagram case. Georg

Thanks, using that now.