Hi, I am using PPL for my project which is related to verification. In this context I require my computations to be exact or, if that is not possible, to be over-approximating. From what I read in the documentation, PPL ensures such an over-approximation, which is nice. I am using C_Polyhedron as my data type and currently interested in using rational arithmetic (mpq_class) in combination with that. From what I see, you can use the mpz_class type for C_Polyhedron, but not mpq_class directly. Is there a way intended to overcome this problem? In my current setup I use the double interface, but this requires conversion and most probably introduces rounding errors. Furthermore a usage question: Is there an intended way on how to compute a linear transformation of a polyhedron with a matrix? Best regards, Stefan -- Stefan Schupp M.Sc. RWTH Aachen University Computer Science Department, Informatik 2 D-52056 Aachen, Germany http://ths.rwth-aachen.de/people/stefan-schupp/ Tel.: +49 241 80 21244