We prove what we want to prove, and the real difficulty is to know what we want to prove.