==========Fons Rademakers mail======================= Rene, I think trying to integrate a not complete C++ interpreter makes no sense and will not add any additional value to ROOT. We should work with Masa in a parallel rewrite of the CINT kernel in C++ all the while keeping CINTs good features and backward compatability in mind. Having spent yesterday a whole day in debugging CINT I agree that a rewrite would be nice (CINT clearly started as a much smaller project than it currently is, i.e. from simple C subset interpreter to (almost) full ANSI C++ interpreter. It is already amazing that Masa managed to evoluate the code like he did). The ideal would be to have Masa + a VERY GOOD Fermi lab guy work for a year on this project. The benefits of a rewrite should be: - thread safe - robustness (better error handling for trivial typing errors) - better template support - more transparent, i.e. more maintainable code - exception handling all the while understanding that C++ is a huge monster of a language (largest and most complex computer language ever) normally only implemented by fairly large teams of people (as I know from HP experience). As always,implementing 90% of C++ (like we have now) will be fairly well doable. Getting to 100% might be almost impossible. Fons