Out of curiosity — has anybody compared the speed of 1. interpreting a parsed syntax tree, versus 2. interpreting byte code, versus 3. interpreting a RISC CPU ? Regards, Adriaan van Os