Totally agree, and that's why clx static analysis and optimizer steps permits some performance gains