> well, sometimes (haven't checked here yet), we find > compilers/interpreters become more strict about rejection broken > code (like gcc4). As a gcc developer, I must say that the only way you can get away with doing this is if you have a support infrastructure to help people understand what they need to do to fix their code.