Fixing security vulnerabilities in every piece of software ever written (or to be written) seems like a hell of a lot more work than hardening the processor. N

Fixing security vulnerabilities in every piece of software ever written (or to be written) seems like a hell of a lot more work than hardening the processor. Non-executable memory, for example, eliminates a whole class of vulnerabilities in which illicitly added code is run from the stack or heap. To suggest that further work on the processor security model is a waste because people ought to have done it right the first time is blind to the fact that people will screw up no matter what.