Can we honestly hope to attain the 99% in uncovering those faults, undetected errors, or are we simply chasing the impossible? If these numbers are truly accurate, they're more what we would need, I'd say. Mind you, our 5-year old Mirtec machine seems to miss a bit too often, and without diving into a deep sea of statistics, it appears to only pinpoint about 94-96% of errors?