Dick,
While I did work in the QA biz for many years, I've been in the engineering side for the last 10 yrs where our guys set the criteria for inspections or more often I'm involved in evaluating what happened on failed parts and what to do about preventing failures thru inspections.

The studies I've seen suggest that probabilities of detection (POD) at 95% is about a .050" crack/flaw for FMP on most stuff. On most parts we don't allow for any detectable cracks/flaws. That says the minimum detectable crack size is acceptable for strength up to at least twice that size. But, we allow for 2 opportunities to detect. Establishing the inspection interval is at least as important as the type of inspection. The repeated stresses need to be assessed along with the strength of the part in order to intelligently establish intervals. As a NDI pro you're undoubtedly familiar with all of this, I just wanted the crowd here to understand that there's more to it than just inspecting a barrel with a NDI method and declaring it crack free. It ain't necessarily so. It may have a crack below the detectable threshold or may develop a crack right after being inspected. That's why critical components on commercial airplanes get regularly scheduled inspections for cracks.