Engineering-risk Books

The work proposes a new technology from the marriage of in-service inspection, probabilistic fracture mechanics and probabilistic risk assessment. It founders on the violation of an axiom of probability theory by in-service inspection.

In brief, the preservation of lives and property sometimes depends upon the integrity of a structure which functions under mechanical stress. In-service inspection (ISI), using ultrasound, xrays or other nondestructive techniques for detecting flaws, offers a mechanism by which these risks might be mitigated. Engineers apply ISI to a variety of critically important structures, including the pressure boundaries of nuclear power reactors.

Seemingly, one could improve decisions about when and how to use ISI by marrying it to a pair of recently developed disciplines. One of them, probabilistic risk assessment, estimates the probabilities of events with dire consequences for people, such as reactor meltdowns. The other, probabilistic fracture mechanics, estimates the probabilities of failure of structures which are subjected to ISI. Does ISI not have an effect on the probability of reactor meltdown? Don't do it, would be the conclusion coming from the new technology. Does it cut this probability 10-fold? Do it, would be the conclusion.

However, there is a fly in the ointment. A recently published paper demonstrates that ISI violates an axiom of probability theory (see "Erratic Measure," in NDE for the Energy Industry 1995, pp. 1-6. The American Society of Mechanical Engineers, New York, NY.)

This is a very disappointing and shallow book about a very deep and important subject: software failure risk. It seemed to me to be nothing more than a mediocre survey article that was expanded into a book, with no real meat in it. Basically, after reading the book, I had very little in the way of new ideas for how to measure software failure risk for a real industrial software system, nor how to manage software failure risk. There was a 10 page chapter on software testing. Since testing should reduce software failure risk, I would have expected a lot more information in this chapter, but it was almost all superficialities. It is hard to know who this book is aimed at - certainly not software developers or testers.