I felt otherwise. I believe the author not only pointed out other alternatives but talked with those people trying to do things differently. Namely the course o

I felt otherwise. I believe the author not only pointed out other alternatives but talked with those people trying to do things differently. Namely the course of action is simply to try different and more varied things. The whole problem that is being decried is that of following a rigid, fixed formula that tries to generalize for all cases. That of using the mouse model for everything, even those things for which it is not the best fit. In software terms this is akin to Netflix's implementation of Chaos Monkey. Mainly that one shouldn't only check for certain failure conditions that have been previously specified, but to also create a framework where you try to look for faults in areas that are outside of the expected.