So posts about the Pants Bomber are currently dominating my RSS reader.
This post from Professor Bainbridge is very good. Co-sign most of it, this despite my fear of flying. See also this much-linked-to post from Bruce Schneier.
No, I am not inclined to call for Napolitano's head. One of my Facebook friends posted a status update to the effect that counter-terrorism security regulations ought to be subject to cost-benefit analysis, just like everything else. That strikes me as roughly right; that is, yes, it may not be appropriate to try to deter 100% of terrorism accidents if 100% deterrence comes only at enormous cost and difficulty.
That said, counterterrorism requires particularly weird kinds of cost-benefit analysis. The typical terrorist would rather kill 100 passengers in a sensational, much-publicized air disaster than kill 100 people via car bombs scattered in tiny towns throughout the country. The point is to sow fear among the general population that's disproportionate to the actual carnage inflicted. In some ways, terrorists are good behavioral economists; they manipulate the availability heuristic like nobody's business. This makes it hard to determine what the relevant benefits are: is it lives saved, or is it some inherently amorphous feeling of security?