I think you are absolutely right about the optimal response / change in procedures (basically, don't change anything) given these awful events. Given the rarity of these types of events, the expected value of any security program is essentially identical to what it was two weeks ago. It's a crying shame that this kind of cost-benefit analysis is not the universal reaction to proposed changes in policy arising out of unusual events.

But it is a logistical fact that faster killing devices kill more people per incident.