> First, you cannot draw such a causal conclusion from statistical data alone. From what else but (suitable) statistical can you draw such a conclusion? > Doe

> First, you cannot draw such a causal conclusion from statistical data alone. From what else but (suitable) statistical can you draw such a conclusion? > Does that sound stupid? I hope so, because it is. Yes. > But that's exactly the type of argument that this article is making. No. The argument says that the attacks on society as a whole will likely remain constant regardless of airport security because there are many other, unsecured attack vectors. Going back to your analogy, airport security is like allowing only passphrase-secured SSH keys of 8192 bits (say), while at the same time running a telnet daemon which gives root immediately – the latter corresponding to all the unsecured roads, malls, trains, you-name-it. That is, if we don’t want to specifically protect airports but human lives, we have to consider not only the security gain from increased security in one area, but also whether this reduces the effectivity of attacks overall. If we want to protect one whole server and not just the SSH daemon, we have to consider not only making the SSH daemon as secure as possible but also all the other software running on this server – at which point we notice that it cannot be possibly feasible to secure this thing 100%, because nobody likes going through metal detectors to be allowed to leave their house.