One of these days, I want to do a long post about quantum cryptography. Actually, I want to do it now, but I don't understand it well enough.
But anyway I can do a short post, to give me an excuse to link to this discussion of practical risks by Michael Nielsen.
As a baby experimentalist, I find the prospective practical problems pretty interesting. Right now making these things work at all is a slow, fickle process that involves ritual sacrifices to the vacuum gods (at least, it is in our lab.) It's all jury rigged and improvised. How could our ad hoc solutions be exploited? Are the lasers beams that we use to push the atoms around carrying an information about the system afterwards, for instance? (Probably this is a dumb question, but it's one I've never thought about before.)
Actually what we're working on is not quite quantum cryptography, but it's related. Michael Nielsen says, "The basic problem is for Alice and Bob to share a secret key which only they know." And although that is not really our problem, our two vacuum chambers are indeed named Alice and Bob. This is because all of the papers ever written on this subject call the communicating parties Alice and Bob. If you were to write one now referring to Andrew and Betsy, your audience would be hopelessly confused.
The principle behind all of this stuff is just that in quantum mechanics you get to know statstical stuff about particles without actually looking at them or measuring anything -- and that once you do measure something, you don't have the same statistical spread anymore in later measurements. So Alice can send Bob a bunch of particles, and then they can check, either by revealing some bits or by comparing the kinds of measurements that they made, that all of the statistical distributions are what they should be. If they're not, they've got an eveasdropper doing extra measurements. If all is well, then they've sent information that they can be sure only the two of them share. They can then use this, for example, as a code key. (It's also possible for Alice to send Bob information that even Alice doesn't know -- this is what's going on in what we somewhat misleadingly call "teleportation.")
But what I want to know is, when these systems are standardized and made available to the military, is the person who ends up sending this secret data going to object being called "Alice" by the engineers?