I'm a geek. There's no other way to describe it. Because I'm a geek, I read geek things. Today, I was reading an article about confirming sensitive information.
You have two managers, Bob and Alice. One of Bob's coworkers complains about something, but requests confidentiality. Alice later tells Bob that a coworker has complained to her about something and Bob realizes it's the same problem. The problem: Bob and Alice want to know if the same person complained, but they can't just blurt out who it is without violating confidentiality. How do they confirm the person's identity without revealing the identity if it's not the same person?
My two favorite solutions are as follows (yeah, you can poke holes in them, but they're still fun):
Bob calls the airline and reserves a flight in the complainant's name. Alice later calls and tries to cancel the flight in the complainant's name. If different people complained, Alice will be told that the person didn't make a reservation, but she won't be told who did (this is probably a hell of a lot safer with restaurant reservations).
A pool of candidates is assembled (all of the coworkers who could have complained) and each is assigned a random phone number. Bob calls the random number assigned to the complainant and tries to leave a message for Alice. The person at the other end doesn't know who Alice is. Alice then calls the phone number assigned to her complainant and asks if anyone left any messages for her :)
Any other cool ideas?
Incidentally, the Bob and Alice problem was based on a real-life problem. The actual solution that was implemented was much different from those above and pretty funny in its own right.