[eru]

Scott Aaronson has some good motivation for "QM falls out naturally if you try and use a 2-norm for your probalities instead of a 1-norm." See http://www.scottaaronson.com/democritus/lec9.html