I generally agree with the consensus that the lock washer should butt against the nut. I have had times where this was not practical (inadequate space on the nut end, for one), and I did not hesitate to use the lock washer on the bolt side. This is no sin, and functions almost as well as the other way.
Consider: What you have is a steel ring, cut in one place. Then, the two ends are bent so they are not in the same plane, then tempered to form a basic spring. Therefore, when the nut is threaded against the lock washer, it takes substantial force to flatten the lock washer back to a single plane. Because it is spring steel, it does not simply bend the washer back to flat; the force applied is retained as potential energy stored in the washer. Pushing back against the nut, hence discouraging the nut from coming loose. It will not allow the nut to vibrate loose, as this counter force prevents any space developing where physical vibration can occur.
The same affects apply, if one puts the lock washer against the bolt. I think it is safe to say that the control of vibration (loosening) would be slightly better on the nut side, because the same spring force need constrain a somewhat smaller mass of the nut. I posit that the difference in effectiveness would be hard to measure, whatever side one puts the lock washer on. Usually, if there is only one flat washer, it too, will go on the nut end. Besides applying a more even fastening pressure, this practice provides a nice flat surface for the lock washer to contract against.