Thanks for this. I am most interested in mathematical comparisons. That is to say, ways in which optimal analog computation (idealizing away from noise propagation, among other things) might be sometimes better than its digital counterpart. Your last point about diff equations is directly relevant to this.
