I mentioned the range constraint of the algorithm previously (which I guess you ignored), My assertions are that it can be used with many sets of integers in the real world, and whenever it can be used, it is very fast, indeed.
In academia, huge and negative numbers really have their place. In the rest of the world, not so much. Did you ever go to the store and buy -10 loaves of bread? Pick up > 65,000 boxes of nails, or any other supply? Even when you're drilling in the earth, you measure the depth as a positive number, never a negative. I don't think it's difficult to see where this algorithm can be used wisely, or in where it's inappropriate to use it. In those cases, it's easy to use Quicksort or one of it's derivatives, instead. adak
