Grader's comments on Problem Set #2:
   
   c1. Can you show that your run-time analysis doesn't change when
   you vary epsilon and p?
   
   c2. You need to specify how many iterations your algorithm takes.
   
   c3.  Please be more complete. A good proof is one that has a very
   clear line of reasoning (in words) that's also concise.
   
   c4.  Remember Horner's rule in your runtime analysis (O(n)).
   
   c5.  Please clearly specify AND PROVE the termination condition in
   your algorithm.
   
   c6.  Just arguing "this is similar to QuickSort" is not
   sufficient. This algorithm's analysis is actually much more
   complex. See solutions for a simpler and more elegant way of doing
   the ranking.
   
   c7.  This isn't exactly comparison sorting. The difference between
   this problem and QuickSort is the fact that the pivot element is
   from another set. You need to show analysis that you've taken this
   fact into account.