QuickSelect
 =============

  Background
 ------------

The problem we are trying to solve here is that of finding
the k'th largest (or smallest) of a collection of data
(hereafter referred to as the array).  The array is
therefore considered to be unordered because, of course, if
it were ordered, it would be a matter of mere indexing to
get the k'th largest value.

  The Plan
 ----------

The idea behind quickSelect is to run a partial quicksort on
the data.  Partial in that instead of recursing down both
halves of the pivot-partitioned data, we only recurse down
the half where the k'th largest would be.  This requires a
slight modificaton of the partition algorithm to return the
position where the equal partition starts so we know what
half to recurse into.

  Analysis
 ----------

Recursing only down half the array at a time makes this turn
out linear instead of n*log(n) overall and so is a goodly
speedup from sorting the data before finding the k'th
largest.

  Smallest Finding
 ------------------

Finding the k'th smallest can be done by just finding the
n-k'th largest.  Easy peasy, right?