Creates a symbol-table data type whose keys are two-dimensional points to support efficient range search and nearest neighbor search (NNS).
PointST.java is a brute-force range search and NNS implementation that uses a red-black binary search tree.
KdTreeST.java implements range search and NNS using a 2-d tree. This implementation is much faster than PointST: in a trial run, KdTreeST made over 317,000 calls per second to nearest(), compared to just 8 calls per second for PointST.