Reminder: hw4 written is cancelled, but the hw4 program is due
Wednesday, and I urge you to get started on it!  There are
several challenges:

  * compute the first tour (do a DFS on edges of the MST)

  * represent your tour in such a way (array or list, at your choice)
    so that you can quickly check whether an edge e (from G)
    defines a good 2-opt move.  This means you need:
       * for a given point index v or w, find its corresponding
         position in the tour order (say i and j)
       * given i+1 and j+1 (maybe wrapped around), which are
         indices into the tour order, find the corresponding
         indices into the points array

    I have some example output in the subdirectory hw4-outputs/.  Note
    that the "TSP2.png" files might still have some crossing edges,
    because we are not doing a perfect 2-opt.

  * Finally, if that all works (for run10, run100, run1000)
    then can you try some of the extra credit (faster, better)?

OK, after that, we get back to Section 4.4, computing shortest paths
where we possibly have weighted edges, directed edges, and maybe even
negative weights.

We'll hit the slides again (started last time).