Sorting Lab Optional Extensions

If you wrap up the first parts of the lab and wish to think a bit more about sorting, here's a few ideas.

Measure the time it takes to sort integer vectors of various sizes with Selection Sort. Do the measurements match your expected running times? Generate plots of your data with Gnuplot, which is a very handy tool to have at your disposal.
Gnuplot is a tool to make graphs. If interested, try out some of the examples from the gnuplot tutorial.
The file start.gp from the handouts page contains a script for gnuplot to get you started. Once you generate a postscript file, use the command open file.ps to open it in a viewer. You can print it to the printer using the lpr file.ps command.
To practice, plot functions that are O(1), O(n), O(n^2). In addition, you can copy the file unknown.dat and plot the data in that file. (What do you think it represents?)
Quick Sort is typically considered the most versatile and fastest way to sort. Replace Selection Sort with Quick Sort, and measure the performance of your new algorithm. Does it match what you would expect?

A few hints:

When timing your program, you may call System.currentTimeMillis() to get the current time in milliseconds. This returns a long value, and you would use it as follows:
```
  long startTime = System.currentTimeMillis();
  // run test here
  long endTime = System.currentTimeMillis();
  long duration = endTime - startTime;
```
Timing is very sensitive to what else is happening on your machine (ie, network traffic, background processes, etc.). To minimize these impacts, do not run any other programs (like iTunes, Safari, ...) at the same time as your program. Also, make multiple tests and average the results.
Also, when measuring running times, run your program using a command like java -Xint SortTest. The -Xint flag disables certain dynamic optimizations that may throw off your timing measurements.