Hm. That's a good question. You're right that the setup we gave you doesn't really lend itself to JUnit testing.
As you've probably realized, in order to have a high-scoring naturalist, you need to capture your own representation of the graph so that you can run algorithms that need global knowledge, e.g., path-finding. If you make your graph algorithms suitably generic (using type parameters instead of a fixed Node class), you could make individual JUnit tests that test those algorithms on fixed graphs that you create, using your own version of the Node class.
You could also invoke Simulator.main from inside a JUnit test, for example on running on custom maps to make sure the naturalist works on maps with strange topologies.
--
You received this message because you are subscribed to the Google Groups "Cornell CS 2110" group.
To post to this group, send email to cornell...@googlegroups.com.
To unsubscribe from this group, send email to cornell-cs211...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/cornell-cs2110?hl=en.