SWE 437 Homeworks 6 and 7
Fall 2018
Final Tools Assignment

There are various options for the final two assignments. Pick one for assignment 6 and a different one for assignment 7.

• If you have never done this before, work through a simple TDD example such as Koskela, chapter 2: Koskela is available online through Mason's Safari subscription:
  Direct Safari Link
  Off Campus Safari Link.
  
• If you have never done this before, set up a continuous integration service such as Jenkins or Travis CI. and use it to build and update a small example system. You should intentionally break and restore builds as part of your exercise.
  Note: This works best as a collaborative exercise.
• Use Evosuite to generate tests on a small project. You should analyze the tests to see what Evosuite comes up with and what it misses.
  Note: Evosuite is aimed at Java.
• Use Klee to generate tests on a small project. You should analyze the tests to see what Klee comes up with and what it misses.
  Note: Klee is aimed at C.
  
• Read the book How Google Tests Software and determine where your organization is on the "Test Certified Level" (Chapter 2) scale. Create up a plausible plan to move your organization up to a higher level. Ideally, you would want to actually start implementing this plan, but that is not required for credit.
  The book is available electronically through Mason's Safari subscription:
  Direct Safari Link
  Off Campus Safari Link.
  Note: This option really only makes sense for students who have a software development job or internship.
• Consider the PIT mutation tool. Apply PIT to a program of your choosing and generate tests to kill the PIT mutants. Your tests should be as close to mutation-adequate as you can make them. That means you should be able to justify unkilled mutants in your report. Note: This tool is aimed at practicing developers, so it is relatively easy to deal with.
• Consider the Major mutation tool. Apply Major to a program of your choosing and generate tests to kill the Major mutants. Your tests should be as close to mutation-adequate as you can make them. That means you should be able to justify unkilled mutants in your report. Note: This tool is far more powerful than PIT, but it's a research tool, and it has its quirks.
• For those interested in a game, play Code Defenders. Obviously, this only works as a collaborative exercise. For full credit, you need to play the game until your test set is pretty good. You'll need to justify why you think that is the case.