Interface for runners that allow sorting of tests. By sorting tests based on when they last failed, most recently
failed first, you can reduce the average time to the first test failing. Test sorting should not be used to
cope with order dependencies between tests. Tests that are isolated from each other are less
expensive to maintain and can be run individually.