Maintainable Tests

Just some thoughts on how to structure tests (specifically setup) in a way
that makes them easier to understand in the future.