What We Just Learned
We began this chapter by reviewing the recommended distribution of test depth as described by the Testing Pyramid. We then looked at the separate concern of who the tests should be readable by, as captured by the Testing Iceberg. This led us into a discussion of how we can keep our scenarios running quickly, while still keeping their documentation of our application complete. Finally, we looked at concrete techniques that allow us to vary the amount of the application that our steps exercise without changing the scenarios themselves.
The example in this chapter focused on switching between the web UI and the business logic. Try to introduce a similar choice somewhere else in our application. For instance, you could ...