Why is automated verification valuable?

There are two answers to this question. The first one should be easy to guess: because automated verification saves time and effort. We could verify our system's behavior by having someone sitting in front of a screen, running our system with one hand, while holding our specification with the other, and checking that everything works as we describe in our features. This approach will take a lot of that person's time and it runs the risk of that person getting tired or distracted and missing things or making mistakes. Alternatively, we could automatically run code that verifies our system consistently, reliably, and faster than a person can.

The second answer is much more appropriate. We have mentioned repeatedly throughout this book that features must reflect domain-level behavior, things that are important to an actor in the context of a particular feature's scenario. So, for instance, if our scenario for publishing a blog...