The testing pyramid

Mike Cohn is credited with having created the concept of a testing pyramid in his book Succeeding with Agile. The concept is that your cheapest (fastest) tests to run, which will be your unit tests, go at the bottom of the pyramid; service level integration tests are on top of this, and at the very top, you place full end-to-end tests, which are the costliest element. Because this is a pyramid, the number of tests gets smaller as you move up the pyramid.

In the early days of automated testing, all the testing was completed at the top of the pyramid. While this did work from a quality perspective, it meant the process of debugging the area at fault would be incredibly complicated and time-consuming. If you were lucky, there might be a complete failure which could be tracked down to a stack trace. If you were unlucky, then the problem would be behavioral; and even if you knew the system inside out, it would involve plowing through thousands of lines of code and manually...