Providing concrete BDD-style examples as acceptance criteria for user stories is a powerful way to influence the testability of your architecture. This is especially true if developers are provided the examples before starting development and they are responsible for the automated execution of those examples. This talk illustrate how the examples can have this benefit.
Examples are a powerful way to communicate what "done" looks like for agile user stories. When they are automated, they can also act as very useful regression tests/checks. But most automated checks are too detailed to act as useful examples. This can be addressed by using a layered approach to examples in which we provide less detail for broader scope examples and narrower scope for more detailed examples.
Can we stop obsessing over code coverage numbers? Test automation is good, no doubt, but we need to deliver software don't we? Let's talk about a test automation strategy that actually works and how to leverage manual testing in the most effective way.
Code coverage metrics... please go away!