Track Talk F4

Concurrency: Why End-to-End Testing Isn’t Enough

Rick Tracy

09:00-10:00 Friday 14th June

In almost every project I worked on, End to End tests (also known as E2E or “chain tests”) are held up as the holy path for building test strategies and the ultimate means of assessing quality in a system. They are also a source of frustration for testers and users. E2E tests fall apart because of mismatched environments and configurations, and they take considerable coordination and communication too.

Is it worth it? In this talk, I’m asserting that No, it is not. In my experience, integration testing costs far too much, telling us things we can find in other ways, while not going far enough on its own to test an entire system flow.

We’ll explore the limitations of integration testing and some of its alternatives. Then I will present a new testing style: Concurrency Testing. This style of testing is designed to emulate real production actions, trigger multiple actions, and combine exploratory tests to reveal much more than a traditional E2E test. You’ll be able to use this model to enhance your tests, to provide some real value for effort, and maybe find a new or unique bug that’s been hiding away. At the very least, I hope to relieve some of the frustration people express about E2E testing — and give us some fun new tools to play with