There is a best practice of writing unit tests during the development. Some experienced developers say that developing and writing unit tests together is faster than to develop without writing unit tests. When we tried to adopt this method of writing unit tests, we found that it costs more time with no benefit: The tests never failed due to a real bug.

In this talk, I will present when unit testing gets wrong and why. I will describe what method we used to write unit tests with a positive outcome, how it relates to DDD and clean architecture and how to apply it to systems with legacy code. By the end of the talk, you will have the strategy for unit testing against the living products that seek to continue growing, even though they are all built of complex legacy code.