I’ve recently been following the
#isTDDDead debate between Kent Beck (@kentbeck), David Heinemeier Hansson (@dhh), and Martin Fowler (@martinfowler) with some interest. I think that it’s particularly beneficial that ideas, which are often taken for granted, can be challenged in a constructive manner. That way you can figure out if they stand up to scrutiny or fall down flat on their faces.
The discussion began with @dhh making the following points on TDD and test technique, which I hope I’ve got right. Firstly, the strict definition of TDD includes the following:
- TTD is used to drive unit tests
- You can’t have collaborators
- You can’t touch the database
- You can’t touch the File system
- Fast Unit Tests, complete in the blink of an eye.
He went on to say that you therefore drive your system’s architecture from the use of mocks and in that way the architecture suffers damage from the drive to isolate and mock everything, whilst the mandatory enforcement of the 'red, green, refactor’ cycle is too prescriptive. He also stated that a lot of people mistake that you can’t have confidence in your code and you can’t deliver incremental functionality with tests unless you go through this mandated, well paved road of TDD.
