Empirical test-after development by Mark Seemann
A few techniques for situations where TDD is impractical.
In Epistemology of software I described how test-driven development (TDD) is a scientific approach to software development. By running tests, we conduct falsifiable experiments to gather empirical evidence that corroborate our hypothesis about the software we're developing.
TDD is, in my experience, the most effective way to deliver useful software within reasonable time frames. Even so, I also understand that there are situations where TDD is impractical. I can think of a few overall categories where this is the case, but undoubtedly, there are more than those I enumerate in this article.
Not all is lost in those cases. What do you do when TDD seems impractical? The key is to understand how empirical methods work. How do you gather evidence that corroborates your hypotheses? In subsequent articles, I'll share some techniques that I've found useful.
- Empirical Characterization Testing
- Empirical software prototyping
Each of these articles will contain tips and examples that apply in those situations where TDD is impractical. Most of the ideas and techniques I've learned from other people, and I'll be as diligent as possible to cite sources of inspiration.
Next: Empirical Characterization Testing.
