Running experiments and checking the results seems intuitive. No one likes redoing his work. Yet, people become hesitant about tests, when they don’t validate the hypothesis. In those cases, the experiment feels like a waste. Humans are hesitant to throw their work out. Doing so feels wasteful. Yet, a test isn’t a waste, if you learn something from it (how to achieve a goal, what the target should be, etc.). From my experience, that idea can hard to explain to people.
I always found the practice of TDD to be attractive, since it was basically tight loops of the scientific method. Make a hypothesis that the software is deficient, run an experiment (create/execute a test), validate the hypothesis with a failed test, make another hypothesis about how you could fix the situation, run an experiment (write some prod code), validate the hypothesis with a passing test.
Running experiments and checking the results seems intuitive. No one likes redoing his work. Yet, people become hesitant about tests, when they don’t validate the hypothesis. In those cases, the experiment feels like a waste. Humans are hesitant to throw their work out. Doing so feels wasteful. Yet, a test isn’t a waste, if you learn something from it (how to achieve a goal, what the target should be, etc.). From my experience, that idea can hard to explain to people.
I always found the practice of TDD to be attractive, since it was basically tight loops of the scientific method. Make a hypothesis that the software is deficient, run an experiment (create/execute a test), validate the hypothesis with a failed test, make another hypothesis about how you could fix the situation, run an experiment (write some prod code), validate the hypothesis with a passing test.
Yeah. The tracer bullet idea in the Pragamatic Programmer is quite similar but on a macro level.