How much reason supports the rituals and mantras often repeated as coding guidelines? It turns out that the advice often fails, even for the novices they are intended to guide. Let’s reason through these rather than accept them as unquestioned habits.
How many asserts should a test case have or not have? How much work should a constructor (not) do? What mantra guides test-first programming? How do you name your classes and other identifiers? How do you lay out your code? These questions and others have standard answers based on received and repeated mantras, practices that are communicated in good faith to be passed on as habits. But how much reason supports these assertions? It turns out that the advice often fails, even for the novices they are intended to guide.
This talk has little respect for ritual and tradition and takes no prisoners: What actually makes sense and what doesn’t when it comes to matters of practice? What guidelines offer the greatest effect and the greatest learning?
Video producer: http://www.javazone.no/