Test-Driven Development vs. Test-After Development

stephenwalther.com

3 pointsabl3 comments

I wonder if, one day, people will realize that none of this stuff really matters. Write and comment your code clearly, think your designs through, and comply with your team's conventions so that people know what to expect.

None of this stuff (TDD, TAD, XP, Agile, whatever) is computer science. None of it is science period. They're all essentially shamanistic rituals which "work" by giving you a direction so that you can start writing code. If you find a methodology works for you, great! Use it! But all this squabbling about the one true path is nonsense.

If something as simple as a checklist [1] can have a meaningful impact on patient care in hospitals, I think it's worth talking about the equivalent in software.

Neither claims to be any kind of science, but both can save the day by compensating for the natural human foibles of the practitioner.

[1]: http://www.newyorker.com/reporting/2007/12/10/071210fa_fact_...

Sure, having an organization strategy is a good plan, and I didn't mean to imply otherwise. The problem I have is with people claiming that their way is the "right" way without having any legitimate scientific backing for the claim.

Reply on news.ycombinator.com