4 ms·
This logic is flawed, and I'm not surprised it's coming from Microsoft. If the expectation is that (repeatedly?) making blind guesses quickly and (optionally) c
by kevwil 9y ago
This logic is flawed, and I'm not surprised it's coming from Microsoft. If the expectation is that (repeatedly?) making blind guesses quickly and (optionally) cleaning up the mess later is better than expressing an understanding of the problem domain before writing 'real' code, then yes TDD will not live up to those expectations.
- seanmcdirmid 9y agoNote the site is http://blogs.msdn.microsoft.com http://blogs.msdn.microsoft.com. Any MS employee can start a blog here, and none of the content posted by an employee goes through a vetting process. That being said, the logic is not really flawed at all. There will always be the real scottsman fallacy to deal with (what is real code? When does code become real?). TDD promotes writing tests first, whereas you might not even know what the spec is until you've gone through multiple iterations of the code, and along the way you'll have to refactor and rewrite to reduce coupling and reflect better understanding of the spec. Tests aren't bad, but writing the tests at the beginning doesn't really make sense in that context.