【TDDを再定義したほうがいいって話だったのさ】UncleBob, Martinfowler, DHHのツイートまとめ
@unclebobmartin "Fundamentalism is the demand for a strict adherence to orthodox theological doctrines…" Seems to me @dhh chose a mot juste
2014-04-26 14:52:36@pragdave Since there’s no Theos involved with TDD; that definition does not apply. His use of the term is, and is meant to be, offensive.
2014-04-26 22:01:14@unclebobmartin there are no kalashnikovs in TTD either, but you choose that definition. I think it's clear @dhh meant following blindly.
2014-04-26 22:33:06.@pragdave not using TDD or a better alternative seems equivalent to 'I don't believe in science' @unclebobmartin @dhh
2014-04-26 23:31:53@bdruth @pragdave @unclebobmartin Requiring test-first as the driver of design or you "don't believe science" is exactly the fundamentalism.
2014-04-26 23:38:07@bdruth @unclebobmartin @pragdave "Writing your unit tests after the fact is not TDD. It has nothing to do with it", https://t.co/TtmBUiYsYN
2014-04-26 23:54:03@bdruth @unclebobmartin @pragdave "Writing your unit tests after the fact is not TDD. Has nothing to do with it", https://t.co/TtmBUiYsYN
2014-04-26 23:54:43Additional context for TDD is dead: "Writing your unit tests after the fact is not TDD. Has nothing to do with TDD", https://t.co/TtmBUiYsYN
2014-04-26 23:57:40I completely agree with @unclebobmartin that you shouldn't be afraid of your code, but differ strongly on the tactics to get there.
2014-04-26 23:58:38"Every line of code as written as a result of a failing test" is just not an approach I find helpful. Neither is "decouple all slow" mantra.
2014-04-26 23:59:47@bdruth @unclebobmartin @pragdave Bob says TDD = Test-First. I say Test-First leads me places I don't want to go, thus TDD is dead to me.
2014-04-27 00:00:36@bdruth @unclebobmartin @pragdave So by defining TDD so narrowly, as in "only test-first", leads to its demise amongst many.
2014-04-27 00:01:36@bdruth @unclebobmartin @pragdave Yes, I've heard the sermons. Now we're back to square one.
2014-04-27 00:02:21@bdruth @unclebobmartin @pragdave My talk had examples. Have a look at that.
2014-04-27 00:05:29You're probably sick of hearing about testing. That's actually what inspired me. Sick of having every code design talk center around tests.
2014-04-27 00:08:28And sick of having every unclear architectural setup or pattern misapplication justified by "the tests made me do it", "it makes tests fast"
2014-04-27 00:09:36Jack TDDer: Sympathetic to the ideals of testing, not hung up on specific practices like test-first or mock-for-speed http://t.co/pGn9JAiH6l
2014-04-27 00:21:23@ketzusaka Yes, that's the dogma I'm pushing back against. "easy to test" = "you're doing it right" = "good code". Nope.
2014-04-27 02:17:45@ketzusaka I think test-first often leads to bad/convoluted designs. I think over-focus on test speed does the same. Two main tenets of TDD.
2014-04-27 02:27:23@ketzusaka Where I concur is with the value of having automated regression suite and confidence in changing your code.
2014-04-27 02:28:05@dhh @bdruth @pragdave Who has demanded that TDD is required as the driver of design? Tests are one among many such drivers.
2014-04-27 05:25:54@wanzuhao Tests are wonderful. Better tests are more wonderful. Thanks for helping!
2014-04-27 06:10:29@thatrubylove Thanks. We all err from time to time. Tackle the ball, not the man, as they say. s/ball/idea/g. All good.
2014-04-27 06:14:22@unclebobmartin @bdruth @pragdave Traditional TDD is sold primarily as a design tool, secondarily as regression protection. Many bought it.
2014-04-27 06:20:14