My biases may be guessed at in my summarization, but let me make them more explicit.
I think it odd that some of the techniques he rails against are inspired by mathematics: TDD tries to preserve invariants; RESTful design tries to impose the invariance of idempotency. If the question is whether those techniques make those who use computers more productive, then a scientific answer would involve a stunningly expensive human subjects experiment involving large numbers of people and complex problems. The likely result would be this: https://xkcd.com/1445/
I'm a sysadmin doing my best to automate (e.g., puppet); I rarely have the luxury of collecting data sufficient to the immediate problem, so I rely on math and (unreliable) heuristics. I write perl/shell/puppet/ruby/anything in small fizz-buzz complexity chunks; an "artisan" if you will. I support CAE environments with low-latency and poor parallelism opportunities, and until that changes (e.g., becomes cloud compatible) I don't see my tactics changing significantly.
I think it odd that some of the techniques he rails against are inspired by mathematics: TDD tries to preserve invariants; RESTful design tries to impose the invariance of idempotency. If the question is whether those techniques make those who use computers more productive, then a scientific answer would involve a stunningly expensive human subjects experiment involving large numbers of people and complex problems. The likely result would be this: https://xkcd.com/1445/
I'm a sysadmin doing my best to automate (e.g., puppet); I rarely have the luxury of collecting data sufficient to the immediate problem, so I rely on math and (unreliable) heuristics. I write perl/shell/puppet/ruby/anything in small fizz-buzz complexity chunks; an "artisan" if you will. I support CAE environments with low-latency and poor parallelism opportunities, and until that changes (e.g., becomes cloud compatible) I don't see my tactics changing significantly.