Comparison
Broken window theoryvsNormalisation of deviance
Broken window theory
the first person skipped the test, and within a month nobody in the directory was writing them.
Visible neglect invites more neglect: one unfixed mess signals that mess is acceptable here. In a codebase it is why standards decay faster than anyone decided, and why a failing test left red for a week is more expensive than the test. It is an argument for fixing small things immediately and, more importantly, for not letting a broken build become normal, which is the same mechanism as normalisation of deviance.
Full entry →Normalisation of deviance
the deploy has warned about that for months and everyone knows to ignore it.
The process by which an unsafe practice becomes normal because it has not yet caused harm. Each individual acceptance is rational — it worked last time — and the standard drifts until the anomaly is invisible to the people inside it. It is the mechanism behind most large organisational failures, and the practical countermeasure is periodic outside eyes plus taking small deviations seriously while they are still small.
Full entry →