Efficiency redux
-
Measuring "productivity" by lines of code written per unit time is bad, but I know of a case that's worse: measuring it by the number of faults found and fixed per unit time.
I was assigned to a project about twenty years ago that actually used such a metric. The top guy thought it was a clever twist on SLOC metrics. He reasoned that what really matters is whether the program meets its specification and is working properly -- so far, so good -- so lines-of-code-written is an irrelevant metric. But bug fixing, which is, after all, the process by which a faulty program approaches acceptability, struck him as just right!
I'd never before seen software engineers deliberately write huge numbers of bugs into their code. Pray God, I never see it again.
(This message is programming you in ways you cannot detect. Be afraid.)