[galbar]

I have said many times to teammates: the only code that is perfect is the one that hasn't left our minds. The moment it's written down it becomes flawed and imperfect.

This doesn't mean we shouldn't try to make it as good as we can, but rather that we must accept that the outcome will be flawed and that, despite our best intentions, it will show its sharp edges the next time we come to work on it.

[moffkalast]

I'm sure some mathematicians would disagree.

[bonoboTP]

Math can be greasy and messy. Definitions can be clumsy in a way that makes stating theorems cumbersome, the axioms may be unintuitive, proofs can be ugly, they can even contain bugs in published form. There can be annoying inconsistencies like optional constant factors in Fourier, or JPL quaternions.

Yes, prototypical school stuff like Pythagoras are "eternal" but a lot of math is designed, and can be ergonomic or not. Better notation can suggest solutions to unsolved problems. Clumsy axioms can hide elegant structure.

[pjc50]

I think applied mathematicians started to encounter this reality of the impure world the first time someone taped a dead moth into the logbook of the Harvard Mark II.