At least that's why I voted for it...
That is simply false. There are an infinite number of algorithms that might contain the (sub)expression X/X for which zero is a valid value of X. To assume it's a programming error is sheer unmitigated stupidity that I might expect from a mathematician that has never written a real program in his life.
As someone with a degree in mathematics and a degree in computer science (with special academic honors, I might add), I strongly disagree. Fix your damn program to check for a dividend of zero, or at least trap the exception and handle it then. If NaN or any of the infinities are useful in your computation, do it outside the normal math libraries or choose a language that explicitly permits them.
To assert that it's not a programming error is sheer unmitigated arrogance that I might expect from a code monkey who barely scraped by his high-school math courses, assuming you even attended any.
And yes, I've made my living writing programs, many of which benefitted from my knowledge of higher mathematics.
... OpenSSL and SSSL protocol flaws
South Shore Soccer League has a protocol of their own?
... password reuse is a larger danger to users than is having a weak password.
The best of both worlds: use a six-to-eight word diceware password for your password manager, and generate a long, random password for everything else.
Another megabytes the dust.