Hacker News Reader

I wonder what percentage of bugs in general are from people not understanding how floating points work. I think property based testing (QuickCheck) should be used whenever floats are involved. Nobody ever seems to get them right.

View on HN

At least -0.0% are.

Would QuickCheck know to try 999999 as input? Most of the possible inputs would give correct results, and those that don't, aren't very 'special', such as 0,1,-1,MAX_INT, and so on.

Don't many QuickCheck-inspired libraries have special cases to ensure they generate common numbers like those? I could be misremembering, but I would have sworn I read that in the documentation of the last library I looked at (whose name escapes me).

Sorry, I phrased myself poorly - in this case, the inputs that give incorrect output aren't very special, as opposed to 0,1,INT_MAX,...

Although including numbers such as 10^N - 1 isn't out of the question either.

Reply on news.ycombinator.com