Someone please explain this, my brain is too dainty and full of frolicking poetry to understand the joke.
Has to do with floating point addition. https://adacomputerscience.org/concepts/number_floating_point_arithmetic
Way I understand it, computers can internally represent floating point numbers only as binary approximations with a finite number of binary digits. So since mantissa and exponent are stored only with an arbitrary precision, rounding errors happen.
Hope that's basically correct.
(post is archived)