Note that f + 1 > f can also fail for integers in many languages (e.g. it does not hold for signed integers in C or C++, because the behavior is undefined when you add 1 to INT_MAX, and unsigned integers always wrap around). This particular gotcha is not unique to floating-point.