This doesn't have anything in particular to do with C/C++. It's how IEEE floating point works. Anything that provides access to an IEEE 754-compliant FPU will display this artifact. The author is confused and thinks this is some feature of C, betraying his lack of knowledge.