Addition of two floating point numbers is more complicated than you might imagine, mostly because the exponents have to be the same before the two numbers can be added. This forces the computer to do exponent adjustment which can completely zero out one of the addends. Overflow is also discussed, as is subtraction.