The only thing I can think of that might be an issue is precision. Remember that binary is not good at representing floating point values. So the square root of 8 might be 2.8284 or it might be...