Answer from cs61c-ad (Maung Maung Aung 16764789) for Question 4
If we follow IEEE754 to do the floating point arithmetic, there is chance to have some innaccuracy because of rounding.