The problem you describe here arises from using binary fractions, that is, a power of 2 as a denominator. You cannot represent the decimal fraction 0.1 as a bin

The problem you describe here arises from using binary fractions, that is, a power of 2 as a denominator. You cannot represent the decimal fraction 0.1 as a binary fraction. You would have the same problem representing 1/3 with decimal fractions. It just does not work. You can solve it by switching to decimal floating points, they are defined by IEEE as well: https://en.wikipedia.org/wiki/Decimal_floating_point https://en.wikipedia.org/wiki/Decimal_floating_point