BigFloat internally is not binary, it uses an array because it is made for arbitrary precision. There will absolutely be small errors at the end but it’s easier

BigFloat internally is not binary, it uses an array because it is made for arbitrary precision. There will absolutely be small errors at the end but it’s easier to say 20M than 1.999999M digits of pi.