basically, arithmetics with an added limit to precision. as it stands, it seems useful mainly for those who could benefit from assumed uncertainty to every operation on numbers (as some numbers always get lost which every step of solving and equation)