I'd like to do calculations with floats and at some point equality of two number will be checked. What is the best way to make sure that equality of floats will be detected, where I assume that mismatches beyond a certain point are due to truncation errors?