A floating point number represents an exact rational number. Trying to convey precision of the value is a mistake that many people do when designing fp systems. Please stay away from this. Gaston.