Math::BigFloat provides control over the accuracy and precision of floating point values. Update: that way you wouldn't need "dynamic" format control anyway!