This article addresses the problem of computing $\circ(\frac{x}{y})$ from $\circ(\frac{1}{y})$, targetting single and double floating-point precision.