Accepting both fixed-point and floating-point numbers