I want, here, to give an alternative point of view. My view is that for software management in practice, there already exist more powerful ways of getting control over quality and cost parameters than the above-mentioned prediction models. The important question is, do you want prediction, or do you want control?