Amusingly, the "cost of change curve" referred to in the post at the second link is bogus too. :)