On Tue, 25 Sep 2007, John Fouhy wrote: > You've got upper and lower bounds - maybe you could do a binary search > to find the max exactly? It should only take the same number of steps > again... I thought of that; and then I thought I'd rather go home and have dinner.