> I think this has a undesirable quality: ... Oops, ignore that - now I see Christopher's experiment I realized I misread the algorithm. Sorry for the noise, Darren