A050296: Eating crow!

[John Layman]

Sometime over the weekend I realized that a blunder was likely in my
calculations communicated last Friday, but email facilities were not available
at the time.  I now see that Rob Pratt quickly uncovered my errors.  

Rob, your results are very impressive.  Would you be willing to make your
algorithm available somehow?  If I use my original fail-proof algorithm
(instead of my faulty efficient algorithm) I find that for n somewhere in the
range of 35-50 the calculation takes longer than I'm willing to wait (days).

John