Amusing! I wonder if this is related to any of the examples in the paper "High Precision Fraud" by the Borwein's: http://www.jstor.org/pss/2324993 Victor On Sun, Mar 14, 2010 at 11:30 PM, Bill Gosper <billgosper@gmail.com> wrote:
I was showing a kid n 2 i k pi ==== -------- \ n > gamma(e r) / ==== k = 1 limit ------------------------ = - %gamma, 0<r<1, n -> oo n
because it approximates a contour integral of Gamma(z)/z around 0, modulo the additional crudity of replacing
pi sin(--) n pi ------- by --, i pi n ---- n e
when he observed that the crudities somehow cancel: n = 1/r = 69 gives 126 digits! --rwg _______________________________________________ math-fun mailing list math-fun@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun