This is a note about a result from Xavier Gourdon : http://numbers.computation.free.fr/Constants/Algorithms/nthdigit.html see the bottom of the page. He has found a way to exploit the Newton-Leibniz series for Pi : (arctan(1) = 1 - 1/3 + 1/5 - 1/7...) and make a relatively efficient algorithm to compute Pi in BASE 10 at the n'th position. The efficiency of the alogrithm is sub-quadratic. He could reach the position 4,000,000 in 2 days of computation on a PIII with a very small amount of memory. This result is interesting since it opens the door (if well implemented) to distributed computing of Pi over the internet and many users. The result is somewhat based on my previous results on pi in base 10 and also Bellard improvement. http://www.lacim.uqam.ca/%7Eplouffe/Simon/articlepi.html in 1996 I had found a way in O(n*n*n), Bellard improved the result and made a better program in O(n*n) and now we have something better, that's neat. This is remarkable that the more we find efficient algorithms for Pi the simpler are the formulas to get the digits. Simon Plouffe