On 03/06/2013 09:13 AM, Patrice Scattolin wrote:
Since we have no idea of emacs' algorithm
We have the source code, so we can get an idea. The code can be found here: http://bzr.savannah.gnu.org/lh/emacs/trunk/annotate/head:/lisp/calendar/cal-... The function calendar-islamic-to-absolute is the key part. I read it, and found that GNU Emacs uses the most-popular arithmetical approximation to the Islamic calendar. By "most-popular" I mean that it's the approximation used by the majority of the Moslem world, according to page 84 of Dershowitz N, Reingold EM, Calendrical Calculations, 3rd ed. (2008). Cambridge University Press, ISBN 978-0-521-70238-6. Dershowitz and Reingold mention that there are other arithmetical approximations. In particular, some Microsoft products use a variant approximation. So it appears that no matter which approximation we'll use, we're bound to disagree with somebody's else's prediction. I'll soon circulate a proposed patch to the 'africa' file that makes predictions based on the GNU Emacs approximation.