From: Derek Bell <dbell@maths.tcd.ie>
Subject: Re: (urth) Re: lunar calendars [Digest urth.v028.n014]
Date: Sat, 28 Aug 1999 22:15:38 +0100

In message <199908122009.NAA08988@lists1.best.com>, John Bishop writes:
>Mantis says:
>
>> Lunar calendars have thirteen months (13 x 28 = 364).  Period.
>
>Not that I ever read.  The islamic calendar still offically
>depends on a witness to the new Moon, the Judaic calendar used
>to (before it could rely on calculation), and the Classic 
>world's lunar calendars did also (in particular, the calendars
>used to calculate possible eclipse times), though in later
>times astronomers calculated a sequence of short and long
>months that would be predictable (by reference) but never
>more than half a day off the real new Moon.

      The Islamic calendar is also exclusively *lunar* based with a
mean year shorter than the solar year. I once did a
back-of-the-envelope calculation about when the two calendars would
have the same year simultaneously - sometime after the 30th century,
IIRC. 

>See Abell's astronomy text and various encyclopedias.  I once long
>ago programed up a calendar maker which used modern data to construct
>the 19 and 57 year calendars, but I don't know if the Classic ones
>would have had the same distribution of longs and shorts.

       I'd also suggest _Calendrical Calculations_ by Dershowitz and
Reingold - a web site with sample material is at
http://emr.cs.uiuc.edu/home/reingold/calendar-book/index.shtml

	Derek Bell

*More Wolfe info & archive of this list at http://www.urth.net/urth/