Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83 (MC840302); site kuling.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!mcvax!enea!kuling!andersa
From: andersa@kuling.UUCP (Anders Andersson)
Newsgroups: net.games.trivia
Subject: Re: Answers to my trivium...(trivium?)
Message-ID: <759@kuling.UUCP>
Date: Thu, 14-Feb-85 18:23:57 EST
Article-I.D.: kuling.759
Posted: Thu Feb 14 18:23:57 1985
Date-Received: Wed, 20-Feb-85 01:56:12 EST
References: <1595@gondor.UUCP> <874@ihuxk.UUCP> <328@sdchema.UUCP>
Reply-To: andersa@kuling.UUCP (Anders Andersson)
Distribution: net
Organization: Computer Systems Laboratory, Upsala
Lines: 20
Summary: 

> As I recall, there is an inherent (negative error) of about 3 days every
> 400 years.  This is solved by NOT having leap years in years that end
> in "00", unless the year is divisible (sp?) by 400.  Thus there is
> a leap year in 2000, but not in 1700, 1800, 1900, or 2100.  The errors
> past this point are so small that they just add or subtract the 
> minutes (or seconds) between years.

The method of excluding exactly three years from being leap years within a
period of 400 years will result in a mean value of 365.2425 days per year,
quite close to the length of the tropical year, 365.2422 days. The error
will not be more than 1 day in about 3000 years, which probably seemed
small enough to Pope Gregorius' mathematicians (and does to me too).
However, this error may NOT be corrected by adding or subtracting seconds;
some day in the future our descendants will have to exclude yet another
year from being a leap year. The reason to manipulate with seconds is that
the Earth doesn't really care about Homo Sapiens' idea of how long a second
should be when rotating around her axis, and that's a completely different
problem.

/Dolphin