Xref: utzoo comp.databases:6781 comp.software-eng:4042
Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!zaphod.mps.ohio-state.edu!sdd.hp.com!elroy.jpl.nasa.gov!hacgate!ashtate!dbase!awd
From: awd@dbase.A-T.COM (Alastair Dallas)
Newsgroups: comp.databases,comp.software-eng
Subject: Re: Calendar Algorithm
Summary: CACM
Keywords: calendar algorithm
Message-ID: <669@dbase.A-T.COM>
Date: 10 Aug 90 03:18:37 GMT
References: <140352@sun.Eng.Sun.COM>
Organization: Ashton Tate Development Center Glendale, Calif.
Lines: 30

In article <140352@sun.Eng.Sun.COM>, jasonf@cetemp.Eng.Sun.COM (Jason Freund) writes:
> 	I need to be able to change dates in a database cell with commands like
> "+4w" == "Find the closest weekday to the day 4 weeks from a certain date", 
> "-12d" == "Find the closest weekday 12 days earlier than this date.", and so 
> on.  I'll be writing in Pascal -- the code is no problem -- I'll give my 
> function a "+/-", a number, and a "d/w/m" (day, week, month), and a date 
> "MMDDYY".  The function returns the closes weekday by adding/subtracting 
> according to its parameters.  Does anyone know of a general algorithm for 
> dealing with calendars?  It must that take into account leap years and be able 
> to find what day of the week it is (MTWRF) so that it can make approximations 
> to weekdays. I've seen some good ones before, (for computing how many days old 
> you are and what day of the week you were born on) but I can't find any now.  

The collected algorithms of the ACM has an algorithm for converting
m-d-y into an integer number of days from some start point a long time
ago, and back again.  That makes it easy to add days or weeks.  For
months, you're pretty much stuck adding to your m-d-y months and 
dealing with overflow into years.  The ACM book describes the weekday
values for the resulting integer modulus 7.  (That is, if no_of_days % 7
equals 0, it's Sunday, or Saturday, or something like that.)

> BTW, when is the next leap year?

That's easy.  Leap years occur every four years, except years evenly
divisible by 100.  The last exeception is that years evenly divisible
by 400 _are_ leap years.  Therefore, '92 and '96 are leap years and
for the first time in four hundred years, the turn of the century will
be a leap year.  All because the solar year isn't exactly 365-1/4 days.

/alastair/