Time zone rules change

Paul Eggert eggert at CS.UCLA.EDU
Thu Sep 14 17:02:42 UTC 2006

Jesper Norgaard Welen <jnorgard at prodigy.net.mx> writes:

> the article says that Ramadan presumably will start the 24.th. of
> September 2006, and on the eve of that day, they will switch back to
> winter time.  The rule change should therefore be  Sep 25 0:00.

and "Paul Schauble" <Paul.Schauble at ticketmaster.com> writes:

> It may not be midnight. Some Middle Eastern countries have the next day
> starting at sunset.

"Christmas Eve" is the day before Christmas, and that's how I
interpreted that story's "eve".  I chose midnight since our DB has
previous-year changes at midnight.  If the change is Sept 25 00:00
rather than Sept 24 00:00, or is at some other time than 00:00, I hope
Hammad Hassan will correct us.

Robert Elz <kre at munnari.oz.au> writes:

> it is pretty clear that the "eve of Ramadan" rule for ending
> summer time won't last past the next few years, 2011 at the latest
> I'd guess, but perhaps more likely just this year and the next couple,

My guess was different: that it affects only this year, because the
otherwise-as-usual change directly interferes with the preparation for
the start of Ramadan this year, but wouldn't in other years.  It's
just a guess of course.

Oscar van Vlijmen <ovv at hetnet.nl> writes:

> (My calculations, based on Reingold/Dershowitz "Calendrical Calculations"
> algorithms).

I came up with the same results that you did.  Here's the source code
to my calculations, in GNU Emacs Lisp (tested under version 21.4).
The output is text that you can feed directly to zic.

   (defun generate-ramadan ()
     (let ((year 1427))
       (while (< year (+ 1427 33))
         (let ((date (calendar-gregorian-from-absolute
                      (calendar-absolute-from-islamic (list 9 1 year)))))
           (insert (format "Rule\tEgypt\t%d\tonly\t-\t%s\t%d\t 0:00\t0\t-\n"
                           (extract-calendar-year date)
                           (calendar-month-name (extract-calendar-month date) 3)
                           (extract-calendar-day date))))
         (setq year (1+ year)))))

I came to the same conclusion that you did, which is that a lunar
calendar is not likely to be the basis of a daylight saving scheme.
(Though it might be the basis of a _moonlight_ saving scheme, for
those of us who admire moonlight!)

More information about the tz mailing list