Kent M Pitman  <pitman@world.std.com> wrote:
+---------------
| Leap seconds are what bother me the most.  The fact that politically
| an organizational body can insert an extra second between two times
| means that any representation of a future time is virtually, not
| actually, correlated to the calendar.
+---------------

Are you sure you're not talking about Daylight Savings Time instead of
leap seconds?  Except for the one discontinuity in 1972 when everyone
switched to UTC (Coordinated Universal Time) -- which *did* require
"political" agreement among a bunch of nations -- leap seconds are only
added as needed to keep "noon" within a second of noon, especially when
*non*-periodic (and therefore unpredictable by direct extrapolation *or*
political bodies!) variations in the rotation in the Earth occur. And
they're only added (or removed, but so far none have been) at midnight
December 31st or June 30th, so a fairly small vector of (4 bits/year
since 1972) can be used to adjust.

Note: I'm not sure how far in the future they can reliably predict when
leap seconds will have to be inserted. Maybe the following URLs will help:

	http://bul.eecs.umich.edu/uffc/fc_utc.html
	http://www.boulder.nist.gov/timefreq/faq/faq.htm#8
	http://www.gfy.ku.dk/~iag/iers95.txt
	http://www.grdl.noaa.gov/GRD/GPS/DOC/arc/leap.html

Looking at this last one, it's pretty clear that the insertion rate of
leap seconds is *not* a simple periodic function...


-Rob

-----
Rob Warnock, 8L-855		rpw3@sgi.com
Applied Networking		http://reality.sgi.com/rpw3/
Silicon Graphics, Inc.		Phone: 650-933-1673
2011 N. Shoreline Blvd.		FAX: 650-964-0811
Mountain View, CA  94043	PP-ASEL-IA