On Apr 9, 2008, at 6:53 AM, Poul-Henning Kamp wrote:
And the interesting one of course is:
http://www.iers.org/plots/FinalsAllIAU1980-UT1-UTC-BULA.png
This is an up to date version of the first figure (labeled figure 3)
from my screed of exactly seven years ago today:
http://iraf.noao.edu/~seaman/leap
Remarkable how issues that aren't resolved keep coming back at you :-)
How to optimize the scheduling algorithm for leap seconds has always
been one of the more interesting - and potentially more productive -
questions, whether it is your 10 year look ahead or the other extreme
of permitting monthly leap seconds as I described.
What clever chap was it who suggested ALWAYS having a leap second at
the end of each month - toggling positive and negative? An actual
leap transition then becomes the omission of a (negative) leap. Since
systems would have to deal with the monthly cadence they would be
forced to work correctly. (Evolution trumps design any day.) And
since a leap would be the absence of an event, nothing would break.
More to the point, all of these strategies are permitted under the
current UTC standard :-)
Should we make an informal bet on when the next leap second happens ?
I'll take December 2008. A Carlsberg Elephant (what do the locals
drink?) versus your choice from the Nimbus brewery in Tucson?
Rob
---
http://nimbusbeer.qwestoffice.net/thebeers.html
_______________________________________________
LEAPSECS mailing list
[email protected]
http://six.pairlist.net/mailman/listinfo/leapsecs
