On Wed 2007-03-07T20:04:04 +0000, Zefram hath writ: > In the absence of DST or > other changes (the story predates DST), "day" in a civil context and > without any qualifier must be taken to mean civil days at any fixed > location. Which location doesn't make any difference, of course, when > using the day only as a unit of duration. You'd have a hard time arguing > in court to measure solar days at a moving location.
But this seems exactly the de-motivation for the IAU and IERS folks paying any attention to the notion of Universal Time anymore. The current paradigm for earth rotation eschews the validity of any geographic entity as a reference point (other than in a least-squares sense for defining the underlying and more "Platonically ideal" ITRS). This notion has been growing ever since BIH was given authority to monitor all broadcast time signals, and it can be traced back at least as far as remarks between Janssen and Newcomb at the International Meridian Conference. In a world with tectonically moving plates there can be no location which defines "day" at the microsecond level. And, just to confuse the notion of day and night further, a pair of papers by a guy who notes that even the IAU 2000 resolutions preserved a lot of "classical cruft" with the implicit use of an ecliptic in the precession model and the new explicit definition of an equator. http://arxiv.org/abs/astro-ph/0611781 http://arxiv.org/abs/astro-ph/0611782 The upshot being that if you throw away all those classical notions you can fit the earth rotation using 31 parameters instead of 46000. Of course in this scheme there is only the earth and the quasars; the sun need not apply. It's also not clear whether the method leads to anything like a useful long-term predictive model. -- Steve Allen <[EMAIL PROTECTED]> WGS-84 (GPS) UCO/Lick Observatory Natural Sciences II, Room 165 Lat +36.99845 University of California Voice: +1 831 459 3046 Lng -122.06025 Santa Cruz, CA 95064 http://www.ucolick.org/~sla/ Hgt +250 m _______________________________________________ LEAPSECS mailing list [email protected] http://six.pairlist.net/mailman/listinfo/leapsecs
