Re: System.Time.Clock Design Issues

On Feb 4, 2005, at 8:34 PM, Ashley Yakeley wrote:

In article <6c0017bff5beac04cf224af544c840b4@comcast.net>, Gregory Wright wrote:

...

In essence, civil time in the US before the introduction of UTC had seconds which were usually 1000 SI milliseconds long, some that were 1020 SI milliseconds and a few that were 1060 SI milliseconds.

Did they ever have shorter seconds, 980 or 940ms? UT1 wobbles all over the place.

Yes, on five occasions shorter seconds were used to advance the timescale. These used seconds with lengths 940 ms, 950 ms, 900 ms and 999.8 ms (yes, a 200 microsecond advance). The retarded seconds had lengths of 1020 ms, 1005 ms, 1100 ms and 1001 ms. (There was no retarded second of 1060 ms as I mentioned in my last note; the 60 ms was a clock advancement and I added when I should have subtracted.) Before 1960 20 ms steps were common, happening almost every month. After 1960 until the adoption of UTC 100 ms steps were most common. (The objective was to keep broadcast time signals within 100 ms of UT2. This procedure was followed from 1962 until the introduction of the leap second scheme with UTC in 1972. More info in the Explanatory Supplement to the Astronomical Almanac, section 2.57 et seq.) Greg

-- Ashley Yakeley, Seattle WA

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries