[LEAPSECS] pick your own length of second

[Warner Losh]

On Jan 11, 2012, at 2:33 PM, Tom Van Baak wrote:


> Here's a plot that shows how a non leap second UTC would

> look if the cesium resonance were other than 9,192,631,770.

> 

> http://www.leapsecond.com/pages/ut/ut-ani-v2.gif

> 

> In retrospect it's too bad |DUT1| had to be so tight. If Essen

> and friends had made it 10 s we wouldn't need leap seconds

> in a lifetime.


Yea, I sure like 9,192,631,950 a lot better than ,770. :)  But then again, any value > 840 or so would have required positive leap seconds if UTC was defined like it has been, but the SI second's value just slightly off.  '770 has a fractional frequency error of 2e-8, which is about what we'd expect from 1.3ms/century drift from the Ephemeris Second's origin in 1820.

Warner