[LEAPSECS] All of this has happened before

[Rob Seaman]

On Jan 28, 2015, at 10:02 AM, Poul-Henning Kamp <phk at PHK.FREEBSD.DK> wrote:

>> We each wear multiple hats.  Two of mine are 1) to point out that
>> physical reality trumps standards and software,  [...]
> 
> And one of my hats is to point out that you have no monopoly on
> defining "physical reality" and have a great tendency to define it
> to support your opinions.

Physical reality defines itself :-)

Is either assertion in my previous email incorrect?  The residuals on the y-axis of these plots from the IERS attest in great detail that Mean Solar Time differs from Atomic Time:

	http://www.iers.org/IERS/EN/Science/EarthRotation/LODplot.html

And length of day most assuredly derives from the synodic day, i.e., mean solar time.  Is there not one fewer day in the year than sidereal rotations?

> Likewise, as for your personal "physical reality" isn't trumping
> anything, unless you're also the one paying the piper.

We have leap seconds (or would have some other accommodation) due to the accumulation of the residuals in the IERS plots.  But those residuals are as small and manageable as they are precisely because the duration of the SI-second was chosen to closely mimic the fraction 1/86,400 of a solar day.  Over the history of UTC the residuals have stayed below 4ms.  That is within 1.00000004629 of the (varying) length of a solar day.  This is not coincidence.  And LOD is not a free parameter.

I think, rather, that it is my statements about systems engineering that you wish to redefine.  Leap seconds are a means to an end; one could consider other means.  However, attempting to ignore engineering requirements outright is a recipe for trouble.  I infer bigger trouble than you are willing to admit.  But the people who are not paying the piper are the ITU.  It would have been significantly less expensive / more effective to have spent the past 15 years performing coherent systems engineering than in single-minded pursuit of a pretense that two different things are the same.

I am delighted at the recent small engineering project surrounding DNS as a conveyance for Bulletins C & D.  It seems likely to lead to something at least modestly useful.  The tzdist group is making great progress with a larger project as Steve has described.  Is there anything else this group can do along these lines, for instance to verify and encourage uptake of the fixes to the issues encountered during the 2012 leap second?

Rob