[LEAPSECS] drawing the battle lines

[Warner Losh]

On Mar 20, 2013, at 2:20 PM, Rob Seaman wrote:

>>> 	2.   a continuous reference time scale corresponds to UTC without leap second discontinuities;

> 

> And also corresponds to UTC with leap seconds.  There are no discontinuities.


discontinuities here means "irregularity" not the a violation of the contrived continuity of the variable radix UTC.


>>> 	3.   the concepts of continuity and uniformity should be applied strictly in a reference time scale;

> 

> This is not a fact, it is a statement of policy and requires detailed definition.


Viewed from a fixed-radix standard, a leap second is a discontinuity and a non-uniformity.


>>> 	4.   the unit for any quantity in metrology is unique, and as such, a single time scale should also be unique;

> 

> Two things.  The SI-second is derived from a more fundamental frequency standard.  There are and will remain vast numbers of time scales that are directly or indirectly layered on the SI-second.

> 

>>> 	5.   in the event of a redefinition of any quantity in metrology, the unit should be invariant, and particularly for the second of the Système International the respective scale should be continuous and uniform;

> 

> This scale already exists in TAI.  If TAI has issues, it isn't obvious why UTC needs to be changed.  Does TAI actually have issues?  It is an opinion, not a fact, to assert that practical time scales need be uniform.  And uniform with respect to what?


TAI isn't disseminated.


>>> 	6.   the name “Coordinated Universal Time” be maintained in the case of a redefinition of UTC without leap second adjustments;

> 

> This is an opinion, not a fact.  Many disagree with their opinion.





>>> 	7.   the term “Universal” in “Coordinated Universal Time” implies that the time scale is to be used throughout the world;

> 

> No.  There is no term "Universal" in Coordinated Universal Time.  Rather UTC parses as Coordinated "Universal Time".  "Universal Time" is a prior term that has always been approximately equivalent to "Greenwich Mean Time".  Relying on sophistry in definitions reflects a weak underlying position.

> 

>>> 	8.   the term “Coordinated” in “Coordinated Universal Time” implies coordination among National Metrology Institutes and not a relationship to the direction of the Sun from a position on the surface of the Earth;

> 

> The second half of this sentence belongs with "Universal Time".  Again, definitions are not facts.

> 

>>> 	9.   the angle UT1 used to relate celestial and terrestrial reference systems should not be considered as a time scale, but as the angle that characterizes the variable rotation of the Earth;

> 

> UT1 can be both an angle and a time scale.  That the rotation of the Earth and other bodies varies does not invalidate the identification of the word "day" with "synodic day".


UT1 is a time realization of an angle. It is an irregular time scale because it is based on an imperfect oscillator whose frequency error and time error are not predictable. That's what makes it not a suitable times scale.


>>> 	10. the International Earth Rotation and Reference Systems Service (IERS) provides a means of accessing UT1 in real-time by means of routinely available predictions of UT1-UTC with precision 100 000 times better that the coarse approximation UT1 = UTC currently provided by means of coding UTC to match UT1 within 0.9 second;

> 

> This is irrelevant and is meant to imply that any issues with implementing a redefined UTC will be minor.  They will not be minor for my community.


If you take the product of inconvenience and probability of that inconvenience summed over all users, you'll find the change has a low expected impact.


>>> 	11.  a wider dissemination of UT1-UTC is to be encouraged;

> 

> Again, a policy, not a fact (whether or not desirable).

> 

>>> and further recommends

>>> 

>>> that the International Committee for Weights and Measures (CIPM) and the International Telecommunication Union – Radiocommunication Sector (ITU-R) consider the formation of a joint working group to study the possibility of coordinating future actions in the definition of the continuous world-reference time scale.

> 

> This may well be a good idea, but perhaps additional institutions should be involved?

> 

> The function of point 9 (in combination with others such at #5) is to argue that civil timekeeping doesn't need to remain tied to Earth rotation.  Rather, the fact is that time in society depends on both atomic and synodic time scales.  To control "Universal" "Time" (as opposed to "Universal Time") they first must argue that a single time scale can rule them all.

> 

> It is not obvious why TAI does not already fill this role.  And if not TAI, define some other continuous time scale (by whatever definition of "continuous") under a different name, and leave UTC (~ GMT) alone.


TAI is a paper clock. It has no real-time realization. UTC is a real-time realized clock, and has traceable chains of time transfers.

Warner