[LEAPSECS] ACM article

[Rob Seaman]

Poul-Henning Kamp wrote:


> Tom Van Baak writes:

>>> Rob Seaman wrote:

>>> It is simply fact that time-of-day and interval timekeeping are two different things.

>> 

>> Help me out here. That ACM generated time-stamp in your posting; which is it by your definition: time-of-day or interval timekeeping?

> 

> Or is it by any chance an earth orientation estimate ?


Time-of-day is indeed an Earth orientation estimate.  Civil timekeeping is trivially recognizable as a realization of Earth orientation.  (Gedanken upon request.)  UTC currently provides this.


> Rob Seaman writes:

>> 

>> 

>> 

>> Universal time is time-of-day.  The current definition of UTC permits it to be used to recover an interval timescale.

> 

> To claim 40 years later that timeintervals are merely a unintentional sideeffect of the UTC definition, is a very tough row to hoe, and you're not even close to making it.


UTC provides a realization of both Universal Time (= Earth orientation = time-of-day) and interval time (as TAI).  I have never attempted to suggest that evenly tempered SI time intervals are either an unintentional or unimportant side effect.  They are, however, indeed a *side* effect of any time scale in the happy UT family of timescales.

The ITU wishes to change this fact.  Obviously I object.  However - if the ITU does seek to change UTC into something other than a realization of Universal Time, then blatantly obvious system engineering best practices say that they should plan to address the consequences.  You guys claim without evidence that there are no possible consequences.  Astronomers know otherwise since you are aiming to break all our software.

There is also the inconvenient little fact that civil time does just happen to be based on time-of-day as I have described it over-and-over, that our calendars count synodic days, that mean solar time is the point of the exercise, and that apparent solar time has nothing to do with it.  The ITU might be able to cheat - if they plan for the contingent issues.  But it is a cheat, plain and simple.  The radiocommunications community does not own time-of-day.

Tony Finch wrote:


> Rob Seaman wrote:

>> 

>> The current definition of UTC permits it to be used to recover an interval timescale.

> 

> Not if you are scheduling events in the future.


Righto!  Time-of-day and interval timekeeping are two distinct things.  No single system can convey both with 100% self-consistence.  "Let's stop issuing leap seconds and see what happens! Yippee!" is not a coherent plan for attempting to resolve this conflict.

Rob