[LEAPSECS] WP7A status and Re: clinical evidence about time and sun

Rob Seaman seaman at noao.edu
Mon Dec 22 09:25:03 EST 2008

Tony Finch wrote:

> On Fri, 19 Dec 2008, Rob Seaman wrote:


>> To synchronize two clocks (Earth and Lunar in this case), you can

>> adjust

>> the rates on one end or the other, or you can reset the zero point of

>> one or the other on some sort of schedule. Additionally, if the

>> differential rates continue to vary, then the scheduling has to vary.

>> If the clock rates are too far apart, the best solution is to put two

>> clocks on the wall.


> As well as the different rates of rotation of the various extra-

> terrestrial bodies, isn't there also the problem of differing

> relativistic

> frames of reference? (Is the magnitude of the difference big enough to

> matter?) The local atomic clocks on the Moon or Mars will not run at

> the

> same rate as a time signal transmitted from the Earth. So you will

> need to

> put THREE clocks on the wall, or maybe four: local atomic time, local

> rotational angle, terrestrial atomic time, and perhaps also

> terrestrial

> rotational angle.

In addition to Zefram's excellent comments, there are several ways to
approach this.

1) The references from this message a couple of weeks ago:


rather thoroughly put the last nail in the coffin of overreaching
assertions about the origin of leap seconds having anything to do with
relativistic effects.

2) Also, Dave Mills has been working on the issue of timekeeping in
space in the NTP context:


This adds a nicely engineered shell on top of the physical model.
Basically, the ephemerides of the planets and the known velocity and
location of spacecraft allow the relativistic (Newtonian and Einstein)
corrections to be computed.

3) My own point of view focuses on the requirements for "wall
clocks". Civil timekeeping has (heretofore) been mean solar time
because the clocks in question are general purpose clocks on people's
walls, wrists, laptops and phones. For instance, our conference rooms
have multiple clocks because the observatory has sites or partners in
Hawaii, Arizona, Chile, Illinois and a corporate HQ in Washington, DC
(no surprise there). Separate wall clocks are needed for human

Whether four clocks as you list are needed is a question of whether
civilians need to keep track of atomic time versus earth orientation
time (times two planets). Until now, the answer has been "no",
because atomic time has been a technical timescale that only folks
like us have to keep track of.

4) The ITU proposal is basically an assertion that people don't care
about mean solar time. (We can continue arguing that point in some
other thread.) They would answer "no" to your question since a key
point of the notion is to substitute one time scale for another while
recycling the namespace.

5) The zoneinfo idea is to preserve civilian access to mean solar
time, while providing enhanced expert access to atomic time (interval
time of whatever origin). It does not imply that both clocks have to
be displayed - or indeed, that one would choose to represent the data
in the same format at all.


More information about the LEAPSECS mailing list