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

Zefram zefram at fysh.org
Thu Dec 18 17:43:22 EST 2008

Poul-Henning Kamp wrote:

>A timescale which takes earth rotation into account should be

>called "Terrestial Time Coordinated" (TTC ?) and the timescale

>that takes into account the rotation of Mars should be MTC.

The name "MTC" has already been used to refer, not to the Martian
equivalent of UTC, but to the Martian equivalent of GMT. Obviously, the
people who did this are among those who don't think of the leap seconds
in UTC, but instead use "UTC" as the label for vague UT. They were in
fact people at NASA working on the Mars rover missions.[0]

I think the fact that professionals in such a closely-related discipline
can't keep UT and UTC straight says something about the viability of
working with leap seconds. It chimes with how we've seen all manner
of technical systems, even including NTP implementations, misbehave
around leap seconds. It also matches my experience in editing the
Wikipedia articles on time scales: my plan to merge [[leap second]]
into [[Coordinated Universal Time]] was overwhelmingly rejected, by
people to whom the distinguishing feature of UTC is its Universality,
rather than its Coordination.

The Martian time scale in question, the local solar time on the Airy
meridian, is better described by the older name "Airy Mean Time" ("AMT").
We have a pretty good idea of the time on that timescale now, to within
a second. (Though we did not have that precision when the rovers landed,
which explains the odd offsets of the mission timezones.) Creating a
Martian equivalent for UTC is a much more complex job: it requires the
concepts of Marticentric Coordinate Time (TCM) and Martian Time (TM),
and an atomic time scale Martian Atomic Time (TAM) that realises TM.
That'll require actual atomic clocks on Mars. Then MTC would have the
job of coordinating TAM with AMT (possibly also to be named MT1).

I find that the mental exercise of the previous paragraph drives home
just how artificial UTC and TAI are compared to UT1. Pondering MTC
also gives some insight into the far future of UTC: current MTC, if we
could implement it, with TAM ticking SI seconds, would already require
more than one leap second per *minute*. A minute of AMT occupies about
61.65 SI seconds. We'd certainly need some arithmetical scheduling.

>But the universal timescale should depend on nothing that is not

>uniform throughout the Universe. A good choice would be an

>easy to measure and well defined atomic resonance under well

>defined relativist circumstances.

Current physical theory does not identify any preferred reference frame.
The closest time scale we have so far is TCB, which is based on a
reference frame in which the barycentre of the solar system is at rest.


[0] Michael Allison and Robert Schmunk, "Technical Notes on Mars
Solar Time as Adopted by the Mars24 Sunclock", Dec. 13 2005,

More information about the LEAPSECS mailing list