Re: [LEAPSECS] Some personal notes on the Torino meeting

From: Markus Kuhn <>
Date: Tue, 03 Jun 2003 17:41:50 +0100

Steve Allen wrote on 2003-06-03 02:26 UTC:
> On Mon 2003-06-02T17:58:36 +0100, Markus Kuhn hath writ:
> > backgrounds, one of the first presentations that gave a case for
> > abandoning leap seconds was given by William Klepczynski. He took the
> > ADS-B (Automatic Dependent Surveilance Broadcast) system as an example,
> > in which GPS receivers installed in civilian aircraft broadcast their
> > own position and velocity vectors, such that any receiver can implement
> > a "virtual radar" from this data. He gave ADS-B as an application
> > example, where a high-precision time scale has (or is about to) become
> > an element of a real-world safety-critical air-traffic control system,
> > and where the parallel use of several time scales such as UTC and "GPS
> > time" could lead to rare and difficult to test malfunctions related to
> > UTC leap seconds. UTC would be used to display data to operators
> > (pilots, controllers) in real time, but a uniform timescale will also be
> > needed to solve motion equations.
> This begs the question:
> If the system presumes that it has access to GPS, why would its
> system design bother to do any operations or data exchange in UTC?
> Within the system the designers should choose a single timescale that
> makes sense, and for such systems UTC is obviously not it.

The term "system" as used in today's engineering practice is usually
nothing but a gross oversimplification or even a verbal joke on the
original meaning of the word ("organized or connected group of objects
or principles"). Real world "systems" are bewildering semi-organized
hierarchies of subsystems with independent history and situated in
disjoint domains of administrative responsibility. It would be close to
impossible to avoid or at least contain a global standard such as UTC in
a "system" as large as the global air traffic control infrastructure.
ADS-B and GPS are just examples of the many data sources that need to be

> The Galileo GNSS is going to use TAI, not "GPS time"

There are those who will argue that at one level of abstraction TAI and
"GPS time" are sort of the same thing, whereas at another level of
abstraction TAI(GNSS) and TAI(BIPM) are completely different things
(different enough to cause tens of meters difference in pseudoranging
solutions). Others will actually look into the specification and read
there that GPS was specifically designed to provide you with UTC, which
makes UTC the "GPS time". I'm not sure where terminological nitpicking
will lead.

There was some discussion at the meeting on people easily confusing
different levels of abstraction, in particular when people who worry
about time offsets in the order of a second talk to people who worry
about time offsets in the order of a nanosecond.

> I don't see how crippling UTC resolves the problem of having multiple time
> scales, and adding "TI" into the mix in 20 years only makes that problem
> worse.

Me neither, though I do understand the logic and assumptions behind some
of the arguments. There are systems for which having a ubiquitously used
single leap-second free time scale would be an advantage, and there are
others where a close coupling of the time to UT1 is of an advantage.
Some believe, the former are sufficiently more important to justify a
mildly radical break with civilian timing tradition. There are others,
who believe that this will not eliminate the problem that we need both
uniform time and earth-orientation time widely diseminated.

> > One claim made was that in many
> > real-world systems that run on UTC, the insertation of the leap second
> > is today actually delayed until the current operation has finished. This
> > leads temporarily to a loss of synchronization by one second for some
> > time after each leap second.
> I hope that they bothered to name these systems, mostly because it
> would set a precedent in the history of this mailing list. So far
> nobody has named a system here which cannot tolerate leap seconds.

If memory serves me right, this came out of William Klepczynski's talk,
but I don't recall specific examples being given there. I hope the
slides of all the talks will be posted at some point.

> I have previously posted about civil lawsuits that might arise when
> the difference between GMT and "leapless civil time" diverge by more
> than a few seconds.

Ron Beard's view on this was that we should at present focus on
technical needs and technical difficulties. Legal needs and legal
difficulties should be discussed later and separately by lawyers and law
makers and can in this particular case probably be solved fairly easily
by updating the relevant legislation to catch up with technical
developments. That sounded very sensible to me, though I did observe in
the past that countries with a napoleonic legal tradition (most European
countries) are considerably faster in updating laws to match technical
progress in science and metrology than countries that follow more the
British legal tradition. Just look at the sorry state of the use of SI
units in United States legislation to get the idea.

> The announcement that Galileo will fly implies that atomic time will
> be available just about everywhere on the surface of the earth.

Receiving weak UHF satellite signals inside buildings/tunnels/basements/
inner-cities/mountain-valleys, where much equipment of interest is
located, is not always a trivial and ultra-low-cost issue. That's one of
the reasons, why I try to keep satnav enthusiasts reminded that
low-frequency time-dissemination and navigation systems equally need
continued improvement and updating and should have a long and prosper

> For applications deep within buildings and underground, the long wave
> broadcasts such as WWVB could continue. They do not provide voice
> modulation, and are for rather specialized applications. For the
> purpose we use our WWVB receiver, I can say that we would not mind
> if the coded modulations started reporting TAI instead of UTC.

With "our WWVB receiver", I suspect that you mean some relatively rare
piece of specialist lab equipment found in observatories that you can
order for just a few thousand dollars. The price of a DCF77 receiver
(the Central European equivalent of WWVB) is now down to 1-2 euros or
dollars. More than half of all clocks sold in Germany today now contain
one. All the wall clocks in the university building in which I work
contain MSF receivers (the UK variant). If you changed the time in the
LF transmitters to TAI or TI, you would change defacto the basis for
civilian time.

> Unless I'm mistaken, none of the above time broadcasts is regulated by
> existing ITU recommendations.

On the contrary! ITU-R TF.460 is exactly all about what type of time
these transmitters are supposed to provide, and all the time
transmitters operate in bands that have been allocated by the ITU for
exactly that purpose.

> That leaves short wave radio, which has
> historically been of great importance to navigators who actually need UT1.

The beauty of short-wave radio is only that you can get reasonably good
global coverage of both transmitter and receiver have medium-sized
antennas via ionospheric reflection. For really strong global coverage
to get predictable delays with a single transmitter and a simple
low-cost receiver, you should use the lowest frequency available to use
the entire atmosphere as a waveguide, preferably send a few MW at
~20 kHz, as is used for communication with military submarines.

> If atomic time becomes available via all these other means, how
> relevant are the recommendations of the ITU-R regarding "time signal
> emissions"?

I'm not sure I understand exactly what you are aiming at here, but I
believe that international standardization is a very good thing in

I fully agree that all time dissemination means should be upgraded to
provide both uniform and earth-orientation timing information, and these
upgrades should be done based on international standards that still need
to be written. That is the technical issue. Whether the definition of
civilian times should be changed to follow uniform as opposed to
earth-orientation time is an orthogonal issue. I am not yet convinced
that there is a strong case for doing it, but if it is done, I could
certainly live with it.


Markus Kuhn, Computer Lab, Univ of Cambridge, GB | __oo_O..O_oo__
Received on Tue Jun 03 2003 - 09:44:47 PDT

This archive was generated by hypermail 2.3.0 : Sat Sep 04 2010 - 09:44:54 PDT