[time-nuts] Re: UTC - A Cautionary Tale

Rob Seaman seaman at noao.edu
Sat Jul 16 11:20:53 EDT 2005


On Jul 16, 2005, at 12:07 AM, Poul-Henning Kamp wrote:

> No, you can not tell me today how many seconds between now and  
> 2010-01-01 00:00:00 UTC and that is the whole problem.

That is *part* of the problem - a part that is intrinsic to living  
within a non-inertial reference frame.  Folks who need interval time  
such as you describe should use TAI.

> The entire point here is to make UTC be a usable timescale without  
> having to put all military personal, all civil servants, all school  
> teachers, all emergency response personel and all programmers  
> through a reeducation.

UTC is a usable timescale - in fact, it conveys two usable  
timescales:  TAI for interval time and an approximation of UTn for  
time-of-day.  In point of fact it conveys TAI with at least two  
orders of magnitude higher precision than UTn - it is already more of  
a TAI standard than a UTn standard.  All of the people you describe  
expect civil time to be responsive to the rotation of the Earth, few  
care about leap seconds one way or another.  Most clocks are, and  
will continue to be, unaffected by leap seconds since they are  
embedded in systems that require resetting more frequently than the  
secular drift can build up.  Most clocks and users of clocks are  
deeply affected, however, by the underlying nature of the civil  
timescale.  Try to "reeducate" the public to understand that each day  
will no longer begin and end at midnight - that the prime meridian is  
drifting out to sea - that "atomic" clocks (with all the 20th century  
baggage of that word) are more important than mother Earth.  The  
"leap second deauthorization proposal" is not only a naive technical  
"solution", it is astoundingly bad public relations for the precision  
timing community.  Having castrated the world's clocks, should the  
PTC expect new funding to be forthcoming to clean up the mess they've  
created?

Those clocks and systems and projects and people that *are* affected  
by leap seconds should have implemented the civil time standard that  
was in effect when the clocks were designed and built.  A  
professional builds to spec even if he has a philosophical  
disagreement with it.  The real issue here is that the "medium- 
precision" timing community built lots of clocks and clock-like  
systems without ever bothering to investigate the underlying issues  
of timekeeping.  The clocks aren't built to spec, because nobody read  
the specification.  It was the ITU's responsibility to promulgate the  
UTC specification.  Are we now to trust them when they want to  
abandon it?

There will be no need to reeducate anybody if the civil time standard  
is left unchanged.  Part of what is going on, however, is the simple  
and inevitable *education* of technical personnel whose projects have  
run afoul of previously unexamined real-world constraints.

Rob Seaman
National Optical Astronomy Observatory



More information about the time-nuts mailing list