[fitsbits] Leap Second questionnaire

[Steve Allen]

On Tue 2002-06-25T05:10:22 -0400, Arnold Rots hath writ:
> Also, remember that there is this unnecessary time-scale, GPS, that
> leaves the strong impression that the GPS folks had not realized the
> existence of the leap second mechanism - if they had, they would (?)
> have adopted TAI.

I believe that there might be another, and more reasonable explanation
for the difference between GPS time and TAI.

For practical purposes GPS time is 19 s behind TAI.  However, GPS time
is not exactly 19 s behind TAI.  GPS time is kept by a different set
of clocks than TAI/UTC.  Its exact offset from TAI/UTC drifts.  The
specifications call for the drift not to exceed 1 microsecond, though
lately it has managed to stay within 50 nanoseconds.  A polynomial
giving the current estimate of the difference between GPS time and UTC
is broadcast as part of the signals from the satellites.

The current value of GPS time is something that can be obtained
instantly, but the current value of TAI/UTC is not actually available
until about a month from now when the clocks have all been averaged.
If the GPS system had adopted a time which was offset by TAI by zero
seconds, then many users of GPS time might incorrectly come to believe
that GPS time was exactly TAI.

By choosing an offset, albeit a trivial one, the GPS system designers
have made it very clear that GPS time is not TAI.  The GPS system
designers appear to have chosen to make it necessary for users of
their signals to bother to understand how they must be interpreted.
This is a Good Thing (TM).

Unless someone knows of a better historical explanation, I believe
that the GPS system designers were astute enough to give them the
benefit of the doubt on this possibility.

>                                               Doing away with the UTC
> leap seconds turns UTC into an unnecessary duplicate of TAI.

Agreed.

--
Steve Allen          UCO/Lick Observatory       Santa Cruz, CA 95064
sla at ucolick.org      Voice: +1 831 459 3046     http://www.ucolick.org/~sla
PGP: 1024/E46978C5   F6 78 D1 10 62 94 8F 2E    49 89 0E FE 26 B4 14 93