[time-nuts] Re: UTC - A Cautionary Tale
seaman at noao.edu
Fri Jul 15 23:17:01 EDT 2005
Warner Losh says:
> We already have ambiguity in when something occurs, as defined by
> Earth. Each timezone is 15 degrees wide, and thus something may
> happen at 11:59:59pm local standard time, but really happen at
> 12:01:01am the next day 'solar' time.
Ambiguity cuts both ways. Standard timezones provide a simple
mechanism for converting between local time and the single worldwide
standard of UTC/GMT/Zulu time. Standard time is no less "real" than
local time. Mean time is no less real than apparent solar time. I
have no argument with having a single worldwide standard - I just
argue that the support that UTC supplies for both time-of-day AND
atomic time is a better standard than abandoning half of the
equation. If we have to abandon anything, abandoning TAI would be
the better choice.
> We lost earth local time when we went to a standard time years ago.
No - by standardizing the meaning of the terms, we made it possible
to easily convert between all the flavors of solar time using closed
form algorithms accurate to whatever precision is required. Give me
the explanatory supplement to the astronomical almanac - or one of
Jean Meeus' books - and I can code up conversions between dozens of
physically interesting timescales. The main fallacy in this whole
debate is the idea that a single timescale can be useful under all
> That introduced 30 or more minutes of ambiguity between the mean
> local solar time and the standard time. Given such a large
> ambiguity that people accept today, it is hard to believe that they
> can't accept a few more seconds (oreven minutes).
This argument confuses periodic with secular effects.
Straightforward algorithms (a few lines of C) can convert standard
time to local time and mean time to apparent time. It is the
proposed abandonment of time-of-day (whether called UTC, GMT or Zulu)
that will introduce the ambiguity you are concerned about. Rather
than some Chebyshev approximation of arbitrary precision that can be
relied upon for years or decades or centuries in advance (precisely
because it is keyed to a priori knowledge of the orientation of the
Earth relative to the Sun) we will only be able to make predictions
with elaborate extrapolation of civil (TAI+34s) time to UT and
involved numerical integrations over the intervening epochs. Read
what Jean Meeus has to say:
Civil time should remain based on time-of-day. Failing that, civil
time should be called anything *except* UTC. They can call it UBT
for "Universal Bureaucratic Time" if they want.
The proposed change would be a fundamental shift in the philosophy of
time, not some little bureaucratic realignment as it is being marketed.
National Optical Astronomy Observatory
More information about the time-nuts