Ken Pizzini ken at ember.spry.com
Fri Nov 1 01:25:41 UTC 1996

In <199610032320.QAA04986 at shade.twinsun.com>, Paul Eggert
<eggert at twinsun.com> said:
> I also used to believe that UTC was established in 1972, but I've been
> corrected.  UTC was introduced in 1961.  It was originally kept close
> to UT1 by periodically adding or subtracting steps of a fraction of a
> second.  1972 is when the current leap-second regime was instituted.

Hmmm, so POSIX's epoch *is* well defined then.  Do you know if there
are any records of when and by how much the various UTC adjustments
were made, particularly for the interval 1970-01-01T00:00:00 until
the first leap second (1972-06-30T23:59:60)?

I'm interested in revisting the issue wollman at uvm-gen.EMBA.UVM.EDU
raised in <9402062245.AA18471 at bajoran.emba.uvm.edu>:
> ...According to the bulletins that I've seen, |TAI-UTC| is about
> 28 seconds now.  However, the leapseconds file only lists 19 steps.
> This is probably due to the fact that the first step was 10 seconds,
> not just 1.  Is there an easy way to fix this, so my users who use NTP
> don't complain about their clocks being nine seconds off?

[Note to new readers: this is somewhat misleading.  The "first step" of
10 seconds was actually the defintion that UTC-TAI would be exactly 10s
when the new leap second mechanism was introduced.  This was because
UT1-TAI was about 10s at the time and UTC had been tracking UT1 and
so was therefore also about 10s off of TAI at that time.  Any actual
stepping (or leaping) involved was of a fraction-of-a-second nature.]

Given the "UTC established 1972" misunderstanding at the time, this
was considered an awkward problem to address.  But it sounds like it
may actually be possible to invent some "phantom leap seconds" which
are backed up by some version of "truth" in an implementation of the
suggestion made by Paul Eggert in <9402072149.AA05650 at spot.twinsun.com>:
>       B.  Invent some leap seconds for the period between 1970 and
>       1972, as if UTC had been in effect then.  This lies about UTC
>       but repairs the 2 or 3 s error.  The conversion correction
>       would be correspondingly reduced, to 7 or 8 s.

Provided, hat is, that someone can find records of how UTC was kept
almost-in-sync with UT1 prior to 1972.

		--Ken Pizzini

(NB: referenced articles are in the tzarchive file on elsie).

More information about the tz mailing list