Re: [LEAPSECS] Monsters from the id

From: Rob Seaman <seaman_at_noao.edu>
Date: Mon, 16 Jan 2006 19:22:59 -0700

On Jan 15, 2006, at 10:40 PM, John Cowan wrote:

> I realize the ALHP has severe problems with this, but I don't
> approve of the ALHP anyhow (save perhaps tactically, as explained).

Nothing to add to this. I just continue to enjoy the fact that folks
with completely opposite points of view about civil timekeeping have
the same low opinion about leap hours :-)

> Indeed. But the sensible approach would be for each State
> government to fail to omit the hour of the normal spring transition
> in the year 2700, say. In that way, AEDT would become TI+1000 and
> a normal-looking autumn transition would cause AEST to become TI+0900.

You get points for calling it "TI" instead of "UTC".

> Countries without DST transitions would have to actually repeat an
> hour, of course, just as Algeria had to do in 1940, 1956, 1977, and
> 1981 (the country has repeatedly flipflopped between UTC+0000 and
> UTC+0100).

...and Algeria had the freedom to do so precisely because UTC existed
to function as a worldwide civil timescale that continued
uninterrupted "in the background" while the local authorities
extemporized. Presumably you would assert that TI (TAI + constant)
could serve this same purpose. That was presumably the notion of the
folks in Torino. I'm delighted they recognized that UTC should not
be redefined. I disagree that interval time can indefinitely serve
as a stand-in for solar time. It is when the first leap hour (or
timezone migration event) occurs that interval time fails the test.
And precisely because it is a timescale that is designed to simply
tick, tick, tick in even intervals.

> By the way, I re-counted all the secular time zone transitions
> worldwide. According to the Olson timezone database, there have
> been 516 of them since the beginning of standard time (when that
> is, of course, varies with the country or subdivision thereof).

I think we're using the word "secular" in different ways:

        secular 3 a : occurring once in an age or a century
        b : existing or continuing through ages or centuries
        c : of or relating to a long term of indefinite duration

The ~2 ms per century lengthening of the solar (or sidereal) day is a
secular trend. Overlaid on this trend are all sorts of interesting
more rapid periodic and aperiodic effects. Whatever adjective one
might attach to the 516 ad hoc time zone transitions - secular isn't it.

Rob Seaman
NOAO
Received on Mon Jan 16 2006 - 18:23:15 PST