As others have frequently indicated, the current UTC standard would permit a
lot of flexibility in implementing different scheduling strategies. While I
personally find options with a frequent cadence (eg, alternate positive and
negative intercalary events) to have attractive qualities - after all, this
describes NTP - this flexibility could alternately be used to better stabilize
the long term scheduling as PHK has suggested.
Let's talk about relaxing the bounds on DUT1, not about eliminating them
entirely.
On Jan 7, 2011, at 5:54 AM, Ian Batten wrote:
> What stops leap seconds being applied at local 2am?
A key question. Ultimately the whole point of UTC is to serve as - well - a
universal clock. UTC is the same everywhere. Timezones are different
everywhere. The former is stepped globally. The latter, locally.
This distinction is why I don't like the notion of hiding the perpetual drift
in the timezones. A locality may decide that their local clocks not be
conformant with UTC - they cannot redefine UTC to their own purposes.
That said, many arguments here have also been made that most clocks are not
only ok with a certain amount of error - they are designed with that error in
mind. My wristwatch doesn't step at midnight UTC (or at 2am locally for that
matter). The essence of my position over the years is that it is more correct
to state that most clocks keep synodic time (mean solar time) than that they
keep UTC. The timezone overlay is an orthogonal issue.
Acknowledging the true requirements of a project is the biggest step toward
freeing it to explore useful and interesting variations. The former is part of
understanding the problem. The latter of posing potential solutions. One
problem, many solutions. Individuals or institutions are not able to redefine
the problem to their own purposes. The result of confusing the two was best
described by that great system engineer, Sir Walter Scott:
Oh! what a tangled web we weave
When first we practise to deceive!
Scott also said:
But time and tide o’er all prevail
I'm not saying civil time should be, or even "must be", based on the synodic
day. I'm saying it "is" synodic - a simple recognition of patent fact. The
requirement exists whether on not all seven billion of us pretend otherwise.
Or deem’st thou not our later time
Yields topic meet for classic rhyme?
On the other hand, leap seconds are on the solution side of the equation. We
are free to tweak them - even eliminate them - IF the requirements are
respected.
Rob
_______________________________________________
LEAPSECS mailing list
[email protected]
http://six.pairlist.net/mailman/listinfo/leapsecs
