Please Look before you Leap a second

by Michael S. Kaplan, published on 2008/05/03 03:01 -04:00, original URI: http://blogs.msdn.com/b/michkap/archive/2008/05/03/8451692.aspx


I am going to talk about Leap Seconds for a bit. I'll explain why I am doing so after I blather for a bit.

Leap seconds are essentially a force of correction between two different philosophies of time measurement that are fundamentally different, with those differences causing the results to seem further apart over time.

The models?

Well, think of it this way.

There are two different ways of looking at time (and the passage thereof):

Method number one looks at time in terms of the duration of a day -- the fundamental idea being that by staying in line with the rotation/spinning of the earth, one is in touch with the units of time that pass. In this model, the DAY is the most important unit of time. A new day is just that -- a new day -- and it being 86,400 seconds from the same "time" on the previous day is interesting to some people but mostly it isn't as interesting as thinking about how I get up each morning and go to bed each evening. The second was indeed thought of as the fraction 1/86400 of the mean solar day. I can look outside and see the time of day and that all makes sense.

Method number two looks at time in terms of the duration of a second -- the fundamental idea being that by staying in line with the international System of Units (also known as SI), which defines the second as "The second is the duration of 9 192 631 770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium 133 atom, at rest at a temperature of 0 K." So if I multiply that value by 86,400 then know how long a day is. I can look at my computer or my watch and see the seconds tick away and I know that all makes sense, too.

So we have two ideas that both make a lot of sense, that are looking at essentially the same thing, right?

Now we hit the problem.

Because while atoms behave quite precisely (as SI indicates), the planet does not. The planet is (usually) slowing slightly, and therefore to make sure that these two different systems of measuring time make it back in sync, the idea of Leap Seconds was added. As described in Wikipedia:

A leap second is an intercalary, one-second adjustment that keeps broadcast standards for time of day close to mean solar time. Broadcast standards for civil time are based on Coordinated Universal Time (UTC), a time standard which is maintained using extremely precise atomic clocks. To keep the UTC broadcast standard close to mean solar time, UTC is occasionally corrected by an intercalary adjustment, or "leap", of one (1) second. Over long time periods, leap seconds must be added at an ever increasing rate (see ΔT). The timing of leap seconds is determined by the International Earth Rotation and Reference Systems Service (IERS).

When a positive leap second is added at 23:59:60 UTC, it delays the start of the following UTC day (at 00:00:00 UTC) by one second, effectively slowing the UTC clock.

Now it is interesting to watch the debate on the proposal to abolish leap seconds which is being voted on now in the wider context of all this, sure.

But I am thinking of with my Heinlein hat on, and of course method one looks like the "old way" of doing things -- even though it gets more accurate than sundials and such, the products that go down the road of method two (watches and such) that are second based rather than day based definitely look like products of technology.

And with my Heinlein hat on I must consider the day when humans (if we manage to avoid blowing ourselves up) will be on other planets, at which point method two is clearly more sensible than any terrestrial-based corrections. Which may or may not be a reason to want the proposal to succeed, but certainly is a reason to recognize that the notion that humans in Alpha Centauri are unlikely to be picking up Leap Second corrections from Earth in 500 years. :-)

Now this all came up in the context of someone looking in the Microsoft Knowledge  Base at Q909614: How the Windows Time service treats a leap second, which is a great example of how a second-based system might choose to apply day-based corrections, which are otherwise completely out of scope for them to deal with. The person asking about Q909614 wanted to know what effect the proposal would have on Windows and NTP (Network Time Protocol) if it was adopted.

And the answer is simple: no effect whatsoever.

Leap Seconds here in this context are a fudge factor to the system that corrects computers based on the actual time. So if the motion passes to stop adding the fudge factors, then the system will do just what it did before, without modifying the time by the use of Leap Seconds.

So you may have really good reasons to want the proposal to be accepted, or really good reasons to want it rejected. But there is no effect on Windows using NTP.

In my personal experience, the clock skew I see on my Windows machines is so great that turning off the Time services for even a week will lead to a clock so far off of the time that I doubt I would have even noticed the presence of absence of Leap Seconds anyway. :-)

 

This blog brought to you by 𝄵 (U+1d135, aka MUSICAL SYMBOL CUT TIME)


HenrySkoglund on 3 May 2008 9:50 AM:

Even if we're not living on other planets yet, the're some spacebased human infrastructure today that benefits from going with method two, consider the GPS satellites for example.

(Because of method one-based leap seconds, they currently differ from UTC by 14 seconds)

Rgrds Henry

Jim DeLaHunt on 3 May 2008 2:50 PM:

I look at this as human idosyncrasy meeting the enforced simplicity of technology. This theme comes up in your blog a lot, mostly as cultural conventions meeting Windows structures. My own meditation on second-based time versus planet-based time is "Times Change, or, We Live In Complex Times". at http://blog.jdlh.com/en/2008/03/24/times-change/ .

John Cowan on 3 May 2008 3:11 PM:

"The planet is (usually) slowing slightly, and therefore to make sure that these two different systems of measuring time make it back in sync, the idea of Leap Seconds was added."

To clarify: even if the planet could be persuaded to straighten up and fly right, we'd still need leap seconds, because the SI second was defined in terms of the year 1900 C.E., which was an unusually short year.

We *need* leap seconds because SI seconds are shorter than 1/86400 of an average day right now (they were approximately right in 1865 or thereabouts).

We can't *predict* the leap seconds very much in advance because the planet behaves unpredictably.  (The steady slowing of 1.7 ms/century from tidal friction is quite predictable: it's the smaller variations that make leap seconds messy).

We haven't *had* any leap seconds lately (causing an overflow in the GPS count of weeks since the last one, which was limited to 255) because the planet has been running fast of late.

moksha on 14 Mar 2011 8:30 AM:

please give the explanation of the proverb[look before you leap]


referenced by

2012/07/03 Please Look before you Leap a second (take down your servers edition)

go to newer or older post, or back to index or month or day