U.S. Moves to Kill Leap Seconds

blacklite001 writes "Not content with merely extending Daylight Savings Time, the U.S. government now also proposes to eliminate leap seconds, according to a Wall Street Journal story. Their proposal, 'made secretly to a United Nations body,' includes adding 'a "leap hour" every 500 to 600 years.' Hey, anyone remember the last bunch of people to mess with the calendar?"

Re:Unfair to clockophiles!

[Anonymous Coward]

At least he's got http://leaphour.com/ [leaphour.com] as well.

Torino conference notes

[Anonymous Coward]

I found this mailing-list message to be very interetsing and informative:

http://www.mail-archive.com/[e-mail] .mil/msg00163.html/ [mail-archive.com]

Brief excerpt:

I also gave a presentation of leap second issues in distributed
computing, presented the UTS proposal and argued that something like it,
together with more carefully implemented NTP software, would in practice
eliminate computer worries about leap seconds, without a need to change UTC
arising from this area. I also argued that the message formats of
pre-GPS time broadcast services such as the various LF and HF time
stations leave much to be desired and that work on a globally
standardized state-of-the-art signal format would be a timely and
important project.

http://www.cl.cam.ac.uk/~mgk25/c-time/torino2003/u tc-torino-slides.pdf [cam.ac.uk]

Finally, on the afternoon of the second day (Thursday, 29 May), the
agenda moved to writing up a draft conclusion of the colloquium, which
was then to be refined and phrased out more carefully by the
invitation-only SRG meeting on Friday.

Ron Beard with William Klepczynski drafted in PowerPoint on the
presentation laptop a list of objectives and conclusions for the
meeting. They started out with a few very pro-change statements, that
quickly attracted criticism from the audience as perhaps not being a
quite adequate reflection of the discussion at the colloquium.
Throughout the subsequent discussion, I had the impression that they
were rather happy to include pro-change arguments and statements that
were proposed by participants into the draft, but were very reluctant to
include any of the more sceptical/conservative statements that were, as
far as I could tell, proposed equally often. In the following coffee
break, a number of participants noted on their impression that the
organizers of the colloquium probably had already made up their mind on
the death of UTC and would push this through ITU in any case.

The connected geek question

[astrashe]

The article talks about lots of problems that leap seconds cause with software.

The problems don't come from the complexity of the underlying problem of adding leap seconds, but rather because leap seconds are added so infrequently that the code to handle the leap seconds isn't well tested.

So the real question here (to me, at least) is this: what do the leap second problems tell us about how software is developed?

Are people not thinking about leap seconds when they write code? Or are they thinking about them, but not testing the leap second cases properly? What's going on?

And how does the emergence of really big collections of APIs affect this? I mean, if people use standard routines for calendar functions, and if people keep their tools up to date, shouldn't these problems be mitigated? Shouldn't we be able to have some hard core calendar geeks solve the problem once in the API, and carry the rest of us?

If that doesn't work, why not?

We can solve this particular problem by changing the calendar. But what if we couldn't, and we had to try to address it with engineering practices? How would we proceed?

Re:now correct me if im wrong

[Tekoneiric]

The problem is that it isn't working fine. To begin with we should have 13 months in the year, not 12. Months are supposed to reflect lunar cycles and there are 13 of them a year. The year is one day and some change longer than 13 (28 day) months a year. Ever noticed how the business world works off 13 periods a year? and of course the menstrual cycles too. Take a look at this [wikipedia.org] sometimes.

Re:Apparently not...

[Entrope]

Resetting the clock is not complicated, but the current system means there is a 61st second in a minute, as your quote of TFA mentions. People -- including software developers -- are strongly used to dealing with 60-second minutes, and software sometimes makes that assumption. It just requires attention (sometimes a lot of attention) and extra code (sometimes a lot of extra code) to get it right, but since very few people pay attention when a leap second happens, bugs are easily overlooked.

Since leap seconds are based on changes in the time period of Earth's rotation (the sidereal day), and the decay is both very slow and influenced by hard-to-predict factors, leap seconds are not reliably predictable. They can only be announced when they are necessary -- and so it is easy for the displayed time to drift if a leap second announcement is missed or ignored.

Leap hours, though, are different beasts. Virtually every piece of software in the world that displays time knows how to deal with the hour jumping forward or backward. That transition happens predictably and affects a huge number of users, so errors are easily noted.

Last bunch of people?

[nwbvt]

The Romans were the last bunch of people to mess with the calander? You mean nothing concerning it has changed since the fall of the Roman Empire? I seem to remember something about some guy named Gregory in there somewhere...

And does this mean the Romans had leap seconds where they adjusted their atomic clocks to keep in synch with the sun?

I know much of /. will be complaining about how this is about the Bush Administration attacking science in their quest to please big business, but in reality from a purely scientific stance this makes sense. The definition of a second hasn't been linked to the Earth's orbit since 1967, so why should we keep on pretending it still is?

Re:Can we say what we will think 500 years from no

[Rich0]

In China, there is only one timezone, but it works terribly since half the country wakes up in the dark and the other half wakes up in bright sunlight. They have adapted to this by "unofficially" setting work hours according to the longitudinal timezone rather than the government-mandated timezone.

My feeling is that they should simply have a chronometer which keeps ISO standard time. Go ahead and use an hours-minutes-seconds based system so that people get used to it. Forget leap-seconds - no need for that. Forget time zones - no need for that either. We'd probably go to 24-hour time and ditch am/pm since they'd have little meaning in most regions of the world.

An office would set their working hours as 1830-0230 and that would be it. No changing the time in the summer/winter/etc. They could change their hours in the summer/winter though.

An office on the other side of the country might start work at 1700 instead.

There would be no official countrywide designation of starting and stopping time, although most people would expect businesses to be generally open between sunrise and sunset. In 500 years nobody will care that the whole clock has drifted an hour, since the number on the clock doesn't mean anything in the first place. It is just a reference, and it would work fine for that purpose under such a system.

I can't really think of anybody who would be negatively impacted by such a system other than traditionalists. Astronomers would be fine - their star-tracking software probably calculates everything in some internal time format anyway, since the leap-year/leap-second/23h56m business already makes the current 24-hour clock useless for them. If anything, the software would be easier to design since the rules would be deterministic.

13 moon peace calendar

[bennyp]

Why not use the more sensible and natural 13 moon calendar of the mayans?

Re:Heh

[Spetiam]

And false. We use the Gregorian Calendar, which was instituted by Pope Gregory XIII. So he's really the last one to mess with the calendar, and as far as I know, he died a natural death. [wikipedia.org]

Not that a little reality will get in the way of bashing the US...

Re:Lunar Calendar Is Better

[Anonymous Coward]

I said this in an other post. Lunar calendar will also have problems. But I think that will be much better. Here is an example from Hinduism Lunar calendar.

Ancient calendars were based on lunar months, but in order to keep the calendar in step with the seasons, it was necessary to insert extra months, because 12 lunar months are 10.8751234326 days short of a tropical year. The point to be noted is that the Vedic astrology paradigm uses the sidereal zodiac where the relative motion of the solar system itself, in this universe is noted and its precession has been measured @ 50.23 seconds of arc per year. This translates into an additional 20 minutes of time in a solar year.

    The Vedic Calendar is the oldest and tries to cover this shortfall of 10.87 days between 12 lunar months and a year by interpolating an extra month every third year called the Adhika Masa.

This was the first approximation and had an inbuilt error of 3.095 days in every 3 years that would tend to shift the seasons back by as much time. This inbuilt error can be rectified by having another Adhika Masa every 30 years i.e. every 30th year has two Adhika Masa (leaves an error of about 1.417 days in 30 years) and yet another every 625 years (i.e. every 625th year has 3 Adhika Masa).

There could be many more in different religions but not thoroughly understood. Instead of reinventing the wheel why not take a look at these different religious texts.

Re:Planet

[hazee]

Yep, once we get that space elevator working, we'll be able to ship huge amounts of rock up and down, adjusting the angular momentum of the Earth, and thus its spin rate...

I wonder just how much mass would be required to adjust the length of a day by the required fraction of a second per year?

Good question from a lazy asker...

[Bradee-oh!]

I'm too lazy to go Google it right now, but I think the point is pertinent/interesting to this crowd -
With our current system of leap seconds, does the Unix timestamp actually reflect the CORRECT number of seconds since Jan 1st, 1970?
 
Sure some of the Unices are probably different but I'm guessing that many of the implementations of the algorithm calculate the seconds with basic math using only leap years as the deviation from standard.
 
Ah, hell, maybe I'll go google it, too, but, I'll still ask here. :P

Re:now correct me if im wrong

[gstoddart]

The problem is that it isn't working fine. To begin with we should have 13 months in the year, not 12. Months are supposed to reflect lunar cycles and there are 13 of them a year.

Actually, the 12 months was to align with the constellations of the zodiac so that certain constellations will be in the same place at the same time. It keeps astronomical calendars in tune.

Cultures which slavishly kept to a lunar calendar (another method of timekeeping, but it ignores the fact that we revolve around the sun) found that every bunch of years their months would be in the wrong season.

A month is an abstraction made by humans for timekeeping, there is no 'should have 13 months' that closely aligns with actual astronomical time passage, which is far more important.

Keeping track of solstices and equinoxes are really important when it comes to things like knowing when your seasons are changing.

Re:Apparently not...

[sallen]

That won't work in applications where countrywide synchonization with a precision in fractions of seconds is needed, such as some same-frequency-broadcasts or GSM networks. On the other hand, those probably already have real solutions for this problem, instead of a kludge.

Bingo. You mean synchronization with precision like that which has been used for decades by the telco's (read: ITU jourisdiction). This is a problem what isn't one. Time is relative (sorry, couldn't resist.) I can't believe anyone, let alone Naval Observatory dude thinks in terms of sync to, essentially, a wall clock. That's why internal clocks are used. The time representation for someone looking at the 24 hr clock is simply a representation of an internal clock converted to something us dumb humans can relate to. Is this guy an idiot or what? If he'd been around many years ago, there wouldn't be binary systems, we'd all have to be on decimal systems, becuase he probably couldn't count either (That'll be his next recommendation.) And what happens when they try and add the 'leap hour'? I'm sure he looks at it like the deficit... he won't be around when it has to be taken into consideration. But I guarantee it'd make the millenium 'bug' (ie, original laziness) look like a cakewalk.
As for 'nobody uses a sextant' since we have good old GPS, tell that to the sailors not too many years ago who lost all nav equip and used hmm... a sextant. One NEVER abandons the ability to utilize alternate means of problem resolution. He figures there's no possibility GPS could ever fail or be subverted? Bad mistake. That's what kills people, not leap seconds. He'd probably be the one to say take INS out of planes since GPS works. Right. I'll never fly over the ocean if he's that ignorant.

It's easy to see where this is coming from...

[Stonan]

Potential errors in adding 'leap seconds' is causing screw-ups in computer systems. The main cause is sloppy programming so eliminating them makes everything better. Don't have to worry about it for 500-600 years.

Ask yourself who benefits from this. The only answer I can come up with is software programmers, specifically OS programmers (programs usually read what time the OS is reporting). Which OS manufacturer has the most clout with the US gov.? Which company is reported to have the most liquid cash? To take a quote from Mr. Moore: Who your Daddy?

Re:now correct me if im wrong

[Guppy06]

"Well, this isn't a radical change like decimal time, in that it will have zero effect on John Doe's wrist watch."

My watch and my alarm clock both set themselves from the signal broadcast by WWVB. That signal will tell my clock and watch if it's DST, whether there's a leap-second change coming up, and in what direction that leap second will go. There's no room for leap hours without changin the encoding standard for the radio broadcast. Changing DST has no effect on these timepieces (the radio signal indicates whether it's DST or not), but changing the definition of UTC will break all radio-controlled timepieces.

Additionally, unlike DST, leap seconds are applied globally at once, at 00:00 UTC. Where DST is applied in the middle of the night between Saturday and Sunday, leap seconds would be applied between the hours of 09:00 to 17:00 for literally 1/3 the planet, and at the end of a given month (potentially in the middle of a work week). This simply isn't workable when we're talking about entire hours.

I can support changing DST, and I could support abandoning UTC and just sticking with TAI, but I can't support this.

Re:now correct me if im wrong

[Peter La Casse]

If you made 13 months a year our calendar would go horribly out of sync.

Not necessarily. Make them 28 day months, and then in between Firstuary and Lastcember have a holiday (which doesn't get a day-of-the-week name) that lasts 1 or 2 days (depending on whether or not it's a leap year). Poof: the new 13-month year is exactly the same length as the current 12-month year.

The hard part is coming up with a name for the 13th month, and deciding where to put it. That would be a big political mess.

Re:Can we say what we will think 500 years from no

[Anonymous Coward]

Intercontinental communications would require a degree of pre-arrangement

Intercontinental communications already require a degree of pre-arrangement, it's just that someone already did it for you. Someone already coordinated working hours and time zones.

However, if there were only one time zone I would have to find out when people work, but after I found out when they work I wouldn't have to do any math to figure out what time that was to me.

I also can't figure out why you think businesses would adjust their schedules to local noon, as opposed to adjusting their clocks to matches the businesses that they work with. Companies with nation interests are already open from 8 AM EST to 5 PM PST, because that when they do business with are open. Why wouldn't these businesses continue to be open from the time that their eastmost clients open to the time that their westmost clients close?

The 9-5 standard didn't come out of daylight hours. 9-5 came out of daylight hours, but it became standard because there's a distinict economic advantage to being open at the same time as your business partners. I don't see any reason to think that eliminating time zones would change that.

The Naming of the Months

[Beolach]

Heh, I found this link [wikipedia.org] in the /. synopsis very interesting. The Naming of the months is something that has interested me & I've speculated on a bit. I knew that the first months were named for Roman Gods:
Janus [wikipedia.org]
Februus [wikipedia.org]
Mars [wikipedia.org]
Aphrodite [wikipedia.org] (actually a Greek goddess, but the Romans identified their [wikipedia.org] gods and the Greek gods together)
Maia [wikipedia.org] (another Greek goddess, the Roman name is Bona Dea [wikipedia.org])
Juno [wikipedia.org]

I also knew that July and August were named after Julius & Augustus Caesar. After August, the months are named with their numbers.
September (7)
October (8)
November (9)
December (10)

But wait! Those numbers aren't right! And here began my speculation. I figured the Romans (like most 10-fingered humans) were fond of 10 (X in Roman numerals), so they may have started with 10 months (which actually is the case). I also assumed that August and July were the last months added to the calender, based of their being named after Julius and Augustus Caesar (this assumption turns out to be false; January and February were the last months to be added to the Roman calender: the Romans originally considered winter to be monthless). I found the (incorrect, of course) conclusion of my speculation to be rather humourous: the Roman calender began with ten months, until Julius Caesar came along, and decided he was important enough that he deserved his own month, and so he created July. He wasn't arrogant enough to think he was more important than the gods, but he was more important than just a bunch of numbers, so he sticks July after the months named after the gods, but before the numbered months. That changes the numbering, but the names from the old numbering stuck. Augustus Caesar dittoed Julius Caesar.

Sadly, the explanation based on research rather than speculation that Wikipedia gives for the number mismatch is not so humourous. They simply say that March was originally the first month. But I always thought (incorrectly, it seems) that January was named for Janus, the god of Beginnings and Endings, because it was the first month of the year, that marked the end of the old year and the beginning of the new year. But even when January became the first month, it wasn't because of Janus, but rather because the Roman consuls had a year long term, and took office on the 1st of January.

Re:Can we say what we will think 500 years from no

[adolf]

I wake up and eat breakfast in the morning (after the sun comes up).

I eat lunch at mid-day (when the sun is roughly over head).

I eat dinner in the evening (usually when the sun is starting to descend).

I go to sleep at night (after dark.)

Does it really matter if I wake up at 0000 isntead of 0800? Does dinner taste differently at 1900 than it does at 1100?

Curious.