I fail to see how this is an attack on science or on Europe. That's just hyperbole. Making time consistent to prevent having to add time or remove time just makes sense. Time is a unit of measurement. We don't redefine the foot just because the new king has 10" feet instead of 12" feet. Time being connected to the sun is a relic of the days before accurate time-keeping methods. One day should always be one day.
To rebut my own argument, I suppose one could say that noon should always be noon, regardless of how long the day is. The prevalence of daylight-saving time in the US might disagree with that though.
> I suppose one could say that noon should always be noon ...
Yes, but by allowing our clocks to go out of sync with the natural time of noon, a relationship controlled by orbits rather than clocks, this proposal would prevent that. Over time noon (highest solar position in the local sky) would drift away from noon as given by a clock.
I especially like the part where this proposal's advocates dismiss those who use a sextant to locate themselves on the ocean. I personally required this clock consistency to navigate my boat during my around-the-world sail, after my hi-tech satellite receiver expired.
This is not to argue that there's only one good argument here -- both arguments are reasonable. Unfortunately, they differ in the most basic way.
If we stick to 86,400 seconds in a day and don't allow leap-seconds, eventually daylight will happen at midnight -- either that or we will require leap-hours or even leap-days (as happened in Europe during the transition from the Julian to the Gregorian calendars).
If we continue to allow leap-seconds, all sorts of chronological record-keeping will continue to be more complex than it needs to be.
As with all truly interesting problems, there's no easy answer.
How long would it take for noon to be mid-night (as in middle of the night)? Centuries? By then, one would hope we're dealing with the bigger issue of how to coordinate time between two or more different planets or solar systems. As it stands, a certain section of scientists have to do the translation between the length differences of Earth days and Martian days.
> How long would it take for noon to be mid-night (as in middle of the night)? Centuries?
Much longer. But the practical difficulties would appear right away, when the time "error" is still measured in seconds. As I said earlier, both sides of this issue have perfectly valid points to make.
Right now leap seconds are added well under 1/year (and no negative leap seconds have been added). Let's just assume we need 1 sec/year.
Even to get 5 minutes out of sync, that's 300 years. A leap-hour would need thousands of years. I think it's a rather weak argument that we should worry about noon being "out of sync" for thousands of years in the future.
> I think it's a rather weak argument that we should worry about noon being "out of sync" for thousands of years in the future.
Not if you're on the open ocean and need to know where you are, as I was every day for four years during my around-the-world sail. In that case, "noon" had a very specific, and very important, meaning -- each second of clock error equals 1/4 nautical mile of position error (or 1,519 feet).
I emphasize that both sides of this issue have very reasonable arguments.
Then keep solar time as a timezone and switch into that mode when you really need solar accuracy. This is a rather extreme edge case.
And actually, it can easily, very easily, be done by hand, even if we stopped adding leap seconds to UTC. Just keep a log of leap seconds and then apply them to your UTC clock.
Sticking to celestial bodies for time seems pointlessly complicated to satisfy a few edge cases that can be handled in better ways.
I think a leap hour makes sense because it could be integrated into DST (it wouldn't be an actual leap hour; you'd just change your timezone offset). If DST still exists hundreds of years from now.
In other words, given a point on Earth (or any planet) and it's angle relative to the sun it orbits, when that point aligns exactly with and faces the center of the sun, it is noon.
Consider this: The way everyone refers to time is "X Units before noon", much like the 24-hour clock (no am/pm), except 0 is noon. The key is that X must be flexible, it should have no maximum, and no minimum.
This is not literally true. The period between solar noons oscillates and varies up to tens of seconds away from precisely 12 hours, thanks to the differing orbital speed of the Earth between perihelion and aphelion. What you're referring to is "mean noon", true solar noon averaged over the course of a year. See "mean sun" : http://en.wikipedia.org/wiki/Mean_sun
More complications arise from the position of the Moon. The Earth's rotation is affected by tides (stronger when the Moon is at perigee), and by the Earth's orbit around the Earth-Moon barycenter. Precession of the Earth's rotational axis is also a factor. And many more astromechanical effects also subtly affect the timing of the apparent position of the Sun.
Locking our timekeeping to the actual position of the Sun actually proves to be an intractable problem. At some point, we need abstractions to simplify and assume that that the relation to the Sun is good-enough. Famously, the Julian calendar was good-enough for millennia until the relation to the sun drifted off true by more than 10 days. So arguing over leap seconds could be seen as rather trivial.
This is amazing stuff. Thanks!
I agree that arguing over leap seconds is trivial. We're at a good enough point now, assuming leap seconds are handled correctly, such that accounting for those seconds keeps us on track.
reminds me of "relative time" Jews and Muslims (and Japanese, until 1872) use.
This is a French wikipedia article I stumbled upon: http://fr.wikipedia.org/wiki/Shaot_zmaniot
It may be seven years old, but it's not irrelevant. People are still discussing this proposal. The article is the best explanation I've seen of this issue to date.
UTC and TAI are off by over 30 seconds; people can't accept a discontinuity during the switchover. So what we really want is something like TAI-35. But the real problem is coordinating everyone to switch timescales so we don't end up with half the world on UTC and half on TAI-35. Also, I've heard there are cases where people feel that they're legally obliged to use UTC because that's the legal definition of civil time.
TAI for the timekeepers and as an internal representation, UTC for display, much like how many computers keep time in UTC and apply an offset to display local time. Just because we have to deal with the legacy of POSIX doesn't mean this isn't the right solution.
Yes, there are the obvious problems like the uncertainty of scheduling future events in UTC because the insertion of leap seconds aren't known in advance. But such things should be scheduled in TAI, anyway.
Interesting discussions can still be interesting even if the proposal was ultimately rejected. The fact that this was discussed, and that the reasoning behind it was to benefit "lazy programmers" makes it relevant to me.
We should always be wary of fundamental shifts in the name of "safety", "security", or worse, sloth.
Our current way of measuring time is absurd. It starts as base 1000 with milliseconds, then base 60 with seconds and minutes, then 24 for hours, 7 for days, 4.3whatever for months or 28-31 however you want to track it.
I like that China only has one time zone for the entire country. We should all go by UTC times regardless of our geographical position. But we should also change our means of measuring time to work around whole base 10 numbers.
Single timezone for a large country works very well. It makes it much easier to arrange a phone call or meeting with someone on the other side of it (you still need to know what the difference is to know you're not asking them to be there before breakfast or after supper, but you never get the failure mode where you both think you've arranged a different time). And it can encourage businesses to be flexible about working hours, since mandating 9-5 in all their branches is obviously insane.
Admittedly, a single timezone for the world might be a bit far fetched. But there's nothing intuitive about 60 seconds in a minute, 60 minutes in a hour, then all of a sudden 24 hours in a day, and roughly 4.3 weeks in a month or somewhere between 28 and 31 days in a month. The only numbers in our calendar that are vaguely intuitive is 3 months in a season and 265 days in a year, and even those have problems.
Why must time be be earth centric? Imagine how hard it must have been for Copernicus to convince everyone to use his model. Lazy programmers isn't as compelling argument, but it has similarities.
http://en.m.wikipedia.org/wiki/Copernican_Revolution
Non-rhetorical questions: Does the Mars Curiousity rover use leap seconds? Does it have Martian time too?
> Does the Mars Curiousity rover use leap seconds? Does it have Martian time too?
Both the Mars Curiosity rover, and its operators, use the Mars day duration (which is 24h 39m 35.24409s), but with time measured in "normal" seconds. The JPL people directly responsible for this project even switch to a Mars clock in their personal lives.
> leap seconds present a "risk to air travel in the future" because a glitch might shut down traffic-control systems.
This downright scary if true - a system that's not robust and fault tolerant enough to withstand a leap second bug without putting lives in danger is going to suffer from other bugs which will put lives in danger. Since software development practices actually tend to be very paranoid when it comes to air travel, I suspect this is really a scare tactic to garner support for abolishing the leap second.
> the U.S. has proposed adding in a "leap hour" every 500 to 600 years, which also accounts for the fact that the Earth's rotation is expected to slow down even further. That would be no more disruptive than the annual switch to daylight-saving time
I'm incredulous. Daylight savings time effectively works by changing your timezone offset from UTC. UTC itself doesn't change. Daylight savings time is actually very disruptive to programs, and the solution is to just use UTC for everything. A "leap hour" would presumably require changing the UTC-equivalent itself. I guarantee you that no one will actually code for such a rare event, which will make it a big and scary flag day which will get postponed repeatedly while everyone prepares (probably by bringing a bunch of legacy programmers out of retirement if software still exists by then). The only thing appealing about this is that we'll all be long dead by then.
A "leap hour" would presumably require changing the UTC-equivalent itself.
Earlier I proposed exactly the opposite. Never change UTC, just redefine your timezone to create an apparent leap hour. Since we already do this twice a year in many countries, it should be less of a problem.
Interesting idea! So basically every several hundred years we would shift the center of time farther and farther away from Greenwich. That explains why the British are opposed. I was thinking timezones would stay relative to Greenwich and the "leap hour" would be thousands of accumulated leap seconds applied at once. Your proposal sounds better.
I'm definitely looking forward to this debate when/if we ever have a permanent mars colony. Will they still use earth/solar time? Use mars/solar time? Some compromise between the two? I imagine someone living on mars could give two shits if the earth/solar time was off by a few seconds, and I definitely don't want to be responsible for handling two entire time systems in my code.
40 extra minutes in a martian day definitely make things interesting. Going further, how will time work when we're stationed around Saturn's moons? We will probably need some universal units of time that work regardless of how long it takes you to rotate around the sun. (Stardates?)
Seconds since epoch is good enough for almost all purposes until you're moving at ~10% of the speed of light; they'll be fine for computer timestamps as long as acceleration and deceleration are gradual. So I don't think we'll need any new universal units of time, only new local units of time, and I'm not even sure about that. For mars, just extending the day until 24:40 = 00:00 seems reasonable enough. For Saturn's moons the sun is faint enough that keeping "earth time" is probably reasonable - whether it's solar day or solar night will make very little difference to most people's lives, I would think.
But adding these ad hoc "leap seconds" -- the last one was tacked on in 1998 -- can be a big hassle for computers operating with software programs that never allowed for a 61-second minute, leading to glitches when the extra second passes. "It's a huge deal," said John Yuzdepski, an executive at Symmetricom Inc., of San Jose, Calif., which makes ultraprecise clocks for telecommunications, space and military use...
Isn't this the tail wagging the Dog? Crappy programming leads the UN to redefine TIME for the whole plannet? There are few truly golobal, hitsorical markers of culture shared by human societies. Accross continents, races and millenia. Solar time is one of them.
How about: Don't fix what ain't broke. Fix that which is?
Sorry for the Rant
__________
Edit:
As pointed out elsewhere in the comments. This proposal was from 2005, and is apparently no longer on the table. Cooler heads appear to have prevailed.
you are wrong in your rationale because crappy programming cannot be fixed and together with the current time definition it causes big problems to many, on the other hand changing the time definition causes problems to whom? astronomy purists?
What are your definitions of "cannot be fixed" "Many" and "Problems"? Biological clocks and cellular metabolism are tied to daylight, amongst other things. Talk about programming that cannot be fixed! Does that make 6 billion astronomy purists ;) ?
Its all about point of reference.
When I worked with radar we liked GPS time, always increasing so its easy to get a difference. But for locating the sun and sun rise, earth time is useful (those griping astronomers have a point).
We figured it out, and it wasn't too bad. Some good time libraries would go a long way...
Leap seconds are a work-around to make the 24-hour clock keep time correctly. The reason so many problems occurred was a combination of insufficient programming, but also the affect an inaccurate system.
I'm not sure what a correct system should be, but I know that it can't be based on the Earth, the sun, or any one galaxy. In order to have accurate time, it must be external to all of these things, like a meter stick is to an object.
The possible problem with this is that there will come a time when "3:00 pm" is in the middle of the night, the time to wake up for work will always "change".
> I'm not sure what a correct system should be, but I know that it can't be based on the Earth, the sun, or any one galaxy. In order to have accurate time, it must be external to all of these things, like a meter stick is to an object.
But Einstein's theory of relativity makes that impossible -- there is no absolute time. The larger reason for debates like this is because any timekeeping convention is relative.
As to a "meter stick", that's also relative, and therefore arbitrary, for the same reason -- relativity theory.
"3:00 pm" is in the middle of the night, the time to wake up for work will always "change"..
well, one time .. but do you !really! care about how the folks in +600 yrs measure time? I'd guess they invent some kind of "glacial-period-light-saving-time" by then...
Well, we're still using the Gregorian Calendar, and according to Wikipedia, that was released in 1582 [1]. When designing a system as eternal as time, I believe that it should still be easily usable in 100, 600, or 10,000 years; on any planet, and in any galaxy.
I know it's a joke, but it's lazy programming/lazy thoughts like this that caused the issues with the systems in the first place :)
Although I still don't fully understand why leap seconds caused such a problem - we can handle leap years and we can handle daylight savings, surely this is just another clock correction action such as these?
The problem with programming for a leap second is you don't know when a leap second will occur. Compare this to a leap year, which happen every 4 years, always. Leap seconds are sporadic [1]. Programming a clock to accept the time of 11:59:60 only on occasion is trivial in a web app, but slightly more complex in a firmware or high-reliability system like flight controls.
> Ending leap seconds would make the sun start rising later and later by the clock -- a few seconds later each decade. To compensate, the U.S. has proposed adding in a "leap hour" every 500 to 600 years...
So it's not abandoning the shifting so much as putting it off into a larger interval, the leap-hour.
Today, that interval is the second. Today, we don't adjust for partial seconds; this proposal just moves the adjustment interval way out.
Perhaps somewhere in the middle like leap-minutes would be better. This would thus occur every 60-90 years.
No, Unix Time is the number of seconds since a fixed origin. In human time the origin is usually expressed in terms of UTC time, but Unix Time itself is divorced from Earth's rotation. Leap seconds only have to be taken into account when converting into human dates and times.
But whose clock do we trust? No clock on this planet is in an inertial frame of reference. Over time, clocks at different latitudes and altitudes will diverge because they are moving at different speeds, even if they have the same atomic decay "tick" process. Which do we pick as the standard?
Even if one clock were picked, observations of when the clock ticked would vary over the planet as well.
Astronomical distant fixed points are more stable over long time spans than Earthbound clocks or the Earth's rotational and orbital characteristics, and the relative displacement with regard to Earthbound distance is negligible.
("America! Fsck yeah!" I mean, some of us USans, proud sponsors of WIPO and other fine agreements, never thought we'd see the day our international representatives would be in the right. Sadly that day appears to be... 2005.)
Leap seconds are an abomination against nature: they make math not work on time. In chronological (atomic) time, math on time is stateless and referentially transparent. Like math. In sidereal time, it... isn't.
The general use of sidereal time is a gigantic global complexification with trivial cosmetic benefits to one specialized profession - astronomy. Unfortunately, it is also the astronomers who have been put in charge of time standardization, so the disaster will probably continue.
The right way to treat time as a programmer, if you're really serious about time, is to treat sidereal time as a display mode, like a timezone, and work internally in proper chronological time (eg, GPS time). Unfortunately the leap second system is a timezone that varies over time. But at least the complexity is isolated in the presentation layer.
To rebut my own argument, I suppose one could say that noon should always be noon, regardless of how long the day is. The prevalence of daylight-saving time in the US might disagree with that though.