Slashdot Mirror
Extra Leap Second To Be Added To Clocks On June 30
hcs_$reboot writes: On June 30 this year, the day will last a tad longer — one second longer, to be precise — as a leap second is to be added to clocks worldwide. The time UTC will go from 23:59:59 to 23:59:60 in order to cope with Earth's rotation slowing down a bit. So, what do you intend to do during that extra second added to that day? Well, you may want to fix your systems. The last time a leap second was added, in 2012, a number of websites, Java and even Linux experienced some troubles. Leap seconds can be disruptive to precision systems used for navigation and communication. Is there a better way of dealing with the need for leap seconds?
There is a better way. Just wait several thousand years and then add one leap hour.
and speed up the rotation of the world by 1 second.
Of course, there's a better way. Just ignore the small error until it adds up to an hour, and then skip a DST transition.
No, not really. Leap seconds are a known quantity, make sure your code handles it.
Next question.
It doesn't mean much now, it's built for the future.
"Is there a better way of dealing with the need for leap seconds?"
Sure, use/write software which correctly handles time. Leap seconds with their current, well defined behavior, have been around for over 40 years.
If you have software which assumes a minute is always 60 seconds, an hour is always 60, 60 second minutes, etc., you're doing it wrong.
"National Security is the chief cause of national insecurity." - Celine's First Law
Then engineers build their systems so that they work with leap miliseconds, as otherwise they would fail more often. If your system fails only every 4 years and then only for a few seconds, you won't invest in fixing the problem. If the event occurs more often, then you're forced to.
As far as I understand, the problem is not that Earth is slowing down, but that the speed is slower than our present standard of time.
Therefore the sentence in the fine article above:
"in order to cope with Earth's rotation slowing down a bit"
is not correct.
The reason is not that Earth is slowing down, but that it already is slower, but actually quite stable (at least for now, that is).
There should be a leap second every month, followed by elimination of a second the following month. That way all code would quickly be adapted to work correctly with leap seconds.
Then of course you have to program to handle the months you don't omit a second because you really wanted the leap second to take...
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Dammit, I'll be working! I better get OT for this.
--- Math illiteracy affects 8 out of every 5 people.
There are two domains to consider.... human and computer. Humans won't notice or care about sunrise being off by one second or even much more than that. Computers need exact consistency. So the solution is, as stated, to update the clocks to the actual rotation only infrequently. Everyone, man and machine, will be happy.
"He took a duck in the face at 250 knots." -- William Gibson, Pattern Recognition
>> Is there a better way of dealing with the need for leap seconds?
We just need to make each second last a very little bit longer.
A recent draft of the set of options which will be presented when the World Radio Conference votes this fall is visible at http://acma.gov.au/Industry/Sp... This draft has options A, B, and C, but it is likely to be wordsmithed a lot before it is finished.
Let them add up until we get a full leap day, then everyone can take a day off for an official holiday! Of course, if there is as much as 1 leap second per year, it will only take 86,400 years before we can take that day off! :-)
Solution there, it seems to me, is to inhabit unchangeable-spin planets.
The problem is that you don't understand it.
I'm pretty sure NASA would have to account for all of these additions, but do we live in the same precision world as does NASA? Do command lines like "date +%s" historically count these seconds?
Is that a roll of dimes in your pocket or are you happy to see me?
At some point the earth will stop spinning.
The one who will build a forecast model will know which side of the earth will be .... "sunny way up".
Appropriate investment decisions can be made to purchase a land in the twilight zone, where is not too hot and not too cold.
Some sources say, that earth is slowing down 1.7 milliseconds every 100 years. However the last leap second adjustment took place in 2012, http://en.wikipedia.org/wiki/L...
Anybody know a quick and dirty way, a good formula, that would tell us when the earth will really stop rotating?
At last the french did something right.
You could use Atomic time, which is basically UTC but without leap seconds.
What?
Every large earthquake changes the rotation speed. The big 8.x one a few years ago, the one with the Tsunami, that one significantly changed rotation.
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
First, you've got all those new solar panels absorbing photons from the sun, but only in one direction, and then you've got all of those wind farms and tidal hydro plants adding friction to things. Of course the earth is being slowed down. Every time some do-gooder plants a tree, what do you think happens? More friction. Friction slows down the planet and causes heat. Check it for yourself - try walking briskly to the fridge and back, and then try shuffling your feet on the carpet as you do the same. Heat! Slowing down! Exactly like what happens when the atmosphere has to slide over something that's in the way, like a rainforest. No matter how many coal-bearing mountains we smooth out to help with this problem, the infestation of "tiny houses" clustered around hipster-friendly towns just slows the air back down again.
Don't disappoint your bird dog. Go to the range.
Is there a better way of dealing with the need for leap seconds?
Decrease the radius of the moon's orbit.
It must have been something you assimilated. . . .
Will my server have "time running backwards" errors tonight? If so, I'll have problems with dovecot mail tomorrow.
no, I don't have a sig
how much energy are we talking here?
How about everybody just run the same direction for a few minutes to speed Earth back up?
Table-ized A.I.
No, the Earth really is slowing down very, very gradually. The tidal forces from the moon is slowly leeching off rotational energy from the Earch (as heat). See here: http://en.wikipedia.org/wiki/T...
Note to self: get more sleep before commenting....it's losing rotational energy to pushing the moon farther away. Gah.
61s in a minute! I will definitely try to beat my best laptime on Mario Kart Wii.
...a better way of dealing with the need for leap seconds?"
Well... It could be worse. The last time we let the clocks go off we lost almost two weeks trying to fix it.
Let's just keep up with the leap seconds so that nobody has to cancel Christmas again.
The main problem is the wrong handling of UTC by NTP and Unix.
NTP is simply on the wrong time system. It is a system which is designed to accurately keep a monotonous steady time shared among millions of systems spread across time zones. It does not have any sensible way of dealing with the fact that some systems want to suddenly add a second or subtract one, just like it cannot deal with time zones or Mayan calendars. Those are simply outside its problem domain. Luckily the fix is simple: Stop handling leap seconds at all in NTP, just keep counting seconds. If you must handle it in NTP for those client systems too primitive to do it themselves, do it by transmitting the current offset between NTP time and UTC. The best solution would be to define NTP time to be TAI of course, but it will likely have to be TAI+offset to handle backwards compatibility.
Unix again does it wrong by keeping system time in UTC rather than TAI. UTC is useful for humans but difficult for machines, it should be handled by the human interface libraries, just like time zones. Kernel time should be TAI of course. When leap seconds are inserted, systems must be updated, but that is not particularly harder than keeping the time zone files up-to-date is already. Of course it would be a lot easier if the astronomers would let us know a few years in advance rather than six months, but then the offset between TAI and UTC could exceed 0.9 seconds, and as we all know that would bring Ragnarok.
GNU systems even have sets of time zones for precisely this reason: "POSIX" and "Right". Unfortunately it is not possible to use the "Right" time zones with current NTP.
Finally! A year of moderation! Ready for 2019?
Earth is slowing down, but that's not what this leap second is for, it's for the speed that the Earth has been turning already. That's why it's applied at regular intervals, but doesn't get larger. If it was applied at regular intervals, but got larger and larger, or was applied at smaller and smaller intervals, then it would be because Earth was slowing down.
If you're doing calculations on time using intervals, and one second matters to you, you should be using a raw number instead of calculating the "23:59:59" yourself. If the UTC conversion is too much, use TAI instead and be done with it:
From http://en.wikipedia.org/wiki/International_Atomic_Time:
Hire a Linux system administrator, systems engineer,
23:59:59 is normally followed by 00:00:00 (or 24:00:00). In this case it goes 23:59:59, 23:59:60, and then 00:00:00. One second gets added at the end of the day.
For the vast majority of cases, leap seconds shouldn't be a problem (or even important). Just note that the clock is wrong and use the built in OS functionality to adjust the clock. (in other words, on a well maintained server, do nothing special)
For things that need a constant time base but not a persistent one, just count the ticks and call it good.
For the really hard cases, we probably need a time base that is simply the number of seconds from a defined base. Then maintain a simple table in the public domain containing the correction to the other time bases. Just select the greatest entry that is <= the current time and apply the correction to derive the adjustable time value.
"precisely timed money transactions could go astray"
oh, won't someone PLEASE think of the frontru..., er, I mean "high frequency traders providing liquidity"!!!
not as heat, its transferring into orbital velocity of the moon, which is why the moon is slowly moving away from the Earth as its orbital velocity increases.
Making the generous assumption that this is a brain slip and not a troll, seconds don't normally go up to 60. They stay in the range of 0..59, which means there's sixty of them. Counting from 0 to 60, including the 0, gives you 61 seconds in that minute.
That's why it's applied at regular intervals
That's not true, as many have pointed out already.
Earth is slowing down, but that's not what this leap second is for, it's for the speed that the Earth has been turning already. That's why it's applied at regular intervals, but doesn't get larger. If it was applied at regular intervals, but got larger and larger, or was applied at smaller and smaller intervals, then it would be because Earth was slowing down.
It's not applied at regular intervals. FTFA: "From 1972 to 1979, at least one second was added every year. Leap seconds were added six times throughout the 1980s. But there will only have been four leap seconds added since 1999. "
It is applied based on committee decision to add them based on ongoing measurements of the earth's rotation, and therefore cannot be easily predicted.
Extra Leap Second To Be Added To Clocks On June 30
It's not an extra leap second, it's just a leap second. They're already "extra" by definition.
Yours sincerely,
Captain Pedantic
systemd is Roko's Basilisk.
Heaven forfend! I was under the impression Linux is the alpha & omega of operating systems...or I would be if I was some weak minded nerd.
Leap seconds work perfectly well for most situations. If you need precision monotonically-increasing seconds, use TAI time (or "GPS time", which is at a fixed offset from TAI). Leap seconds keep atomic clocks and the real world reasonably synchronized; any other approach will have its own problems.
- David A. Wheeler (see my Secure Programming HOWTO)
"Not only are the trains now running on time, they’re running on metric time. Remember this moment, people, eighty past two on April 47th."
Explained here.
This post right here is why we need a "Post Inaccurate" option for modding.
Test those boundary conditions! It's just another infrequent event that must be planned for.
Just like leap years, leap seconds are part of the spec. Some Februaries have 29 days just like how the last minute of December 31 and June 30 can have 61 seconds. The only difference is that there's no formula for calculating when a leap second will happen, but they are announced well in advance and the NTP protocol notifies when a leap second is upcoming.
Expand your test suites to include minutes with seconds numbered 0 to 60. If your application is time critical, make sure you're checking for that NTP leap second flag on December 31 and on June 30.
Choose your timescale wisely. Use UTC where it's important that "time" matches some random astronomic phenomenas, and use TAI where consistency matters.
bickerdyke
Computer models suggests that in 20-25 years, Climate Change could cause time to slow to up to half of what it is now making days 48 hours instead of the 24 hour day we are used to. Hold on, it gets better, while simultaneously slowing the Earth's spin causing another 8 hours to the day.
Computer models suggests that in 20-25 years, Climate Change could cause time to slow to up to half of what it is now making days 48 hours instead of the 24 hour day we are used to. Hold on, it gets better, while simultaneously slowing the Earth's spin causing another 8 hours to the day.
There are regular intervals at which leap seconds *CAN* be applied (The last day of March, June, September and December may be one second longer or shorter than other days). However, leap seconds and skip seconds are not always applied. They are irregular because the turning of the Earth is irregular.
When our name is on the back of your car, we're behind you all the way!
use the fucking metric system!!
The protocol allows for a "skip second" as well as a leap second. The clock could legitimately go 23:59:57, 23:59:58, and then 00:00:00.
When our name is on the back of your car, we're behind you all the way!
uh... it is slowing down, by an average of about a millisecond a year, due to tidal friction against the gravitational pull of the Moon.
Political debates have me rolling my eyes so much I think I got optical whiplash. I should sue. - Foamy The Squirrel
Easy way: You can pull the network cable or power off during that time. ... except a shutdown is how more than a few people to get through the DST fallback hour.
I would be joking
the problem is not that Earth is slowing down, but that the speed is slower
Try thinking that through a little more carefully.
And, as was mentioned above in a comment, tectonic forces can affect the Earth's rotational velocity. Our moon is the number one culprit, but other celestial objects also have small effects that will add up over longer periods of time.
He skipped 10 days to make the calendar line up with the equinox.
I love Jesus, except for his foreign policy.
I wonder how many programmers are going to complain that they only coded for leap seconds and never bothered allowing for the reverse. It'll happen one day.
Most of the depleting rotational energy of the Earth goes into the orbital energy of the moon, which means the moon will be boosted in its orbit and recede from Earth until the Earth is slowed down enough that the same side of the Earth faces the moon as it orbits ("tidally locked"). Which, by the way, is why the man in the moon always faces us (i.e. tidal locking happened long ago on the moon). At that point, a day and Earth an an orbit on the Moon would both last about 47 days. Estimates for when this would occur are around 50 billion years, by which time the sun will have swollen up in its death throes and consumed the Earth. There is some heat generated due to friction from tides and distortions of the Earth's crust, but the rotational-orbital exchange is much more significant.
So who is looking into this Slowing Down thing?
You know what that means, don't ya? It means longer days, more exposure to UV radiation, more cancer deaths, more drought, less food, world wide conflagrations!!!!
We need to tax people traveling East to West. They are clearly responsible for taking some momentum away.
Tall buildings! We need to get rid of tall buildings because when the wind hits them from the wrong direction, it also slows the earth down. From now on, no more buildings above 1 story.
We need a U.N. Committee to look into Anthropogenic Slowing Down!!!
ELEVEN.
(That's ridiculous. It's not even funny.)
Addendum: The last minute of the last day of any month can be lengthened or shortened by one second. However, there is a first preference of June or December, and a second preference of March or September.
When our name is on the back of your car, we're behind you all the way!
Well, when you put it like that, the whole thing becomes blindingly obvious.
Thank you :)
...a crime lasting less than one second---it legally won't exist. (Note: this is not at all true.) (It's based on a short New Zealand film I saw decades back in which a man is released because it turned out that when a relevant law were changed there were a gap between the new and old laws' domains, and his offence was in the gap. He and a mate went on to hijack a bus and charge people extra to take them exactly where they went, which proceeds they wanted to use to start a mushroom farm.)
Would 23:59:59 , 23:59:60 , 24:00:00 be possible?
To those who thinks it would be better to use TAI... not going to happen. BIPM actually considers surpressing TAI information. Time and Calender is Solar centric, not Cesium centric. On a sidenote, sometimes earth rotation speeds up which could lead to skipping a sec instead of adding one.
OpenSUSE has 61 sec minute when a leap sec is added.
Eject all Beliebers into space...lower mass = increased rotational velocity = no need for leap seconds.
June 30 is the last day of the fiscal year in Australia; wonder if there will be any issues with our banking / financial systems?
GPS doesn't directly need leap seconds, but it relies on astronomical calculations that do.
I write software that deals with leap seconds. I don't write the stuff that actually calculates them, but I have to make sure not to break it. One of the mottos of the company I work for is "it is rocket science".
A second is a small enough margin of error to be allowable. Most time systems are based on seconds, so that's the smallest reasonable choice.
For the jokers. An error of 1 minute would be 1 nautical mile (almost 2 km). Which is 1 minute of latitude (60 minutes in a degree). Funny how minute and nautical mile are the same.
I'm an astro geek, and a sailor, don't piss in my yard.
We move time forward or backwards an hour during the summer, yet worry about a second? No, its arbitrary.
Also the sun does not sit at that position at noon FOR EVEN A WHOLE TIME ZONE. We use to have local times local to each village to keep noon, yet we standardized at hourly intervals across a time zone. Again arbitrary, and the timezones are arbitrary too, islands and countries decide to be in the wrong zone for political reasons. e.g. most of europe is on CET and it has nothing to do with keeping noon as the suns high point in the sky!
This change makes no sense UTC has become a joke.
It must be stopped for the good of the children
60 seconds = 61 servings
Is this an evil conspiracy by physicists to keep all the good time to themselves while we get leeper time?
... is to install huge rockes on earth and make it spin faster!
All this bullshit about tweaking the time is just ugly hacking!
So, what do you intend to do during that extra second added to that day?
Well there's this novel I've been intending to write but haven't got around to it yet.
Row row fight the power?
"Lack of speed can be overcome. In the worst case by patience." --Znork
England told him to bugger off. Hence, on a unix system:
$ cal 9 1752
September 1752
Su Mo Tu We Th Fr Sa
1 2 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
There are a few thoughts along this line:
- Base time on a basic cosmological constant - perhaps the period of a basic elemental or physical constant.
Then, we regularly calculate the change in Earth's rotation and factor that into the daily calcs for sunrise
and sunset, and live accordingly.
- Screw ANY constant. Base 'regular' time on a percentage of actual Earth rotation, constantly updating
common cronographs (watches), dividing the day up into set portions. Perhaps mid-day is when the sun
is directly overhead (like it theoretically is now). Thus, mid-day -to- mid-day period changes at the
current rate of 1/1000 of an hour. When the rate changes, as time will gradually do (barring a
catastrophic event), lives can also gradually follow the change.
Self-importance and self-indulgence is the root of ALL evil.
We would likely need to keep 2 clocks: One based on a constant period, and another based on events related to Earth's rotation.
Self-importance and self-indulgence is the root of ALL evil.
The better solution, is to actually have programmers deal with these conditions better; and to actually learn something about the domain.
How many issues do we have with DST?
How many issues do we have the 3 years of 7 where Dec 31 is in the following year (ISO Year Week 1); e.g. twitter this year
How many issues do we have when Timezone changes happen (many times a year; some even moved from -12 to + 14, so date went Dec 30, Jan 1; with no Dec 31 that year.
There should never be an assumption that "seconds" will only be a value between 0 and 59 inclusive. It can, and has taken the value 60. This is a known quantity, and should be handled. Much like you can't assume a month is 30 days, or a year 365 (leap year [365], timezone change [364]). You can't assume a day is 86400 seconds. That assumption has been false for many years now.
And for systems that store the date as an offset (linux, being stored as seconds since epoch, rather than BCD), that just means updating libraries that convert unix-timestamp to civil-time as needed, which is much less often than the timezone database.
Some libraries will assume 60s a minute, and thus their times will always be increasingly off by a few seconds as time goes on.
Some libraries take into account the leap seconds, and thus Date Parsing to a Unix Timestamp will be off, or duplicated depending on implementation.
People need to step up and not program in that neat idealistic view.
In 2010, users in one timezone that switched from -12 to +14, only had a 364 day year. As they were missing Dec 31 that year.
For systems that don't their time-functions will be one second off from those that do. Which is currently the case that happens with non-updated system-timeone databases.
Cool! How long would it take to get there? ;-)
Self-importance and self-indulgence is the root of ALL evil.
Leap seconds were applied yearly in the 70's mostly at the same exact date, as I said. However since 2000 we've only had three leap seconds at random times like you said. This actually implies that the Earth is either speeding up, which we know is not the case, or the rate of slowing of rotation has gone down. Interesting.
Strap some big rockets on the earth around the equator and let them speed it up. Bob