Slashdot Mirror
The Future of Leap Seconds
@10u8 writes "Since 1972 precision clocks around the world have ticked using atomic seconds, but earth rotation is slowing down. Leap seconds have been inserted in order to keep noon happening at noon, but they upset some timekeepers. Recent discussions have considered
discontinuing leap seconds in UTC, and a colloquium in Torino next month will present results. It is a matter of international significance."
I can't see why they hate leap second. I'll be damned if I am going to eat lunch at what is called 8:00 in the morning because they don't want to keep leap second. Grow up, we have leap years and human time keeping is not an exact science as the Earth tends to spin the way IT wants not the way we want.
Visit www.seriouslythough.com
first of all, I think it's important to keep on track with time, it's not like we don't have the technology to keep it up. Isn't it amazing that we can even develop the concepts in the first place? Leap years have been incorporated for awhile now, it keeps the seasons from drifting to some "other" part of the calendar. (Winter in July anyone?) Daylight savings wasn't invented to annoy people or make people appreciate the season by forcing you to be awake earlier. It saves energy by having people awake during the daylight hours. This means you're more likely to open a window than cut on a light, and go to bed while it's dark out. While leap seconds are comparatively minute, it's just maintence. (Y2k is an example of what happens when we don't think far enough ahead). I think modern-day timekeeping is the result of centuries of work. It started with us observing the sun, then the stars, and now the earth itself. Needless to say, timekeeping ought to be an exact science. Until we find something more reliable of deserving to serve as a time reference, we ought to keep our ears to the ground. We do happen to live here, and I think the Earth deserves to set the pace.
Why can't all fpga/microcontroller manufacturers just release free optimizing compilers???
Leap seconds are evil. As someone who has spent way too many hours programming high precision time distribution systems to deal with leap seconds, I'd say 'good riddance, don't let the door hit you on the way out. Sites that have to deal with them typically shut down near leap seconds to avoid any glitches. The amount of time wasted on this problem boggles the mind.
I hate them and will not morn their passing.
Leap years work like this:
;-)
One year = the time it takes for the Earth to revolve around the Sun.
One day = the time it takes for the Earth to rotate on its axis.
The problem is, there are really about 365-1/4 days in a year - it doesn't work out evenly to 365 days. So, every four years we add an extra day (Feb. 29), and then it all averages out. Otherwise, if we only had 365 days in a year, over many years seasons would start getting earlier and earlier on the calendar.
One day = the time it takes for the Earth to rotate on its axis
One second = the time it takes for Cesium 133 to oscilate about 9.19 billion times (because it's something constant we can measure)
The problem, again, is that there aren't exactly 86400* seconds in a day. So, we add leap seconds periodically to account for it. As I understand it, this isn't necessarily done at fixed intervals, but rather whenever it's decided that it needs to be done. The Network Time Protocol used to synchronize clocks over the Internet supports leap seconds; they can be announced over NTP in advance, so everybody adds them at the correct moment.
Why is it important? It's not important to most people, but computers like things to be precise and accurate for various reasons, and that means we have to agree on exactly what time it is.
* BIND now lets you write "1d" in a zone file, but how many of you still have this number memorized?
$x='S24;r)>63/* h@<5+oZ)32"5cz';$me='phroggy'x$];
$x=~y+ -xz+\0-Tx+;print$_^chop$me for split'',$x;
seriously, i did! during one of my scientific experiments (I believe it was in Jun-93), they added leap second in the middle of my experiment. The data taken from various places could not be combined together, since they didn't know at what time, leap second was adjusted at which place. So we had a 24 hours experiment on 300 million dollar equipment failed and 100's of manhours were lost in the process.
There used to be a time that a second was something that would fit 24 x 60 x 60 times in one day -- no matter how long the day was. Nowadays a second is something like this-and-that many vibrations of some atomic particle thingy.
:-)
So maybe we should just stretch the number of vibrations of the particle thingy a little, instead of adding extra seconds to days
"We can confirm that Debian does *not* ship the version with the trojan horse. Our version predates it." [CA-2002-28]
Here's an idea, why not fix it on those wierd years, without leap years. For example, 2100 is not a leap year, even though it is divisible by 4 (because it is divisible by 400 and 100). Since many computer programs won't handle that correctly, on those days, adjust for the missing seconds (a few minute change).
Kill two birds with one stone, so to speak.
-Sean
Actually, it's been done. In Kim Stanley Robinson's Red Mars, a Martian colony did adopt a clock customized for the local conditions.
The Martian day is twenty-four hours, forty minutes long, roughly. Mars kept a twenty-four hour clock, with hours, minutes, and seconds remaining the same length. The colony then added a forty minute period (the 'timeslip', if I remember correctly) after midnight. During this period the clocks (all digital) would stop for forty minutes at 24:00, then resume counting at 0:00 the follwing day.
Though neat for dramatic purposes, I would think it more useful to simply run the clocks for a short twenty-fifth hour, forty minutes long. Days could be counted--forget months--for a total of 669 Mars days per year.
The single most useful thing about such a technique is that it preserves the length of the second. Since any human presence on Mars would likely be a scientific outpost for many years, maintaining the second is very important for many measurements. I don't want to have to deal with a kludgy factor of 1.03 in comparing times.
~Idarubicin
Posters seem to broadly agree that we use time in 2 ways: 1, as humans, to "tell the time of day", ie relating the position of the sun in the sky to a time of day, which should not change, and 2, to measure a time difference between 2 events, which may have nothing to do with the sun or times of day. Clearly, for the first application, leap seconds etc. are important to eliminate drift. However for the second use of clocks leap seconds don't matter. So why not use Julian dates (number of metric seconds since some datum) for stuff like GPS, atomic clocks etc. and keep normal times with all the quirky leap seconds for "human-interface" use? After all, we don't find out the time of day from a GPS so why does it need to use minutes, hours or days? Julian timekeeping is already in widespread use in astronomy for "interval between 2 events" uses so I don't see why it couldn't be used in other such applications. We could even have a nice standards-based libtime to convert between the two.