(Phys.org) —Officials at the International Earth Rotation Service have announced that a leap second will be added to the year this summer to allow for syncing up atomic clock time with the Earth's rotational time. The announcement has made many Internet sites nervous, as adding leap seconds in the past has caused problems with services such as Foursquare, Reddit, Linkedln and Yelp—all reported incidents due to the sudden time discrepancy the last time a leap second was added back in 2012.
The advent of computers and the Internet has caused a need for ever more precise time measurement and scientists have responded—atomic clocks are the standard now with accuracy up to quadrillionths of a second. Unfortunately, the spin of the Earth on its axis, which of course is the original source for recording the passage of time—one spin equals one day—is slowing down, losing approximately two thousandths of a second every day. That means that atomic clocks and true Earth time must be reconciled every so often—it has happened 25 times since it became necessary back in 1975. But back then, adding leap seconds went virtually unnoticed by all but the most interested. Nowadays, however, adding a leap second—which is scheduled to occur at 23:59:59 on June 30 (the halfway point of the year)—can cause computer systems to become confused when their clock shows 60 seconds, rather than rolling over after 59—and we all know what that generally means—outages. Other computers will show the 59th second for two seconds in a row, which can also cause problems.
Some companies, most notably, Google, have created a work-around—they call it the "smear around"—it forces servers to use extra time in making updates over the course of the year, which prevents them from ever noticing that a leap second has occurred. That approach apparently worked well enough as Google has already announced that it would use the same technique this year. Whether other sites will be doing the same is still not clear.
Because of such disruptions, some in the technology sector have called for an end to leap seconds—doing away with time based on the Earth's movement altogether. That would not mean much of a difference in the near term, but at some point, people would find their clocks completely mismatched with days and nights, perhaps necessitating the need for a leap minute, or hour which would seem to be even more disruptive. Another possibility is of course, to maintain a dual system, one for technologists, the other for everybody else.
Explore further:
'Leap second' under the gun at Geneva time talks
More information:
International Earth Rotation Service: www.iers.org/IERS/EN/Home/home_node.html

antialias_physorg
4.6 / 5 (8) Jan 07, 2015Only solution: We need to speed up the earth rotation.
Seriously: if the sites are afraid of problems then they should just take themselves offline for a minute. It's not like that would be the end of the world.
TechnoCreed
4.3 / 5 (3) Jan 07, 2015Rockguy
1.3 / 5 (3) Jan 07, 2015Uncle Ira
1 / 5 (2) Jan 07, 2015nkalanaga
4.6 / 5 (5) Jan 07, 2015Thirteenth Doctor
2 / 5 (1) Jan 07, 2015Thirteenth Doctor
2 / 5 (1) Jan 07, 2015Now that I think about it, I think that has more to do with date.
baudrunner
1 / 5 (4) Jan 07, 2015Of course, you'd have the exchange trolls going, "..but do you know how much money is lost and earned in a single second?!!"
..tell it to the hand
nkalanaga
5 / 5 (3) Jan 08, 2015Whydening Gyre
5 / 5 (6) Jan 08, 2015antialias_physorg
5 / 5 (5) Jan 08, 2015C'mon - when you develop software thinking about the particulars of the Earth's rotation and how that could play into your software every few years is not exactly at the forefront of your mind.
alfie_null
5 / 5 (4) Jan 08, 2015Those of us cursed to live in a daylight savings time zone have to deal with this sort of discontinuity twice a year.
nkalanaga
5 / 5 (2) Jan 08, 2015The biggest problem I can see listed in the article is the bit about OS's getting confused by 61 seconds in a minute, which is a timekeeping problem in the software itself. In theory, it could affect digital clocks as well, especially the fancy ones that use radio or internet time services to stay accurate. As you said, well designed software wouldn't have that problem.
nkalanaga
5 / 5 (2) Jan 09, 2015http://phys.org/n...ess.html
Rockguy
3 / 5 (2) Jan 09, 2015@nkalanaga yes you are right, good point, what I said sounded good to me at the time but I guess I didn't think it through lol
IamVal
not rated yet Jan 10, 2015this page says
"losing approximately two thousandths of a second every day."
that page says
"with days lengthening by 1.7 milliseconds per century."
those are about 5000000% different.
if we've had to do it recently, I'd believe the first..
and null has a good point in that any good system will be operating on an absolute tick system(seconds since midnight jan 1 1970 is the 'timestamp' for most computers) . the problem is that if you calculate that straight with only using leap days, Youre interpretation of the time will be off by a few seconds which can be a very big problem for GPS devices and things that need to be synchronized
the way this will work out is that they'll release a new 'base numeral'. so, if 12:59:59 jun 30 is 1500000 (it's not..) then the 'update' would say 1500002 = 1:00:00 and any time after that can be calculated with the basic system.
OZGuy
5 / 5 (1) Jan 10, 2015Y2K certification was the easy money. Scan the code for any date handling routines, check the underlying datatype and manipulation and then produce a report stating whether it was at risk or not. If deemed at risk a quote to undertake mitigation could be supplied, for a fee.
A full blown investigation of a large system including all internal and external data sources/sinks was guaranteed serious cash in the bank. You did have to know what you were doing not just pull a recommendation out of your @rse but it really wasn't that hard and paid very very well indeed.
nkalanaga
5 / 5 (1) Jan 11, 2015"Standard time" did the same thing. Noon used to be defined as the moment the Sun was due south. Since the Earth's orbit isn't round, noon to noon is seldom 24 hrs, sometimes the Sun is fast, sometimes slow. With sundials it didn't matter, with mechanical clocks it does, so we follow the clock now.
TechnoCreed
5 / 5 (1) Jan 11, 2015It just means that now the days are too long by .002 seconds and in one century it will be too long by around .0037 second. Just don't take my words for it, follow my link and read the paragraph named: Did you know... http://www.timean...und.html
bluehigh
1 / 5 (1) Jan 12, 2015