3-2-1-1
-
An extra second is coming this New Years as the world's timekeepers want to insert a second to account for a slowing of the Earth. For most people, it isn't a big deal, but for some real time systems that require accuracy within a second, it could be an issue. The Japanese will actually delay their telephone clock signals by .01 sec starting a few hours earlier until the time is made up.No one knows how big an issue it is, so this is one way to document some problems on a small scale. For databases, it shouldn't be a big deal. For things like medical devices that may inject medicine, I'd like to think it won't be a big deal, but I guess 5 seconds is better than 5 minutes.
I'm hoping that the most disruption is someone like Amazon that might get a few more orders on Dec 31 because of the extra second, but we shall see. It's a little hard to get too excited about it after Y2K was a big bust. Or maybe a very, very small bust.
I still remember sitting in the penthouse of an office building in the tech center on Dec 31, 1999. I'd left a celebration with my wife, non-alcoholic of course
, to go back to work and prepare for something to happen. The lights going out, servers cutting off, or our application blue-screening everything.And nothing happened. After 15 minutes or so of checking servers and pronouncing things healthy, we headed home.
I'm guessing that same scenario will take place again this year for many people. Just glad it won't be me.
Steve Jones
http://news.com.com/Why+Johnny+cant+code/2010-1071_3-5596882.html?tag=nl
Steve Jones
-
I've reached the age when I time things with a calendar.
There is a great story about one of the old British colonies where the accuracy of the town clock was a source of great pride to the natives. In the climate clocks were notoriously unreliable.
They verified the accuracy of the clock by noting that British fort cannon fired when the clock struck 12. It wasn't until the British left that it was found that the clock was no more accurate than any other, it was just that the British used to have a look-out who watched the clock through a telescope and when it read 12 o'clock he used to signal that the cannon should be fired.
-
Well I'm not surprised that nothing happened to any date systems when you had abandoned your wife to watch the turning of the clock because you were 11 months early! On Jan 31 1999
-
In the general spirit of blowing things out of proportion, I am sure that this is going to cause a little spirit of fear in the IT world. I am sure that it isnt that big of a deal, but one that needs to be dealt with. For instance, I am sure that most BIOS manufacturers are going to need to release updates to their BIOSes to account for an extra second each year from now on. I also assume that there will need to be updates to most operating systems to account for this. What is interesting is that the slowdown of the earth's orbit is relatively predictable. Why is it then that the scientists have not laid out a projected slowdown for the next 20+ years and come up with a logical way of adjusting the clock to account for it. That way, software and hardware developers could account for that lost time in their computations, thus lowering the need for last minute and drastic changes to the software that is in each computer in the entire world. Maybe they have, I don't know. It just seems that that would be more logical.
By the way andy, good observation, I seemed to have missed steve's mistake the first time I read it also, its interesting how two digits can make 11 months difference isn't it
Regards,
Aleksei
-
Like so many in IT I heard the uninformed moan about how nothing happend and Y2K was just a big bunch of crap. My answer was, "I guess if we had known the Japs were going to bomb Perl Harbor for 5 years before it happened we would have been prepared as well." Of course we could also add 9/11 to that thing about being prepared for the unexpected, but that's what makes a surprise, it's unexpected!!!
Thanks for the interesting trivia...
Bill
-
"Y2K was a big bust" because businesses spent somewhere around 300 billion pre-Y2K to mitigate the issue.
-
Actually little needs to be done for most computer bios, since they are often a good part of a second (or more) off. If they do any net sync, they will just be reset at the next sync. Network timing devices are generally set to handle the leap second correctly which happens at 23:23:60 UTC (actually more accurately UT1).
There are several misunderstandings about leap second. It is not constant, indeed it is decided on a year by year basis (usually for June, but not always). It is not predominantly the slowing of the earth (which is quite gradual) but the changes in the earth's rotation influenced by shifts in the angular momentum due the moon, tide and other astronomical effects. Not all these effects are known or computable.
The current SI atomic second is probably very slightly shorter than the long term solar mean, which is somewhat due to historic choices (I understand it was originally callibrated to the ephemeral rather than solar second based on the last century or so of astronomical observations, but pragmatically probably not a bad idea.
In any case if TAI (the atomic standard) eventually becomes the overall standard, leap seconds will go away and no further attempts will be made to bring solar mean time into line with atomic time.
[fixed typo]
...
-- FORTRAN manual for Xerox Computers --
-
Uhh, yeah. What Jay just said.
I have to agree with JAlexander. While I don't know what building managers did with their elevators, I do know I spent a LOT of months searching and fixing Y2K issues--and they WOULD have been issues. We prioritized systems, and even use cases within those systems to target our work.
I always feel slighted when people say there was nothing to worry about. Although I never bought in to the world-wide panic and power grid shutdowns, I think most significant systems worked fine because of the hard, tedious work of folks like us.
- Jeff -
-
Uh, yeah .. what Jay said.
In an effort to keep things "real" for a moment, let's get a couple things straight.
1. Y2K wasn't a "big bust", it was a monumental one.
2. Leap second is more easily dealt with by adjusting your own clock than by trying to figure out every possible galactic permutation causing a shift in our axis of rotation. Gimme a break.
Just for a little perspective, realize this; there are sundials centuries old more accurate than our atomic clocks are today. In our never-ending quest for accuracy and correctness, we often ignore the very obvious. Sometimes, when you want to know the answer, just open your eyes and stop listening to all the hock.
Just my 2 pence
-
I too spent countless hours fixated on possible problems due to the mysterious Y2K gremlins. Don't feel slighted, feel proud that you were part of it and learned.
Those that weren't there for the silliness will never understand the great many things that were gleaned from that experience.
-
Just for a little perspective, realize this; there are sundials centuries old more accurate than our atomic clocks are today. In our never-ending quest for accuracy and correctness, we often ignore the very obvious. Sometimes, when you want to know the answer, just open your eyes and stop listening to all the hock.
Huh??? Sundials by their nature vary with the equation of time, with precession, with the earth's fluctuating rotation, with the vagueness of the ability to precisely locate the sun, with the refraction in the atmosphere, even with the shift in ground level that they are installed on (in many places the ground is still rebounding from the last ice age, as well as continental drift). A statement like that is totally wrong.
...
-- FORTRAN manual for Xerox Computers --
-
Thanks, Alex, I think.
Personally, I didn't find anything mysterious. I simply found code whose date calculations failed after 1999 and I modified that code. I did get a lot of chuckles from the experience--mostly from from running tests and actually observing how the software behaved prior to modification.
- Jeff -
-
Is this the beginning of the media hype to create the hysteria over the coming calamity over the addition of 1 second ??? Well if it is, it's only a year or two too late. This is as much an event as a SQL Agent 'bug' in v6.0 that computed 'December 32nd' as the next run date for a daily scheduled task.
RegardsRudy KomacsarSenior Database Administrator"Ave Caesar! - Morituri te salutamus."
Viewing 13 posts - 1 through 12 (of 12 total)
You must be logged in to reply to this topic. Login to reply