UN set to dump GMT for tech-friendly Atomic Time Greenwich could lose its place at the centre of global time if a move to "atomic time" is voted in by the International Telecommunication Union in Geneva in January 2012. Time scientists are discussing the implications of the change today in pre-vote meeting organised by the Royal Society. British newspapers are mulling how …
"One positive outcome for Brits distressed by the loss of our Victorian superpower status is that the end of GMT as an international standard could accelerate the move to keep British Summer Time into the winter, letting us have lighter evenings."
Those of us living in the North would like to register our opinion that this would be a Bad Thing. We don't want to wait until 10 am for dawn to break!
"We" being a selection of those living in "The North"
There are plenty of "we" who would like to try lighter evenings. I hate the current scenario where its dark when I arrive at work and dark when I leave. Leaving in daylight all year round would be quite nice. And summer nights in the beer garden until midnight also appeal :)
Besides. Time is an illusion. Lunchtime doubly so.
The other thing I wonder about is whether they've thought of the effect on outdoor trades... builders and the like work an 8 to 4 day rather than 9 to 5 simply because its daft not to have noon in the middle of the working day. So all the brickies and chippies (not to mention shpeherds and so on) will have to join the rest of you working 9-5 and join the main rush hour...
We experimented with this in the late sixties and a lot of people made money selling reflective gear for schoolkids who had to go to school in the dark. Have they found the abandonned warehouse with the the leftover stock?? Has Del Boy bought it on the cheap?
The public must be told :)
(On a serious note: there was an increase in road casualties in the mornings....)
Surely it would really make more sense (sun overhead at midday) if we were to standardise on GMT.
It's only those who want to regularise our time with Europe who really want to go to permanent BST.
It constantly annoys me when people keep saying that the clocks change for Winter to make the mornings lighter.
The clocks change in Summer, for an entirely out-of-date and arbitrary reason.
If a leap second occurring on a regular basis causes problems, then a leap minutes every century, when people will have forgotten the last time it occurred and it catches everyone by surprise, is going to cause chaos.
If you want lighter evenings, move further south, or to europe. I've very much enjoyed waking up an hour later this week than I did last week and wish we could move the clocks backwards in the Spring too, but we can't that's would be madness. It's going to get dark in the evenings eventually in the winter anyway, no mater what time zone we're using. If we stick with BST year round, then people will start wanting a BST++ in the summer, and a BST+ in the winter and before long, we're on the same timezone as Japan. Deal with it, we live in this part of the world, this is how the daylight hours work. Feck off elsewhere if it bugs you that much.
And these things are so unexpected, aren't they.
If the systems can't cope with a near-annual leap second, the occasional 61-second minute, why should they do any better with leap minutes?
This machine has to have the clock reset once in a while, re-synced to a time server somewhere. I've not noticed anything breaking when I do that. Most of us, I suspect, don't care. If we're in a situation where a continuous sync does matter, not coping with leap seconds is just plain stupidity. If it happens almost every year, anything that might break will have broken, and then been fixed.
Oh, I can see some advantages in going over to something like counting seconds from some arbitrary start point, when you're managing atomic clocks and comparing them to calculate a coordinated time standard, as we do. And if you want that, don't forget to specify what happens when the account rolls over.
But in the end this is about a nationalistic gutter press running a scare story. Greenwich isn't going to stop being the Prime Meridian, and when did they ever get a tech story right anyway? On past form, they'll blame the ITU on "Europe", that amorphous bogey-man of the ill-informed nationalist.
I'm not sure this would actually affect GMT's status in the world, or in fact UTC which is what's used as the reference time for actually doing stuff. E.g. issuing NOTAMs and Weather Forecasts for aviation safety which are all issued in UTC to avoid any confusion when planning long distance flights etc. Atomic Time sounds very useful for people whose work is affected by being fractions of a second out but for the rest of us it's a degree of precision that's irrelevant so I imagine UTC will continue to be used by anyone not in charge of a constellation of GPS satellites or trying to measure the speed of some neutrinos.
Incidentally, I've never understood why some people think staying in BST would solve anything, if everyone's so wound up about having lighter evenings just arrange the working day to start an hour earlier.
"and GPS systems in particular suffer when leap seconds strike"
No they do not! GPS maintains it own 'atomic time' and has a UTC-GPS offset that is steeped as required. It was designed by people who knew what they are doing. Same goes for most astronomical systems that rely of TAI, etc, etc.
This is, or should be, a non-issue as there are plenty of solutions out there. What we have is really a debate about trying to fix broken software by changing how things have been done for years, rather than getting those behind such systems to fix them.
If you NEED to worry about leap seconds, get you own system fixed!
The astronomers have had to deal with the various world timescales for years. The International Astronomical Union has a library for the Standards of Fundamental Astronomy, and there's some good routines in there to convert between utc, tai, ut1, etc. taking proper account of known leap seconds. Only trouble is that it needs updating every time there's a new leap second.
It really would be trivial to adopt that library, add it to ntp or whatever and have every networked computer system automatically updated with a current leap second table as and when necessary.
Putting off leap seconds and having leap minutes instead just makes the problem bigger, though rarer. How does that make things better?
Please refrain from coding anything ever.
A computer is a machine, a machine is absolute, why in the world would you want to feed it with imperfect time subject to random external variations causing more and more issues as processing power grows ?
We have lived with it and even built a fail system around a fail idea : that of feeding machines human time.
There is (almost only) broken software, and it's broken in many ways... but actually designing a software to be based on human time, that's a major WTF.
"A computer is a machine, a machine is absolute, why in the world would you want to feed it with imperfect time subject to random external variations causing more and more issues as processing power grows ?"
Because:
(1) Machines exist to serve *us*, we should not have to change to suit them.
(2) It is already a solved problem, just most programmers have chosen not to fix it, or are ignorant of the issue.
I'm not saying all of current systems are ideal, for example ntp & tz data should be updated so your normally patched OS have all of those troublesome adjustments delivered automatically (and maybe optional dynamic TZ selection for moving installations such as a ship etc), but really it is not such an impossible task to do.
Machines exist to serve us, but they speak only binary.
In order for them to serve us properly (if at all), we have to
1. tell them in their language what we want
2. profit!
It is a problem but it's not solved, many use workarounds or disregard the minor corrections as they are irrelevant to their scope.
In programming, a workaround is another word for more fail code to handle fail.
Human-readable time is a perfectly valid request, and it can be granted to you with MORE ease using absolute time for the machine.
You have to realize this kind of change will change absolutely nothing from a user perspective, while enabling A_LOT_MORE simplicity in anything time-related for any program.
There's no way people could ever agree to insert a leap minute. No systems would be set up to cope with it, so it would be like volunteering for Y2K all over again. As for a leap hour ... forget about it. So in practice we'd be committing ourselves to allowing time to drift out of sync from where the sun is in the sky. So we'd either have to get used to the sun rising at 18:00, say, or we'd have to adjust our time zones to compensate, so Europe would have a 12 hours difference from UTC, say.
UTC and GPS time have diverged since 1980. Currently GPS time is 15 seconds ahead of UTC. TAI is always 19 seconds ahead of GPS time.
It's not clear to me whether the proposal is to switch to TAI, to switch to GPS time, or to stop inserting leap seconds into UTC. I would guess it's the last one.
Given that we have less than a couple of centuries' worth of experience of running any sort of coordinated time, the notion that we should fret unduly about a leap hour, 550 years from now is fanciful, to say the least.
Imagining that standards presently written by the ITU will lead, several millennia from now, to a UK sunrise at 18.00 is like speculating that the ancient Babylonians should have worried about the impact of their timekeeping standards upon the design of today's IT systems.
Oh, wait, hang on. Mine's the one with a sundial in the pocket.
No .. just no.
Today, any decent software uses UTC, not YOUR friggin user time.
Today, you actually use your own personal human time (your tz).
Today, any decent software uses REAL time which it presents to you in HUMAN time.
None of those changes are ever going to change anything for you.
unix epoch = time -> totally worthless to a human
your formatted timestamp in your timezone = human-readable time -> totally worthless to a machine.
The very idea of originally mixing both is a WTF.
So you are worried about coping with leap-minutes. Yet every 4 years, except every 100 years, but including every 400 years we have a leap-year. I don't see that blowing up every time. Most places in the world change the clocks by an hour every year. We have leap-seconds thrown in on an almost annual basis.
Personally, I prefer leap-seconds to leap-minutes, but both are as trivial as anything to cope with. The way you normally cope with time slippage on computers is simple. You routinely check your time against a known reference (time servers or something getting time directly from the atomic clocks). If you are behind, move time up to match, if you are ahead, stall time until you are on time. For most people and most applications you would never notice the difference.
A few hard-realtime systems might have issues that might take a little working around. But these are issues that people deal with already for leap-seconds. There are some systems that need reasonably accurate time but aren't hard-realtime, for example cloud data sources that are updated from multiple locations.
I remember a particular example that was a lot of fun to deal with back in the late 90s. Our network techs botched an installation of a radio sourced atomic clock to our production Netware 4.11 servers. Basically they put the clock in a place where it couldn't receive the radio signal and it didn't fail good. One of our Netware servers ended up deciding it was 30 years in the future. This is a big issue for Netware because writes into the cloud based directory structure (the NDS) are timestamped to resolve contention between multiple writers - the most recent time of an update is taken. When you get a future time written in the NDS, the system copes using a mechanism called synthetic time. Basically Netware stalls time on that particular partition until realtime has caught up with the future time stamp. Unfortunately, when your timestamp is far in the future (like 30 years) this takes too long. So the next strategy is a process called declaring a new epoch. You essentially re-timestamp everything in the partition to a new start epoch time which is basically now. We tried this and it failed. In the end, I remember having to escalate the issue through Novell all the way to the back-end techies, who remote logged in to our network and fixed it with some unpublished switches on admin commands.
Generally though, especially with the design of modern OSs, time has become a much easier topic. The only people who really care about the difference between UT1 and UTC are people working on hard realtime systems with a timing need. GPS is the obvious example. But the GPS people already have a robust way of handling all this.
Overall, it is a bit of a meh story, being whipped up into a frenzy by the British newspapers because of a slow news day.
"One positive outcome for Brits distressed by the loss of our Victorian superpower status is that the end of GMT as an international standard could accelerate the move to keep British Summer Time into the winter, letting us have lighter evenings."
So ... you want to be UTC+1 at winters....
How about going CET/CEST and ditching GMT once and for all?
They are, that's where UTC comes from. But there's a catch - that should be TUC but they came up with UTC as a nice Franglais fudge.
Ditch clocks "springing" forward in the spring and "falling" back in the autumn. Only a proper Victorian bureaucrat could have come up with such well-meant bollocks.
is a waste of effort anyway, since it takes so long for people to adjust physically - twice a year.
As for the experiment in Double Summer Time, I remember it well: it gave a huge boost to the manufacturers of early forms of reflector, who were the only people to gain anything out of it. I went to school in the dark. No thanks!
How much will it cost to have every system in the world updated with the new timezone data? Will Microsoft supply the millions of people with Windows XP? And the Unix Timezone data is being rehosted - so who's going to pay for it all?
If we're going to change time, can't we change it to something more logical?
timezone change data is currently distributed by Microsoft and the UNIX/Linux distributors as updates, it isn't a big problem. That said, I am not sure what other changes will be needed, it MAY require some timekeeping software updates, they would also be distributed as updates by Microsoft and the UNIX/Linux distributors.
When will people accept that change for the sake of change always has a net negative effect. Once you have done it the first time you have set the precedent to change again, and again ad nauseum and change costs both time and money.
They dropped double summertime as it had little value but if you had to go to a single time standard through the year why not adopt GMT. At least the sun would be in the right place at mid-day.
Next you will get people pushing the idea that we could use the same time as France/Germany to help businesses work better.....get real people, there are at least three countries that include multiple timezones and they take advantage of that, they don't whinge about it and if you want to be accurate then France/Spain/Portugal should really be on the GMT too.
Here's a radical idea, those that need really hyper-critical time standards use the funky 'only loses one shake per decade' systems. The real world can stick with simply splitting a solar-based day into 24 hours with 60 minutes each.
Why is it only engineers seem able to understand the concept of using the most appropriate measurement system even if that means mixing them?
I am led to believe that last time we experimented with GMT +1 during the winter, by those that were there, on the grounds that lighter evenings were safer for the kids coming home from school it failed.
They found that all it did was shift the accidents to the mornings and had an increase to boot as it meant that kids were now on the road during the morning rush hour in poorer light levels than travelling home, not in rush hour, in mid afternoon.
I wish someone would explain to me how moving the clocks forward another hour would improve tourism. I mean who is saying "oh, I'd go to Britain in the winter if only the sun didn't go down so early" You'd have thought the weather would be more off putting.
And if doing so would save us x million annually on heating/lighting why isn't the rest of the world campaigning for it too? Surely we could solve global warming and world poverty if we all put our clocks forward 2 hours.
Anything that would bring an end to our days as a "Victorian superpower" would be most welcome.
Hopefully we can then move on psychologically.
Time is merely a reference point, so it doesn't matter if we start using atomic time.
In the summer I will arrive at work at the ninth hour after midnight and in the winter the tenth.
Although time is a reference point, it is currently synchronised with earths rotation (hence GMT). But if atomic time takes over, there's no synchronisation with any point on earth - i.e. the location on earth for each rotation will change slightly and accumulate a significant shift over the years.
I can only understand this working if we accept a shift in the atomic reference point every few years to re-calibrate it to a reference point on earth (e.g. Greenwich Observatory).
BTW, don't knock the Victorians, longitude at sea was solved by the admiralty well before Queen Vic came to the throne.
the end of GMT as an international standard could accelerate the move to keep British Summer Time into the winter,
Or better having decoupled nerd time from real time we can have GMT or even true solar time all year. If people want to get up earlier or latter according to season that is their choice, we don't need no government decreed time to set our watches
The whole idea of time zones is a nonsensical anachronism <- see what I did there?
Farmers don't give a shit about time, they start work when it's light enough and stop when it isn't, so why can't the rest of us in our cosy, illuminated offices do at least as well?
Tourists don't care about daylight hours either, they juggle their days to be where they want to be in the daylight (or it's absence) conditions they want.
I think it's because the Earth is slowing down, so the rate at which leap seconds should be inserted is increasing. It's roughly quadratic, in other words. The rate is not known very accurately and doesn't match theoretical predictions very well, so 550 is presumably a very rough estimate.
Seriously ??? We've been using computers massively for over 40 years and it's only NOW that they think they should use real time instead of bullshit time ????
omg . human race . you fail so hard it's unbelievable.
woot .. we actually built absolute clocks to perfectly determine a non-absolute user-perceived value ... GG -
to all the Froggies. Still haven't got over that one I see.
Exterminate, Exterminate, Exterminate all symbols of national national identity.
Seriously, the radioactive decay of a material doesn't correspond to when we see the sun coming up over the horizon. We're obviously we're seeing the wrong sort of sun, then!
I doubt the spineless one would use the veto.
We have timezones, we don't need a universal system. Why not set up a timezone for your newfangled atomic stuff and let anyone who wants to use it? So what if 8am shows up just around tea-time, its sooo much more accurate, cos its "atomic" see?
Bah Humbug. Go and solve the problem of war or something useful like that.
Quote:
"One positive outcome for Brits distressed by the loss of our Victorian superpower status is that the end of GMT as an international standard could accelerate the move to keep British Summer Time into the winter, letting us have lighter evenings."
Have you ever lived on the West coast of Scotland?
Didn't think so!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
