Slashdot Mirror
U.S. Scientists Call for a Time Change
saqmaster writes "The BBC reported yesterday that U.S. scientists want to change the current system which keeps clocks in sync with solar time by adding a leap second every 18 months or so. This has rattled a few cages with the scientists and operators involved in GMT-related projects and facilities as it would effectively remove the importance of the meridian from timing. "
Interesting, and in China they only have 1 time zone for the whole country (Beijing Time)
All these inaccuracies are the direct result of primitive imperial measurements. It is high time for metric time!
How ya like dat?
I would disagree.... The problem with the current system is that numbers get passed around without the frame of reference, with the tacit assumption that it is known. So, when I say it's "5pm", it's like saying, "the temperature is 13" or "the wind speed is 49". 49 what? Using a timezone suffix (3pm PST) provides the necessary and sufficient information, but still requires extra knowledge to do the math -- what is the time difference between PST and where I'm at right now? It doesn't help that daylight time gets turned off and on, so you have to remember which "mode" you're in, but your wristwatch doesn't tell you at a glance. Using a numerical timezone delta (5pm GMT+5) provides the necessary information in a convenient manner. But, at that point, we might all as well just use GMTs.
Why dont we all take out our photon drives (laser pointer) point em westerly and fix the real problem?
Intead of saying, it 8:00 here, what time is it in Hong Kong.
You would say "We get to work at 1:pm, what time to people in hong kong go to work??"
thus still doing the same math.
Exactly, so for purposes of working out whether it's a reasonable time to call someone around the other side of the world, things would be exactly the same. No better and NO WORSE.
But for other purposes we would get major advantages. If I tell you that I plan to call you at 8:00 am tomorrow then you only have to worry about whether that's a convenient time for you, not whether I mean my time, your time or someone else's time. If I post on my web site that a major announcement will be made at 12:00 midnight tonight then everyone would know when it would be - again no need to worry about time zones. If I know the departure time for a flight and the duration of the flight then I would be able to tell the arrival time without worrying about time zones, or if I knew the departure and arrival times then I would know the planned duration.
Time zones make things more complicated with no real advantages.
To summarise the summary of the summary: people are a problem. ~ h2g2
The overlap may be the correct time to call a telecon, but you don't seem to live in the real world where whoever is in charge calls the meetings and you need to be there. It's one of the biggest problems of being a worldwide organization, or even just nationwide.
If the West Coast finds an urgent problem and needs to solve it before the end of their day, and the East Coast has to be involved, it usually doesn't matter that it will keep the East Coast workers late. West coast already knew exactly when the East Coast was working.
We have both regular and Zulu clocks around. We don't need to adopt a universal time, we already have one--UTC or Zulu. Those who need to use it do so. Those who don't, use local time. Problem solved. Besides, for those not aware, the current system keeps a common reference. If you call someone and they tell you "it's 2am!" then you know you picked a bad time to call. Under a single system, it's 5pm in both locationa--"it's 5pm!" just doesn't have the same impact across the world in that system.
Our founding fathers removed the guys in charge. Be American. Vote incumbents out.
Computers should use an internal clock running in TAI (International Atomic Time), rather than UTC. Time values should be stored and computed in TAI. TAI advances by one second every so many vibrations of that cesium atom, and that's it. No time zones, no leap seconds, no nothing. If you want to find the number of seconds between two points, you just subtract. Then when a user asks for time, the OS should 1) add/subtract the accumulated leap seconds (which gives you UTC) and then 2) add the time zone offset (which gives you "local time"). But most OS's don't do this, and most (all?) programming languages seem to barely support UTC, let alone TAI. POSIX for instance doesn't do this right, NTP doesn't do this right. NTP handles leap seconds by just re-sending the same time values for 1 second. Not important to most of us since the accumulated leap seconds add up to, what, 11 seconds since this was started a few decades ago. But if you want to think precisely about time, that's the answer. Here's a simplified way to think of it: UT = universal time. the "actual" time according to the movements of the planet (a continuous wobbly variable that can only be known by measuring it). The distance between each second is never exactly the same. TAI = atomic time. advances forward one second every second uniformly. UTC = an approximation of UT created by adding/subtracting leap seconds with TAI.