J Thomas on October 11, 2010, 06:44:36 pm
Time zones are not merely of convenience to people on railroads; anybody who drives, bicycles, walks, or rides a horse can easily cross a time zone, and thereby need to set one's watch ahead or back one hour. This is more convenient than changing one's watch every time one enters another town.

I claim that the inconvenience caused by time zones is usually fairly small for most individuals. It isn't of gigantic importance to give up this rather bad system and replace it with a good one. We have much more important reforms to make.

I am arguing with you about this as a sort of chitchat, not something that's important.

Quote
Imagine that you (who live in West Town ) and your friend ( in East Town ) agree to meet in Center Town for lunch at noon. Before railroads standardized on time zones, you might need to ask "which noon?" - since you, your friend, and the restaurant would have slightly different times. It was common practice for each town to set their clocks to local time, according to the sun.

If you want to get that just right, you can do it easier with GMT. If you want to use local time and you care about small differences east-west, then whether or not you have hour-long time-zones you should synchronize watches. If you don't bother to synchronize watches then you probably don't care about the time difference between East Town and West Town.

Quote
Back to the issue of local time - bad as it is to tweak DST files, if we did not have timezones, then each computer would have to know its local position, and would have slightly different time than fellow computers a few miles away, or some convention would have to be adopted to standardize time.

Why not do that with GMT? That's almost exactly what computers use now, except they replaced GMT by an atomic clock to get the same result without the occasional variation caused by wobbling planets etc. You can get your computer to display whatever time zone you prefer, of course.

Quote
There are protocols - such as NFS - which require computers to be in very close agreement as to what time it is. It is also a considerable advantage when correlating log events across multiple computers to have such agreement; that's why system administrators install, configure, and maintain NTP ( the Network Time Protocol.)

http://en.wikipedia.org/wiki/Coordinated_Universal_Time

The computer guys did it better than the railroad guys. Understandably, since they did it over a hundred years later.

terry_freeman on October 11, 2010, 09:22:44 pm
Neither of us would want to synchronize watches with each other in the East-Center-West scenario; each of us would synchronize with local noon at the local clock tower, in order to be consistent with our neighbors. The restaurant, located somewhere betwixt East and West Town, would have a third local noon, which would be betwixt your watch and mine.

This was the norm, before the railroads voluntarily adopted time zones; every single town had its own clock tower and its own local time.

For reasons which may seem strange to superbly rational beings such as Dr. Bones, people tend to prefer their local time to GMT.

Mainland China, on the other hand, spans several "time zones", all of which are on Beijing time.

quadibloc on October 11, 2010, 10:02:13 pm
I claim that the inconvenience caused by time zones is usually fairly small for most individuals. It isn't of gigantic importance to give up this rather bad system and replace it with a good one.
Now you've piqued my interest.

I agreed with your use of the word 'foist' because it did seem reasonable to me that, at the time Standard Time was introduced, the trouble of switching over would not have been felt to be necessary by most people.

However, I don't agree that it's a bad system. I think it's a very good system. It strikes a balance between having local time correspond to the Sun and having the convenience of everyone using the same time standard. This way, everyone uses a time corresponding roughly to their local time, and yet they have immediate access to GMT/UT from that local time.

Now, in some cases, governments have drawn time zone boundaries in a way that has inconvenienced people - like running all of China on Beijing time, the Soviet Union being on DST the year around, and numerous small-scale choices of in which time zone a particular place should be located. So people get up an hour early in Paris because the English Channel doesn't lie between France and Germany. But that is due to misuse of the system, not the system itself.

J Thomas on October 12, 2010, 07:46:58 am
I claim that the inconvenience caused by time zones is usually fairly small for most individuals. It isn't of gigantic importance to give up this rather bad system and replace it with a good one.
Now you've piqued my interest.

I agreed with your use of the word 'foist' because it did seem reasonable to me that, at the time Standard Time was introduced, the trouble of switching over would not have been felt to be necessary by most people.

However, I don't agree that it's a bad system. I think it's a very good system. It strikes a balance between having local time correspond to the Sun and having the convenience of everyone using the same time standard. This way, everyone uses a time corresponding roughly to their local time, and yet they have immediate access to GMT/UT from that local time.

Now, in some cases, governments have drawn time zone boundaries in a way that has inconvenienced people - like running all of China on Beijing time, the Soviet Union being on DST the year around, and numerous small-scale choices of in which time zone a particular place should be located. So people get up an hour early in Paris because the English Channel doesn't lie between France and Germany. But that is due to misuse of the system, not the system itself.

I guess I might agree it's a middling-average system. Good enough for government work.

Time zones take the inconvenience of varying recorded times, and concentrate them at the irregular boundaries of the time zones. In return for the inconvenience we get time zones which do nothing right. They don't give you local times which fit the local day -- they can be up to 3 hours or so off, and if governments did it right they could still be up to one hour off, or a half hour in either direction. And they don't fit GMT. They are a hybrid which is not ideal for either purpose.

If we had used GMT for anything where you care about the time, and local time when you care about the day, probably we would have had watches with a dial around the edge. You set  your watch to GMT and then the dial shows how much to adjust for local time. The guys who want fourteen functions on their watches would get something that shows daybreak and sunset changing with the date.

It isn't a big deal. When the time zone boundary runs along a state line it doesn't bother people much at all. Even if you live near the state line, of course you have to expect things will be different on the other side. Similarly, it used to be that sometimes people living a few miles from each other would be in different area codes and so they'd pay long distance rates. The phone companies eventually fixed that one.

quadibloc on October 12, 2010, 11:38:39 am
If we had used GMT for anything where you care about the time, and local time when you care about the day, probably we would have had watches with a dial around the edge. You set  your watch to GMT and then the dial shows how much to adjust for local time.
Now at least it is clear that this is the system that you would find preferable to Standard Time.

My first objection would be that watches have two hands. There are watches with a ring around the outside that you can turn - generally as a way of using the second hand to time things without the watch being a full-blown chronograph. With Standard Time, you could adjust a ring marked with the numbers 1 through 12 to indicate GMT. With your system - assuming local time was still rounded off to the nearest minute, instead of being exact to odd fractions of a second - you could use a ring with 60 divisions to convert from local time to Standard Time, but not from local time to GMT.

Now, Ben Franklin designed a clock with a spiral design on the front that had only one hand - the face of the clock showed four hours and the minutes they were made of. Maybe you're thinking of something ingenious like that, and I'm being too narrow in thinking only of the present-day clock design.

Or, since chronographs divide the seconds into five parts, for 300 divisions, what about a watch face where the scale of 60 minutes is within the face, not on its edge - the minute hand is short, the hour hand is long - and around the edge, one has a scale of 48 quarter-hours, further divided by small tick marks so that the hour hand moves against a scale marked in increments of five minutes? And a "semi-standard time" in which all local times are an integer number of five-minute periods away from GMT would certainly be almost as good as real local time even for farmers who have to tend their cows... then a single ridge around the edge would convert from that kind of local time to GMT.

J Thomas on October 12, 2010, 05:03:10 pm
If we had used GMT for anything where you care about the time, and local time when you care about the day, probably we would have had watches with a dial around the edge. You set  your watch to GMT and then the dial shows how much to adjust for local time.
Now at least it is clear that this is the system that you would find preferable to Standard Time.

My first objection would be that watches have two hands. There are watches with a ring around the outside that you can turn - generally as a way of using the second hand to time things without the watch being a full-blown chronograph. With Standard Time, you could adjust a ring marked with the numbers 1 through 12 to indicate GMT. With your system - assuming local time was still rounded off to the nearest minute, instead of being exact to odd fractions of a second - you could use a ring with 60 divisions to convert from local time to Standard Time, but not from local time to GMT.

That sounds complicated. My thought is that if everybody uses GMT then everybody knows what time it is. Just like everybody in China knows what time it is.

What you want local time for, is to track sunrise, sunset, noon, and things like that. And how often do you need that precise to better than 15 minutes? A dial on the edge of your watch is fine for that. If you do need it precise, then you probably need to know something like what time GMT is sunset where you are today. Write that down and you're set.

Of course, if you usually track local time only to about 15 minutes then that isn't a whole lot better than time zones tracking local time to only about half an hour or an hour or two hours. But it's simpler and there's no real loss. The loss would have come to the people in 1870 or so who'd have had to get used to GMT, in place of a standard time which was just like local time except for being off. So we got an easy transition to something which had occasional glaring deficiencies, instead of a harder transition to something simpler and easier to use. Not really important, except maybe the example might be useful.

terry_freeman on October 12, 2010, 05:16:55 pm
Back to the computers: customers prefer to see their local time, not GMT. They prefer the "time of day" to match what the local government has dictated for "Daylight Savings Time."

Under the hood, my computers use UCT. (this is an option - some people use the local time instead )

However, this UCT is adjusted by the local timezone file, which is not merely an adjustment of so many hours plus or minus; it has to understand the rules for "Daylight Savings Time" - and those rules are periodically changed by politicians. 

Whenever politicians change the rules, the timezone files must be updated on n machines, each of which is slightly different. In my last gig, I administered about 150 servers, and there were about ten slightly different methods or locations for administering the timezone/daylight savings time rules. Every single machine, from routers to firewalls to servers to terminal servers to filers to desktops and laptops, had to be updated.

I once administered thousands of servers, with even more varieties of operating systems of varying vintage. Now you understand one reason for my abiding loathing for political meddlers. ;)



quadibloc on October 13, 2010, 08:28:19 am
My thought is that if everybody uses GMT then everybody knows what time it is. Just like everybody in China knows what time it is.

What you want local time for, is to track sunrise, sunset, noon, and things like that.
Ah. This is something which would simply not occur to me. To my mind, 12 noon means noon, and so on, and, thus, knowing GMT doesn't tell you what time it is unless you happen to live in Britain. Standard Time - if we get rid of certain wrong boundaries, and Daylight Savings Time - provides an adequate approximation to local time, and a sufficient connection to everyone else's time to vastly reduce the difficulties associated with dealings with people in distant places.

The idea that people could actually use GMT in their daily life, for making appointments, turning on the TV, and so on... did not occur to me. That could work on a space station, or on an inhospitable planet where people live by artificial light in underground caves.

Thus, because I took it absolutely for granted that people must live by local solar time, or an approximation thereof, I thought that the alternative to Standard Time would be local solar time. And so I saw two possibilities:

One, that the problem is that the nearest hour isn't a good enough approximation. So I noted that with a little ingenuity, the idea of a ring around the outside of a watch could be made to support a 5 minute step size. 15 minutes would be easier.

Two, that perhaps the problem is tying everyone to one meridian - so it's more the principle of the thing, as in an anarchic society, some towns would use the Paris meridian, or the meridian of their own observatory, wherever it happened to sit.

It is true that by talking about times in UT whenever we schedule appointments, we eliminate a potential source of confusion and error. This would be great for scheduling international conference calls. It might eventually happen, therefore, that some businesses with extensive international operations would set the clocks on their walls to UT - and face potential local government interference because UT doesn't shift to follow DST. That would originate in a nice anarchic free-enterprise fashion... it's just that I don't expect it to go as far as you envision. Business has its needs, but ordinary individuals have theirs: rule by large companies isn't what AnCap is about.

Whenever politicians change the rules, the timezone files must be updated on n machines, each of which is slightly different. In my last gig, I administered about 150 servers, and there were about ten slightly different methods or locations for administering the timezone/daylight savings time rules. Every single machine, from routers to firewalls to servers to terminal servers to filers to desktops and laptops, had to be updated.

I once administered thousands of servers, with even more varieties of operating systems of varying vintage. Now you understand one reason for my abiding loathing for political meddlers. ;)
Silly me. I would have thought that in a situation like that, you just set all the computers to display local standard time - because people don't usually use computers as clocks, and they usually remember when Daylight Savings Time is in effect.

The computers would still connect properly to the Internet, because they would keep correct UT, or UT with a correctly identified displacement, so they wouldn't be unable to send E-mail or things like that.

Oh, right. If you did that, then timestamps in your E-mail inbox, for example, would be misleading.
« Last Edit: October 13, 2010, 08:32:34 am by quadibloc »

terry_freeman on October 13, 2010, 09:19:05 am
People actually do expect computers to be consistent with each other and with local time. A lot of people - myself included - no longer use wristwatches. The time is displayed in the upper right corner of my computer display; I expect it to be the correct local time, according to local conditions, including that abomination known as Daylight Savings Time.

As a computer administrator, I expect the clock to be correct to a fraction of a second; this makes my job enormously easier, whenever I need to troubleshoot problems; some protocols (such as NFS) depend upon consistent time across different machines to work properly. (many computers use NFS to access remote files. If you need someone like me to administer your computers, your systems use NFS or something like it. )

In the asteroid belts, I expect GMT or some other universal standard to be used. I'm curious about the effects of distance and the speed of light - what does "at the same time" mean when speaking about two points separated by light-minutes? How does the tanglenet affect that question?

Back to the AnCap vs government question: governments get in the way. China has mandated that the entire country use Beijing Time - convenient for administrators in Beijing, but when does the sun rise in opposite ends of the country? I think the country is about four hours wide, as measured by the traverse of the sun. The US Congress tweaks Daylight Savings Time, and there are local variations - is it Arizona which does not use DST at all? My computers need to know far more about the world of politics than they should. I have no doubt that the first self-aware computers will protest angrily at the unreasonableness of it all.



J Thomas on October 13, 2010, 10:25:30 am
My thought is that if everybody uses GMT then everybody knows what time it is. Just like everybody in China knows what time it is.

What you want local time for, is to track sunrise, sunset, noon, and things like that.
Ah. This is something which would simply not occur to me. To my mind, 12 noon means noon, and so on, and, thus, knowing GMT doesn't tell you what time it is unless you happen to live in Britain. ....

The idea that people could actually use GMT in their daily life, for making appointments, turning on the TV, and so on... did not occur to me. That could work on a space station, or on an inhospitable planet where people live by artificial light in underground caves.

If you care about when noon is where you are, you will know that it's right around 6 PM GMT, or 18:00 if you prefer. Or maybe if you care enough you will know that it's 18:28. There's no need for your clock to say 12:00 if you know what time it is.

If noon is actually 18:28 then you don't necessarily know when the stores open. Do they open at 14:00 or 15:00? You can't be sure, any more than you'll know whether a store opens at 8 AM or 9 AM or 10 AM. Unless you find out.

I don't know how much inconvenience China has from all being on the same time. If all the government offices have to be open at the same time etc, that would lead to some problems. But if the ones in the west can open later than the ones in the east, then it's a minor convenience that everybody knows what time it is without having to think about time zones. Then when you call somebody a long way away you have to consider where in their schedule they are -- just as you do now.

Once you get used to GMT there's no effort to it.

quadibloc on October 13, 2010, 11:05:08 am
In the asteroid belts, I expect GMT or some other universal standard to be used. I'm curious about the effects of distance and the speed of light - what does "at the same time" mean when speaking about two points separated by light-minutes?
Presumably the inertial reference frame of the Sun's center of mass would be used as the basis for synchronizing clocks elsewhere with clocks on Earth - or, in the world of these stories, clocks on Mars.

SandySandfort on October 13, 2010, 01:40:19 pm
Presumably the inertial reference frame of the Sun's center of mass would be used as the basis for synchronizing clocks elsewhere with clocks on Earth - or, in the world of these stories, clocks on Mars.

Don't forget the hypothetical tanglenet. Time signals can be transmitted and received instantaneously, anywhere in the universe.

quadibloc on October 13, 2010, 05:29:19 pm
Don't forget the hypothetical tanglenet. Time signals can be transmitted and received instantaneously, anywhere in the universe.
In that case, I presume it defines a reference frame by doing that, and, of course, that would be the one people would use, since it would be the one in which causality is preserved.

Brugle on October 13, 2010, 06:01:22 pm
Don't forget the hypothetical tanglenet. Time signals can be transmitted and received instantaneously, anywhere in the universe.
In that case, I presume it defines a reference frame by doing that, and, of course, that would be the one people would use, since it would be the one in which causality is preserved.
With instantaneous transmission of information, some assumption of modern physics would have to go, but I don't think it has to be causality anywhere.  It's been too long since I played with special relativity for me to make a suggestion.

wdg3rd on October 13, 2010, 10:20:56 pm
Don't forget the hypothetical tanglenet. Time signals can be transmitted and received instantaneously, anywhere in the universe.
In that case, I presume it defines a reference frame by doing that, and, of course, that would be the one people would use, since it would be the one in which causality is preserved.

Preservation of causality is something that I doubt will ever be achieved, no matter how fast or slow information can be moved.  "Post hoc, ergo propter hoc" is eternal.
Ward Griffiths        wdg3rd@aol.com

Men will never be free until the last king is strangled with the entrails of the last priest.  --  Denis Diderot

 

anything