Comment by RegnisGnaw | Hacker News Reader

Comment by RegnisGnaw | Hacker News Reader

Parent Full thread

It is "decided" which is correct. With 24 hours a day and 360 degrees of longitude, it should be simply 15 degrees per hour.

The number of time zone that a country has is a political choice, perhaps based on geography somewhat. The size of each time zone is also variable depending on the choice the government makes.

* Look at China for example, one time zone.

* India is the same with one time zone.

* Mountain Time Zone in Canada actually stretches across 4 time zones (-9, -8, -7, -6)

* The Pacific Time Zone in Canada actually stretches across 2 time zones (-8, -9)

* Newfoundland in Canada should be geographically in the -4 time zone, but it choose to do -3.5 instead. If that's the proper value, then I'd argue Maine should be in -4.5 instead of -5.

* France should be in the UTC time zone with a tiny slice in the +1 time zone. In reality it is in the +1 time zone.

* Heck, why does time zones go from UTC-12 to UTC+14? What's with the values above UTC+12?

See: https://qph.cf2.quoracdn.net/main-qimg-0eb6e2b279845c866f903...

China's time zone is interesting, mostly a reflection of the fact that like 80% of the population lives in a north-south strip along the coast, largely within one "natural" time zone. The further you are from the east coast, the weirder it is. It may be 7 am in both Beijing and Urumqi, but everyone in Urumqi is still asleep because the sun won't be up for a couple more hours.

The far west of China sometimes uses a local time zone, Xinjiang Time that is two hours ahead. That indicates that two hours is too far for one time zone. I get the impression is that usage of local time is on ethnic lines and may have disappeared after crackdown.

And there's also the decision to use hour-aligned timezones. Why not half hours? Why not two hours? It's very much a conscious decision.

Previous to the rise of the railroad, each place's time was based on solar noon. That is, when the sun was at the highest point. Which worked fine. It only became untenable when you needed to manually maintain and publish a railroad schedule. So time zones became a thing. Initially they were on a per-railroad basis, but eventually they got unified.

But now we have computers to do all the bookkeeping. So if we wanted, we could not only return to the previous approach, but take it further to infinite time zones. Your noon is whenever the sun is overhead for you. Our calendars always include locations for events, so the time difference is compensated for. Google Maps knows that if you're traveling east or west a significant distance to make it to an event, it needs to adjust the time. Etc, etc.

We should discuss this in a zoom call. I've scheduled it for eleven-fleventy χm. Don't be late!

...the need to synchronize global tasks means that either you have easily-offset local times, or you just use UTC everywhere and keep it in your head when local wake / sleep is and refuse synchronized tasks outside those times.

No, that's part of the manual-scheduling paperwork. I get that you're used to it, but we could eliminate it. The dialog would go more like: "I looked at our calendars and put it at a time that seems to work for both of us. Feel free to move it if that's better for you." Which is what I often do anyhow with Google Calendar with shared access (as when we work for the same company) or as with Calendly.

https://qntm.org/continuous

That is a good rant, but I think it's wildly overstated. E.g., The whole notion of a timetable is an early-railroad-era artifact. By modern standards, it's a terrible interface. Them not being viable is great.

I should note that I'm not seriously proposing it. My point is more to get people thinking. Our current hodgepodge approach is extremely historically contingent. I personally believe that there's no good solution to this, just different bad ones.

Knowing the longitude between yourself and the event is enough information. A sundial is always locally correct as well. This could be calculated in Airplane Mode using only GPS too. Longitude 0 would be the universal clock for network applications.

Uh oh. I'm starting to think this is a good idea...

1 hour is easy to remember, makes the math easier, and not so far apart that it creates radically different concepts of time for people on the extreme ends.

With one hour time zones:

* If someone is 3 time zones ahead, they're 3 hours ahead

* There are 24 time zones and 360 degrees neatly divides by 24

* 2 hours between time zones means that within a time zone, the sun rises on one side at 8 am and on the other side at 10 am.

We have :30 and :45 time zones though, no :15 time zone. So that doesn't quite apply.

Doesn't it?

I was explaining why 1 hour as the standard makes sense instead of 30 minutes or 2 hours. The :45 time zones would still be an issue for 30 minute intervals and :30 and :45 are still an issue for 2 hour intervals

But the 1 hour is not a standard, we have many places with :45 and :30 time zones. For example, Newfoundland is -3:30 time zone. Nepal is in a +5:45 time zone.

Again, I never said it was the standard. I was saying why 1 hour offsets make more sense than 15 minutes or two hours. I never said different offsets don't exist, I said why they make less sense

And those exceptions demonstrate why it's a bad idea to go beyond hour offsets. It's just added hassle.

Probably because we like round numbers.

But there are half-hour aligned time-zones. Newfoundland is -3.5. It's currently 11:20 am here in Toronto and 12:50 pm in St. John's. Then there's the Chatham Islands of New Zealand with a quarter hour offset at UTC +12:45.

There are also regions where two bordering time zones change by more than 1 hour. At the Pakistani-Chinese border you gain or lose three hours.

Why have time zones at all, and not simply recognize that day and night occur at different times in different places?

This part wasn't a conscious decision, because the reference frame of time and orbital mechanics were discovered long after the practice of timekeeping. We may have a similar problem on a galactic scale some day.

The primary issue is when the day changes in the middle of the local day. Assume everyone adopts UTC. Then in California, midnight will happen at 5pm local time.

  * Friday UTC 2300 = local time 4pm PST  - This is Friday
  * Saturday UTC 0100 = local time 6pm PST  - Is this Saturday?
  * Saturday UTC 2300 = local time 4pm PST next day  - This is still Saturday?

When would Saturday afternoon be? We'd have to change our vocabulary, and I don't see that worth it.

We don't have to change any vocabulary. Noon and midnight happen at different times in different places and that would be fine. We know it's fine, because that's how it currently is in China.

Most servers run on UTC under the hood, and a lot of big organizations have experience already in terms of doing all timekeeping calculations in UTC. It may seem weird, but works pretty well in the long run -- and not having to deal with corner cases like DST or geographical ambiguities around time is extremely valuable.

Have you ever thought about what it's like to live near the edge of a timezone? In a city where half the city uses one time and half uses the other?

"Let's go see a movie Saturday afternoon." "I need this done by Thursday." All these statements are common and completely understandable by everyone. Change to UTC, and the statements are ambiguous. It will confusing for so many people, with no obvious benefit.

China is fine, because the day doesn't change during daylight.

Yeah, computers should use UTC. It's easy for them to convert to/from local time for the user. This is what computers are good at.

There :30 time zones and :45 time zones. Oddly, no :15 time zones.

>then I'd argue Maine should be in -4.5 instead of -5. Maine knows it should be in Atlantic, it just doesn't want to be more isolated from the rest of the country than it already is https://www.pressherald.com/2017/04/27/maine-may-switch-to-a...

I think it should be either a -4:30 or -4:45 time zones instead :)

Reply on news.ycombinator.com