78 comments

[ 2.8 ms ] story [ 152 ms ] thread
Obligatory reply: http://qntm.org/abolish
Yep. Reality is indeed complicated. We have timezones to manage the complexity.
After reading this I changed my mind about abolishing time zones. Thanks.

However, I still think that the time zones we have right now are needlessly complicated. In particular I dislike that they change all the time. zoneinfo is updated several times a year. Why? It makes no sense.

It's political, but not necessarily in a bad way.

e.g. a state decides it wants to be on the same timezone as a neighbour so that trade goes smoothly.

or more frequently: in most countries, it's up to individual states/provinces whether they follow daylight savings or not; so there's always a few each year that decide to start doing or stop doing daylight savings.

I agree there are other problems, but most of the issues in this case study could be resolved by googling "work hours in Melbourne" or "waking hours in Melbourne".

Is it easier to arrange a common time with someone in a different timezone or someone who is in the same timezone but works a night shift?

> resolved by googling "work hours in Melbourne"

...which is the exact same thing as googling "current time in Melbourne" under the current scheme. Abolishing timezones does not simplify anything, it just shuffles the cognitive work differently.

> Okay, it looks like I'm going to have to do this the hard way. Is the Sun up in Melbourne?

> Let me find a webcam somewhere in that city.

Seriously? Can't just Google "sunrise time in Melbourne"? Sure way to lose credibility fast with writing like that. In fact, finding out the sunrise time was all s/he needed to do to figure out if it was a good time to call the uncle.

I came up with a less articulate version of this as I read the parent article and I'm disappointed the interviewer didn't ask a question along these lines.
This won't happen because there is still a need for a local time reference.
Swatch experimentend, not really seriously tho:

https://en.wikipedia.org/wiki/Swatch_Internet_Time

It would have got more traction if they hadn't chosen Biel.
That was a ridiculous decision. Perhaps they were hoping that if it did catch on, their company would be immortalized as the new Greenwich. Nevermind that they were only hurting the chances of an already extremely unlikely success. Reminds me of Sony memory sticks and Apply Macintoshes.
> not really seriously

??

Your comment is contrary to the Wikipedia article. And I recall big advertisement campaigns and hype around Swatch beat in 1999. And as the Wikipedia article mentioned several other companies supported it.

I guess these folks don't read xkcd: https://xkcd.com/927/
They're not proposing a new standard, they're proposing using UTC everywhere which is what lots of people already do. It's not a new standard, just removal of existing standards.

That's just the time zones. I don't understand their calendar and if that's somehow necessary to make it work. At first glance, the calendar seems more confusing than our current one.

> they're proposing using UTC everywhere which is what lots of people already do.

"A lot of people"? I use UTC for a lot of things, because I'm writing code.

But when I travel, I certainly don't use UTC everywhere; my phone shows me the time for "Roaming" and "Home", and that's incredibly useful for me to guess at e.g., what time to break for lunch, when we're nearing close of business hours, when I ought to get to bed so that I get a good night's sleep, etc..

I suppose the really crucial "standard" we're talking about is the associations we have with different times, and timezones are the (official) attempt to keep that meaning consistent around the world.

I doubt it'll work -- partly just because to make this shift:

-- everyone who operates mostly locally would have to relearn those time meanings (all except the English... Hmm; why are they the only ones who don't have to suffer? Perhaps the new "fixed" time should be based on the most populous time zone)

-- everyone who switches time zones regularly will have to learn a new set of time meanings for every zone they travel to.

To be sure, the second set of people have it hardest (until the phone apps catch up and provide that info automatically). But the first group of people -- the vast majority -- will make sure this doesn't happen easily.

I'd imagine abolishing time zones would add billions of dollars in developer productivity back into the economy, doesn't seem like that bad an idea.
This doesn't really make our jobs as programmers easier. For example, if I go for a bike ride in the morning, Strava calls it "morning ride". Morning will still be a concept based on the position of the sun, so it will have to look up my GPS data from the ride track, look up the time, and then compute where the sun was in the sky at the time in order to assign a label. Compare this to the current algorithm of mapping certain hours to certain words.

Similarly, if you have your screen on your phone set to turn redder at night, you now need to know your GPS location in order to do that correctly.

If you travel, this is going to be confusing. "What time do people start their workday here?" "When do stores open and close?"

Finally, this still doesn't solve the problem of "what's a good time to have a meeting with people in other locations". You'll still have to calculate your offset to the other attendees and see if they'll be awake at that time.

TL;DR: I'll eat my hat if this ever happens in my lifetime.

Those problems already exist and have presumably already been solved for people living in the "wrong" time zone, such as western China which is completely out of sync with the sun.

So instead of having to get both the time zone and the sunlight/work hours right, we'd just have to do the latter.

About store opening hours, if you travel to Singapore (very popular stopover country), you'll have to work that one out - they open much later in the morning than western countries. That's just one example - https://en.wikipedia.org/wiki/Shopping_hours

Partially out of sync with the sun. When I was in Urumuqi, everything started a bit later (10AM vs. 8AM) and ended a bit later. I guess Kashgar would be a lot worse.
The program I'm using for setting screen to red during night is actually asking for the coordinates already since the sun rises and sets at different times depending on where you live. It actually doesn't even get above the horizon in some locations on the earth during winter. Where I live we get about 4.25 hours of sun around midwinter.
Right. That's an application which actually cares about solar light levels in a region.

Most applications care about local time, because when it's 7 AM in December in northern Montana, you're getting out and going places regardless of how much longer it's going to be before the sun comes up. Local morning is only loosely tied to local sunrise, civil or otherwise.

Calculating the position of the sun in the sky is easy. It's predictable. Calculating the current UTC to local time offset is hard. Not only does it require knowledge of your location just like calculating the sun position does, it also relies on a patchwork of arbitrarily chosen rules that are subject to change at any time due to the whims of politicians. (And it doesn't tell you the position of the sun in the sky either, as local noon may be anywhere in a ~5 hour range around solar noon, and you have no idea when sunrise or sunset are or even if they occur at all, which they may not above the arctic circle).

It's confusing when you travel anyway. You have to reset your clocks, and know if/when the local time changes, where the time zone borders are, and what the time offset is to other places you care about like home. And after all that you still don't know when businesses are open or when people eat or sleep because that varies.

It doesn't solve the "what's a good time to have a meeting" problem, but it does solve all forms of the "I thought this time was in my time zone but they thought it was in theirs" problem. It also solves the "having to do mental mod 12 math to talk about time with other time zones" and "having to know the dates of everyone's DST changes" problem.

I agree that this will never happen in our lifetimes.

This involves radically ignoring the human factor in time. Things are still mostly local in daily life, except in the small bubble these men live in.

Next year I'm sure they'll be arguing for decimal time.

agreed.

i have tried living on utc several times.

solar days being sunday/monday becomes very disorienting.

Ah, I was really convinced by the article but I completely forgot about days. That really makes it a nonstarter.
How about we destroy the daylight saving time first, and go from there?
I came in to say exactly this. If we can't even get rid of something that is absolutely outdated in the 21st century like daylight savings time, then there is no shot at switching to UTC globally.
Those living in Scotland might disagree that the 21st century has done anything to remove the need for DST.
As someone who's written a lot of scheduling code (seriously, a lot), I'm sooooo on board.

Time zones? Not really all that big of a deal and pretty much a solved problem. Daylight saving time, on the other hand, is the spawn of satan himself.

There is something so autistic about this plan that it's just beautiful. So here's a guy who thinks that if we all used the same number everywhere (and abolish local politics! yay! Politics yuck!), everything would be easier, discounting by comparison that:

1. When we travel, we'd need to learn the local schedule.

2. When we communicate (say, by phone), we'd have to somehow compute where the other party is on their daily schedule.

3. Our understanding of global events would become muddled ("today at 15:30 an explosion rocked Tokyo". Was that the middle of the day? Middle of the night?), let alone of historical documents.

4. We'd need to educate billions who never travel outside their region about the new way the government services they rely on will now be communicating their schedules. I'm sure they'd appreciate the simplicity.

But, hey, we wouldn't miss our flights (never mind that most of the world population can't afford a plane ticket anywhere, let alone to a different timezone) and we'd all use the same number! Numbers are elegant, so if we could agree on a number, maybe the world would become less complicated. Can't we just all agree that the world isn't really complicated, but that icky politicians complicate it by making up their own numbers for "political reasons"?

How is their plan autistic? Hanke and Henry seem idealistic, academic, and likely lack a solid understanding of the sociological consequences that make their plan utterly unattainable, but they don't strike me as sufferers of autism.
I didn't mean to be taken so literally. I simply meant the tendency to "simplify" the world by reducing away the complexity of people and their concerns and focusing on numbers as if they mean anything in isolation. It just struck me as a very extreme case of what I call "spherical cow social science/politics", which is very attractive to some types. And the problem with their plan isn't that it's unattainable, but that it would cause much more harm than good, because their laser focus has led them to try optimize the wrong things at the expense of many right things.
Yes please! I can't even count the times anymore when we had someone over from London and had problems with the 1 hour difference in making appointments. Luckily, Outlook does a lot of the calculating, but then, you take a look at your watch when you arrive in a different time zone (which is just a short time of travel away) and boom, you're off again.

Just define a "Indicative start work time" for every place, like what would be 9:00 today, or something similar to align events during the day with the locals.

Between outlook, phones synchronizing time automatically and https://www.timeanddate.com/worldclock/fixedform.html being available... how do those issues happen? I mean, what's the process that leads to 1h mistake? I'm seriously interested. Having moved to both extremes and the middle of timezones, I don't think I ever had an issue with that.
You fly over from Londen to Amsterdam, having scheduled something at 16:00, which is in the agenda of the other person as 17:00, but then you check your analog watch. True, it would go well when you would've checked your smartwatch.
> but then you check your analog watch

Ah, I see the problem :)

Yes, I see the problem, too, but I suspect it's a different problem from the one you're seeing. I'll explain...

I suspect the problem you're seeing is, "analogue watch". That's not the problem. (Forgive me if I have assumed wrongly! ;)

As every frequent traveller should know, if you have a watch (any technology) that does not self-adjust (for some value of "self") then it is strongly suggested that you adjust your watch to you destination time zone the moment you get on the plane.

This has long been known, so the problem really is "person who unwittingly failed to adjust their timepiece to their destination timezone".

I'm looking at it from the UX / motivation side. There's no trigger to change the time manually and people are great at forgetting things when there's no external stimulus. I see the issue not as people failed to adjust their timepiece, but as people relying on technology which relies on them to remember to do actions after events which they passively observed. (plane landing)

Theoretically it's the same thing, but in practice "don't make me think" applies. If we can make watches adjust themselves, it's a really good idea to do that.

There's no trigger to change the time manually

When you land there is usually an announcement: "welcome to [city], where the local time is [time]."

(comment deleted)
It would help if application developers recognized that GMT is not a time zone, and that the UK uses BST for half of the year!
Destroy the Sun.

All other proposals solve only half of the problems and lead to ugly hacks.

No need to go to that extreme. All this idea really requires is to surround the planet with a dome that captures the sunlight, and somehow distributes it to the entire planet to create a single worldwide day cycle.
Flatten the earth into a disc and rotate it?
> Nepal is –inexplicably – the only country in the world to have a time zone that is set to 45 minutes past the hour.

What? Australian Central Western Standard Time +8:45; New Zealand, Chatham Islands +12/13:45

https://www.timeanddate.com/time/time-zones-interesting.html

ACWST isn't an official timezone and only a few hundred people live there.

http://www.howderfamily.com/blog/australias-weird-little-tim...

Chatham Islands is a bit of a mystery. Again, almost nobody lives there but it's a dirty stain on lists of time zones.

Official enough for me ;)

    /usr/share/zoneinfo/Australia/Eucla: timezone data, version 2, 5 gmt time flags, 5 std time flags, no leap seconds, 19 transition times, 5 abbreviation chars
Seriously though - if people are using a timezone, it exists. We may call it unofficial, but that doesn't make it go away.
And I doubt it's inexplicable, these things tend to have a long history of becoming more and more regular.

The Netherlands were on GMT +19 minutes 32.12 seconds until 1937, so that noon in Amsterdam was exactly right. Then a lot of things happened and only since 1981 are we in sync with other western European countries.

I've talked about this with lots of people over the last 10 years or so. Most people I know (non-programmers) look at me like I'm crazy when I suggest that this is a good idea. I don't think that most regular people see a benefit to this, and since there is cost associated with change, the idea is unpopular.

The thing that strikes me most about this proposal is that they are looking for a top-down solution. We already have UTC, it doesn't need to be invented or implemented, just adopted. Maybe a bottom-up approach will have more success?

If you look at a country like the US that still doesn't use metric, it's hard to imagine having world governments adopting a change like this. I guess there is always hope though.

> If you look at a country like the US that still doesn't use metric

Like? Only the US. All other countries converted to metric at some point.

Same with time zones and DST. Those weren't made by God on the sixth day, governments made the migration to them deliberately.

The problem they are addressing is already solved. Simply say at time X Y, where X is the time and Y is the offset, preferably an established offset.

We meet at 12:00 GMT

Done

Everyone only needs to remember one number, their local offset to GMT. If someone uses an, to the receiver, unknown offset, then look it up and do some first grade math.

This is very hard when you first have to explain what timezones are for some people (no countries mentioned...). When I was playing MMORPGS most, 10 years ago this was the main problem to set a time for a quest.
Small battles first, then the war: first show that you are capable of discarding the stupidity called Daylight Saving Time. Then take on time zones themselves.
Hear! Hear!

DST is an abomination that needs to go. I wish we could make devices that display the correct time without needing an internet connection.

I believe most mobile phone cell towers transmit local time. I'm not sure if you need to be subscribed to a service or not. If not, then you basically only need a cheap GSM module, costing around $10.
> HH: No it is NOT inefficient. It combines the best of both systems: One universal time, combined with local work time connected with the sun being up.

So basically his plan is to rename "time zone" to "local work time", then rename "UTC" to "time zone".

Nice job.

You did nothing, except change the names of things. You still have a time zone, it just has a new name and works slightly differently:

Now instead of the clock keeping track of "local" time, and remembering UTC in your head, your clock remembers UTC, and you have to remember "local" time in your head.

I'm not convinced this does anything useful at all.

Simpler proposal:

All (physical) clocks display two times. Done.

In places with timezones that differ only by the hour you can even make the clock read: 11^8:35:24 - with the number preceding the caret being UTC.

Now the clock keeps track of both things for you.

We are not going _ever_ to solve this politically.

What about educating those who need to think in global time about the existence of UTC. This would let us use UTC for coordinating inter-TZ events.

As an example, say I'm in Tokyo and I've got to make a call with someone in California. I could suggest "my 17:00 (8:00 UTC)" and if this didn't work she could reply "Nah, I've got lunch then. What about my 11:00 (7:00 UTC)"? At the same time we clarify our local time (I remember that she eats at my 5PM) and also the absolute time (8:00 UTC vs 7:00 UTC).

I've done this some, and it works. Where it falls down is practice. We simply aren't used to doing it, but we will learn out of necessity.

Instead of implementing a global time zone, let's just learn to use the one we already have as a common language to describe times. This is a distributed, bottom-up solution to a distributed problem.

Agreed. This movement should be replaced with a movement for remote teams to use UTC as their reference time. Right now, we use headquarters time, but less than a third of the company works there, and no one in the other 2/3rds of the company is used to translating to that time zone. Every programmer has some experience translating to UTC, and if we make it the standard for remote teams, everyone will get used to it. Fewer mistakes will be made.
Does the average American know their UTC offset?

I ask as I doubt the average British person would know, and it's zero for part of the year, and +1 for the summer.

But Americans have multiple time zones in their country, and we don't.

No, only the average American programmer. Most others wouldn't know what UTC is and haven't had to think about GMT since elementary school.
Which would mean time zones do actually work really well for most people, right?
Doesn't it already happen in any sane distributed team?
>This map at the top of this post gives you an idea of what the world looks like now, and what it would like if we instead stuck to single system of Universal Time.

It was not necessary to show a map of what a single time zone on a map looks like :)

(comment deleted)
This is so backwards, it's hard to know where to begin...

We already have this universal standardised time he's talking about called UTC. I can set a calendar invite to UTC, and everyone magically gets it in their time zone.

Time zones are useful because they answer questions like: is it a reasonable time to call? Is it during lunchtime there? Is the office open?

Without time zones, you'd need to mentally store information about each place's habitual times in your head instead.

Seems to me anybody who wants to switch to UTC can simply do this. If aviation people can already do this so can any other company that may find an advantage in it.
How would you deal with the fact that the "day" (Monday or Tuesday, for example) could change at any time? Delaying a "meeting on Monday" by an hour could turn it into a "meeting on Tuesday"!
I wish a week were consisted of a number of days divisible by more numbers instead of a prime number (7). If a week was 12 days, we could do a routine every 2, 3, 4 or 6 days and it would still be the same day every week. E.g. If I wanted to run once three days, I could do that on every 1st, 4th, 7th and 10th days and it would fit the week perfectly or if we want to do a regular business meeting twice a week, we could do that on 1st and 7th days and the number of days in between would always be the same. Our schedules would be so much more neat, which could potentially yield to much more order and productivity.
I think we should totally adopt blockchain time. Typical time will look like "be03d0988bfa799f7d7ef9ab3de97ef481cd0f75d2367ad456607647edde665d6f6f and sixty eight seconds". Very convenient and also VC backable right now.
Dibs on the name Clockchain. We're fully logging every second that elapses. You'll be able to verify how many seconds were in every minute, how many minutes were in every hour, etc. Very handy for generating test cases that expose subtle bugs involving leap seconds. Check us out at clockchain.io.
Utc everywhere, yesterday please