Hacker News new | comments | show | ask | jobs | submit login
So you want to abolish time zones (qntm.org)
159 points by subleq on Jan 17, 2015 | hide | past | web | favorite | 124 comments



As someone who grew up in a country where 24 hours are the norm for digital displays, I'm a bit miffed that the author seems to portray that as alien and weird.

You wouldn't say "oh-four" to indicate four AM. You only need to say "oh-four" if the 12 hours system is the norm and you need to clarify you're talking about "military time". You wouldn't need to say "twelve hundred hours" either.

I'd recon if English speaking countries switched to the 24 hours system, they'd likely just do what everybody else has done: keep counting. It'd be far more natural to say "it's 13 o'clock" (just as you would say "it's one o'clock" when it's obvious from context whether it's in the middle of night or early afternoon).

And "midnight" and "midday" would likely still refer to the "solar time" in the scenario where everybody uses UTC (though it's obvious the author is really just getting at how we would re-invent time zones eventually out of necessity).

I'm not really sure what the article is on about, though. Did anyone ever really argue that we should abolish time zones in general? As far as I can tell most people who are annoyed with time zones (particularly programmers who have to deal with them a lot if they need to deal with date time logic directly) would only want to get rid of DST.


> As someone who grew up in a country where 24 hours are the norm for digital displays, I'm a bit miffed that the author seems to portray that as alien and weird.

As a german, i gotta agree. We have (translated) "four in the afternoon" or "16 'o clock". (4 Uhr Nachmittags, 16 Uhr.) We also never say the leading zero on hours, and depending on context it doesn't even need to be clarified that we're talking about "'o clock", these are perfectly fine exchanges: "When will you go to the beach?" "Three twenty." (Wann geht ihr zum Strand? - Drei zwanzig.) The context makes it abundantly clear which is meant.

There's also the fact that nobody uses the construct "xx hundred" for numbers above 9. The years before 2000 are still referred to like that, but outside of that people will consider you very strange if you try and buy "twelve hundred grams of minced meat". And for times it's especially not used.

The author's awkwardness with 24 hour times entirely on himself.

Edit: To make this post a little useful, this website is extremely useful for dealing with people in different timezones:

http://everytimezone.com/


FWIW, in Swedish it's perfectly normal to say the leading zero on hours, to use the "xx hundred" construct when buying twelve hundred grams of minced meat, as well as to refer to years after 2000 as "twenty hundred and fifteen" (tjugohundrafemton). I think the reason the post uses the "twelve hundred hours" is because the American military does.

That said, I agree that there's nothing awkward about 24 hour times.


> http://everytimezone.com/

It's a great and well designed site. Works offline too!


> Edit: To make this post a little useful, this website is extremely useful for dealing with people in different timezones:

> http://everytimezone.com/

You can now just type "12 pm in Lima" in Google and it will give your the equivalent in you local timezone.


we're talking about "'o clock"

As an aside, in high school I had an English teacher with such clear and correct diction, that people said you could hear the apostrophe in "o'clock"...


'There's also the fact that nobody uses the construct "xx hundred" for numbers above 9.'

Count me amongst those who do. By subjective reflection - and standing significant chance of being wrong - I would say most frequently when I've had to do a bit of arithmetic on values where rounding to hundreds is most reasonable.


> nobody uses the construct "xx hundred" for numbers above 9.

I use it all the time. What's easier to say, "one thousand two hundred" or "twelve hundred"?


This is cultural. In Polish there's separate word for n-hundred. 200 is dwieście, 300 is trzysta, etc. So nobody thinks about number of hundreds and tousands as connected thing, we just divide the number differently because of language.

Imagine someone told you "twelve tens" for 120 :)

Anyway, "tysiąc dwieście" for 1200 is easier than "twelve hundred" or "dwanaście setek".


As a German watching some English podcasts I am now used to it, but at first I pictured the trailing zeros in my head to understand it fast enough in spoken language.


I think it's an American English dialect thing. I'm from Ireland, a dialect close to British English, and always found it odd when they say it on US TV


Depends on the language. In some, "one thousand two hundred" is about as simple as "thousand and twondred", while "twelve hundred" is "twoovertenhundred" and quite a mouthful.


"Twelve hundred" is no weirder than "Ten thousand."


Nor stranger than 1 dekaliter. Some units are just used more often than others.


How about twelve tens for 120? That how it sounds to me ;)


You mean a decadozen?


> You wouldn't say "oh-four" to indicate four AM. You only need to say "oh-four" if the 12 hours system is the norm and you need to clarify you're talking about "military time".

The use of "zero-four" in military time is not to clarify that they are talking about "military time" vs. some other time system but to minimize the risk of ambiguity in environments where hearing might be difficult (noisy environments, poor communications links, etc.) and a mistaken communication could have serious consequences. If you don't have a clear, two-digit hour, you know you've missed something and need to ask for a repeat.

Presumably, the use of "oh-four" by some in non-military use of 24-hour time is an echo of that use.


> Did anyone ever really argue that we should abolish time zones in general?

As I said in response to another thread, yes. People argue to abolish timezones. An example from Vox: http://www.vox.com/2014/8/5/5970767/case-against-time-zones


This map shows the relatively few number of countries still using 12-hour time (instead of the more common 24-hour).

http://www.talkstandards.com/wp-content/uploads/2011/02/12-h...

The term 'military time' is funny. It's not a military thing. It's more like most people on earth has switched, including the US military -- better if both Alpha- and Bravo-squads attack at 07:00 instead of 7 pm/am respectively :)


Yeah 24 hour system makes a lot more sense. My favorite to confuse people not familiar with the 12 hour system is to ask them how many hours there are between 11pm and 12pm.


Hmmm... there is no 12pm? There's 11:59pm and then comes 00:00 (which I'm not sure wether it is AM or PM)?

Is that somewhat correct?


It goes 11:59pm then 12:00am 12:01am ... 12:59am ... 1:00am ... 11:59am 12:00pm so there are 13 hours from 11pm to 12 pm.


No no no no no.

The author is making it easy for himself with a contrived example with a key information already available.

> I want to call my Uncle Steve in Melbourne. What time is it there?

In a world with timezones, if your uncle is on the Internet and says "call me at 13h30" but doesn't tell you where he is, you can't call him, because you dont't know when is 13h30. (Think about that... he just gave you the time, but you still don't know the time...)

With timezones, you can't just give the time, you must also give a location. "Time" doesn't exist with timezones, only "time-location".

It's as if a physicist couldn't give you the mass of something without giving you the color.

"This table is 10kg-blue, which is the equivalent of 20kg-pink." (Because kg-colors aren't equal everywhere.)

On the other hand, with a universal time (UTC for example), 13h30 is the same for everyone, everywhere. The time can be given without any other information. If someone says call me at this number at 13h30, I can just do that without having to know where he is.

Let's look back at the initial problem: > I want to call my Uncle Steve in Melbourne. What time is it there? Well, if I have a universal time, it's very easy: it's the same time as where I am! I don't even need to ask google! Beat ya!

But of course the author want to know if he can call someone out-of-the-blue by knowing where he is. Well, if you can use google to tell you what is the time in another timezone, you can use google to tell you what is the usual waking hour in this location. You don't need additional information in this example.

With universal time, the worst case scenario is needing the same amount of information as we do today (current time + location). The best case scenario is having a sane way of communicating time.


In both cases, you need a location to solve the types of problems humans want to solve. A hypothetical: is now an acceptable time to call a tech business in Asia from the US?

- With timezones: I need timezone (location). Generally I can call 9am to 5pm.

- Without timezones: I need location so I can figure out the range of times people typically work in that area (perhaps it might be 11pm to 7am).

And you're back to solving the same problem, just in a more complicated way with no helpful conventions.

In a world with timezones, if your uncle is on the Internet and says "call me at 13h30" but doesn't tell you where he is, you can't call him, because you dont't know when is 13h30.

Nobody who's used to communicating across timezones has this problem. After being bitten in the ass a few times you learn to say "Call me at 3pm eastern" or whatever. The no-timezones cure is way worse than the timezone disease.


> And you're back to solving the same problem, just in a more complicated way with no helpful conventions.

Not more complicated, exactly the same amount of complicated. You can add the conventions.

> After being bitten in the ass a few times you learn to say "Call me at 3pm eastern" or whatever.

And then you still screw it up because DST changes a week later in one country or another, or some such. Even when you do it frequently, it's not a solved problem by any means.


Not to mention, people have different times that are acceptable. And they travel. So at home I'll get up around 11am to 2pm, local time. When visiting SF, I'm usually up and about by 10am. I work with people all over the world, and they keep different enough schedules that I can't go by their location. And they travel, too.

So the author is right that in some weak contrived scenario, using a highly interruptive comm service, maybe zones help. Event they don't, cause Google could trivially give you the "apparent solar time" going by geography, anyways.

Timezones are an annoying hack. Daylight savings is even more obnoxious. Especially in poorly run counties that make up DST rules on an ad hoc basis.

In the context of programming, you should almost always be using UTC, maybe with an offset. Several times I've run into companies that keep call billing records in local time. Compete with extra hours and time traveling calls around DST jumps. Moronic. (And leap seconds can go get fucked, too. I'm saddened that we have to put up with such s pointless, annoying thing. Maybe we could switch to leap hours. No noticeable effect for millennia, at which point relativity should be a much larger issue with timekeeping.)


The author also doesn't help his case by pointing out the number of exceptions in using the existing system as a rule of thumb. Never mind 4am, what about calling someone at 4pm their time. That might be after the end of their working day in the Baltic states, or just after lunch in Spain.

Depending on location and specific business practices, 9:30am their time might be a good time to call in mid morning, but it might actually be the time they usually arrive at work, or in some Chinese cities the time they usually get out of bed. If it's 9:30 am on Friday it could be, like Europe, the last day of the week in some Muslim countries, the first day of the weekend in others, the only day of the weekend in Afghanistan, and subject to quite a bit of variety somewhere as full of Western run companies as Dubai. Most of these countries changed the days of their weekend in the last decade or so (which suggests problems with religious holidays are not insurmountable...). And that's the standard work week, without getting into the minefield of regional and religious holidays, and individuals' own holiday, sick leave and round-the-world travel...

Suddenly a universal status update system (or algorithmic handling of how inbound calls are passed on) looks far more necessary for the digital age than time zones.

The only thing in favour of time zones as a hack to assess someone's availability [given assumptions about their location] is their universal acceptance.


So either way, we need a global lookup database. Since changing timezones would have massive switching costs, what the benefit of doing it?


Scheduled calls are never miscalculated, no DST, no AM/PM mixups.

I hate timezones, and I especially hate switching times ie: DST. Having to work across timezones I have made these mistakes which were not only inconvenient, but screwed up my schedule for the rest of the day: Prepared for a call 6 hours early (oops I was supposed to add 3 hours, not subtract), missed a call (oops I was supposed to subtract 3 hours, not add), given a schedule that was 8AM, 8AM, 8AM, 8PM, 8AM, showed up 12 hours early for the 4th appointment.

Finally, I don't know if you realize this, but DST time is correlated with higher heart attack rates, increased traffic accidents, and a slew of other bad things.


If you're calling a specific person, you very likely know where they are. People don't move around that much. If he's not in his usual place, he's probably told you where he is if he's expecting a call from you. And even then, if it doesn't want to reveal or burden you with knowing where he is, we already have a universal time stanard he could use for this ("call me at 13:30 UTC").

Decoupling time-of-day for daily life from the position of the sun in the sky is the biggest over-reaction for the smallest barely-a-problem ever.


While this is an amazing and thought-provoking article, I can't help but feel it missed the obvious answer to the question it presented.

Phones should be programmed to accept or reject calls by the owner, because it is a really dumb idea to be able to wake people up knowing only their phone number. Notably, this is already the case: cellphones have mute/vibrate switches, and if you have a smart phone you probably have various "do not disturb" modes that do more advanced routing.

As a followup and to actually resolve the question for the call placer, phones could broadcast their availability status and this status could be queryable by phone number.

Stardates have their own problems but I'm not convinced those problems are harder than the problems we have with timezones now. We've built up a terrible amount of infrastructure around timezone handling, we should not be scared off when stardates require unique infrastructure projects.


In general, the software revolution has made me think back on how much many traditional devices around my house suck.

I pay a lot for phone service and the actual quality of the service they provide to me is terrible. They actually allow illegal phone spammers onto their network. I get phone-calls from robots about free trips, on a service that costs me more than my NetFlix account that I get a lot more enjoyment from.

I get Gmail for free and it filters that crap out.

And don't get me started on the piss-poor UIs of the traditional TV industry. Holy crap.


It's fairly difficult to stop robo dialers. I used to handle about a billion calls a day for providers. Robo calls had to go on separate connections and had other limits and higher prices than regulal ("conversational") calling.

But many customers would slip in lots of dialer anyways. You can tell after the fact, by looking in aggregate. But on a per call basis, nope.

Now, we tried blocking repeated caller numbers. Spammers would just switch to using random numbers. For many reasons (some good) it is simply not tractable to know if s number is s valid source for a call. It's vastly more complicated than say IP spoofing.

The most effective solution is to have an answering service/program that screens calls and allows known callers without further hassle.

Your provider could offer a service to block "known" bad numbers, but a lot of business use the same number for many things. I'd be surprised if some consumer services don't offer such services. Though, there may be some regulations that require them to attempt to complete the call.

Also note that some dialer, like political calls, are expressly allowed and don't have to follow do not call.

The best recourse you have after the fact is to make an FCC complaint and persue it. I've seen tons of complaints that go nowhere because no end user pushes the issue. For really illegal dialer, there's gonna be several intermediaries, say, 5+ isn't surprising. Each one has to escalate to the next. Without a fire under their ass, most providers will close the complaint with effectively "can't repro, dunno".


In the UK there rules about spam calls too, however most of the calls my parents get are international numbers. I told them just to ignore them, but they got upwards of 5 calls a day so it was pretty annoying.

I setup a PBX so that I could block any repeat callers, but that didn't really work because each call is from another random number. There were also a few automated calls so I set it up to play an automated message of 30 seconds saying to press a code to speak to someone, if the number is withheld or international. That ended up blocking pretty much all the calls. I guess the dialers interpret it as a voicemail greeting and move on.

I've since found out there are plenty of SIP providers who if you ask them nicely will let you send any callerid without proving ownership - even for international calls. When it costs them around half a US cent per minute (probably less in bulk) it's not surprising there are so many spam calls.


Another good technique to reduce the number of robo calls is to have your PBX answer immediately, and play a fraction of a second of the "number unobtainable" tone, which is enough for many greedy "robo's" to give up and move on where as a human will barely notice.


Providers have to take calls without ownership. It's simply not feasible otherwise. Sometimes, I'd be handling calls going from one landline neighbor to another. There's just too many intermediaries.

Now, for small, low volume deals, like say, Twilio, yeah you could force an ownership test. But it would have no impact on the real offenders.


I grew up in Ireland. Very few robo calls. Why? It's illegal.

The solution is not a technical solution, but a legal one.


Sure. It's mostly illegal in the US,too. But the FCC simply doesn't pursue the matter enough. Even in illegal things, like guys that setup a line pretending to be the IRS, the worst that happens is someone cuts some little account off.

It could be different. And as a carrier, I've begged other carriers to push. Have their end user escalate. Get LE involved and get them to go all the way till they find the originator of calls and nail em. But out of the few hundred "complaints" I've seen, only one or two even pretended to care, and none ever really followed up.

The US could kill illegal robo dialing, at least a lot if it, just by getting serious and slapping fines down.

It's so dumb, that even the FCC put out a prize for the best anti robo dialer tech someone could come up with. It defies explanation.


We do have a very strong law, the TCPA, that prohibits robocalls to cell phones without prior consent, with a private right of action and a $500-$1,500 _per violation_ (per call at least) penalty. It won't do you much good against scammers, but against actual businesses with assets that could be seized if necessary, it's awesome.

(I personally sued a debt collector in small claims court for wrong number robocalls to my cell phone, and settled for the entire amount I was asking, $500 per call.)


Yes there are 2 kinds of illegal: de jure and de facto. You need it to be illegal both way.


A network that allows anonymous agents to spoof their address is absurdly broken. yes, there are legitimate technical reasons for that problem to exist, but they've had like a hundred years to figure out how to fix them.


> phones could broadcast their availability status and this status could be queryable by phone number

There was a standard for this in the 00's called "Wireless Village" that got so far it was usually integrated into the phone books of dumbphones by Nokia, Motorola and Sony Ericsson, but it never got any uptake.

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


Agreed. Phones are silly; either use a system that provide a status such as "available" or "away", or use a system that by convention is answered by the recipient at their convenience (which for some crazy reason doesn't normally include telephones, though it easily can with current technology). If you need to schedule a meeting in advance, suggest a time that shows up as "free" on their calendar.

I'm often up at 2AM, in which case I'll hopefully still be sleeping at 9AM. Sometimes I have 6AM meetings, in which case I'll (hopefully) be asleep early the previous day. I also travel to conferences halfway around the world. My local timezone is not a good way to guess when to call me.


Or an even simpler solution you can implement today: text your uncle "hey, is now a good time to call?"


Both of you guys missed the point.

The problem that people are trying to solve by abolishing time zones is the annoying need to deal with time differences when scheduling meetings. The "I need to consult a table to schedule meetings" problem.

But we've just moved the problem under another shell. Now, in order to schedule meetings, I have to know what the universal "04:00" means for all participants. This is just the flip side of the same coin.

The real fundamental problem is that people like to work "during the day" and sleep "during the night". Either we solve that problem with time zones tables, or we solve it with "what is the solar time of day" tables. The author of this article is arguing, and I agree, that trading one for the other doesn't improve the situation ("is not simpler"), has a lot of human transition costs, and that the new problem may not ACTUALLY be an improvement over the old problem.


The problem that people are trying to solve by abolishing time zones is the annoying need to deal with time differences when scheduling meetings.

No, abolishing time zones and DST would also remove huge swathes of code in all of today's operating systems, numerous costly bugs, and make time unambiguous without qualifiers like time zone or location. There are lots of reasons to do it.

As others have pointed out, the real fundamental problem is best solved by people indicating their status somehow now and in the future, because not everyone in a country works 9-5, not everyone is available all of that time for phone calls, and why should everyone work to the same schedule anyway?

If you want to schedule calls, use a calendar, if you want to make a call right now, use an availability indicator.


"No, abolishing time zones and DST would also remove huge swathes of code in all of today's operating systems, numerous costly bugs, and make time unambiguous without qualifiers like time zone or location."

Because we can be certain that, going forward, we will never need to talk about a time before we made the change.


> But we've just moved the problem under another shell. Now, in order to schedule meetings, I have to know what the universal "04:00" means for all participants. This is just the flip side of the same coin.

The failure cases are much better though. Say you screw this up, and you schedule the meeting at what you think is "09:00 Bangalore" but it's actually "07:00 Bangalore".

With timezones: You tell the guy in Bangalore to dial in at 09:00, he says fine, then he misses the meeting.

Without timezones: You tell the guy in Bangalore to dial in at 04:00, he either sucks it up or says "umm I'm still having my breakfast then" and you reschedule.


I just want two call buttons. One for "call, but don't disturb if person's busy" and one for "call with a high priority, we have an emergency".

And an option to punch someone in the face over the phone if they had used the second button when they should've used the first.


What a bunch of hoo-wee. Lets start with the simple fact that China, which covers five time zones, doesn't use them. It's that same time in China no mater where you are in China, and they seem to be getting on just fine. Next consider just how insanely inconstant the time zones actually are. They aren't straight lines up and down, rather they zig-zag all over the place --sometimes even skip over one another in "time-zone islands". There is no way to know any of this except to ask Google. And as for the contrived example of calling one's Uncle. Just ask him what a good time is to call. That would be the polite thing to do anyway. And hey, you won't get it wrong when he says "2 o'clock".


> they seem to be getting on just fine.

Maybe they aren't.

> Imagine that you've found yourself in Kashgar, the western-most city in Xinjiang, China’s western-most region. Your friend sends you a text message and tells you to meet him at 3 pm. Sounds pretty straight forward, right? Not in Xinjiang. If your friend is of China's majority Han ethnicity, you can assume that by 3 o'clock he's referring to Beijing Standard Time. But if your friend is a Uighur, the largest ethnic minority group in Xinjiang, he might be referring to “local time,” which is two hours behind.

http://www.theatlantic.com/china/archive/2013/11/china-only-...


The reason it "works" in China is because a totalitarian regime has forced it on the people, and if you live in western China, everyone knows the workday is 11-to-7 instead of 9-to-5.

The thing you're missing is this: How can Uncle Steve know he should say "call me at 2"? Either he has to do a lookup, or he has to internalize that Australia is a place where the workday is 23-to-7.

And where I am, the workday would be 1-to-9, the east coast would be 4-to-13, Paris would be 10-to-6, and Bangalore would be 14-to-23. And this system made everything simpler... how?


How can Uncle Steve know he should say "call me at 2"? Either he has to do a lookup, or he has to internalize that Australia is a place where the workday is 23-to-7

Even now, work days (within a timezone) are not consistent. Most teachers here work from about 8am until 4pm, most office workers work 9am till 5pm, but some work from 9am till 5:30pm. I work 10am till 6pm. I know people who work 6am till 3pm. Lots of people do shiftwork and work completely inconsistent hours. These are just common times, in reality even in the same types of job, there are variations. Its not at all consistent and nobody can know when someone else is working without asking them.

And I certainly don't have to look up when I work (unless I work shifts and have to either way) - I already know when I have to be in work and when I get to leave.

And where I am, the workday would be 1-to-9, the east coast would be 4-to-13, Paris would be 10-to-6, and Bangalore would be 14-to-23. And this system made everything simpler... how?

This is already the case, except that the numbers are confusingly given the same names. What I mean is, even if everybody works 9-to-5 local time, your 9 is not the same as my 9 - you need to know the timezone and if its DST and then add or subtract a number to convert to either your or their local time. How is that easier?

And timezones are not easy. For example, I read someplace that if you happen to live in Israel/Palestine, what timezone your in depends not on location, but on if you identify as Israeli or Palestinian. Simple, right?


> How can Uncle Steve know he should say "call me at 2"?

Um, because that's when he gets off work? If you get up at 16 and go to bed at 8, you say "call me after 16 or before 8". Simples.

> And where I am, the workday would be 1-to-9, the east coast would be 4-to-13, Paris would be 10-to-6, and Bangalore would be 14-to-23. And this system made everything simpler... how?

If you want to schedule a conference call with two of them, you can immediately know what the overlap time is without having to google. (I assume you meant 10-to-18 for Paris)


>> How can Uncle Steve know he should say "call me at 2"?

> Um, because that's when he gets off work? If you get up at 16 and go to bed at 8, you say "call me after 16 or before 8". Simples.

How can you convince Uncle Steve (and the tens of millions of others in his timezone) to set their clocks to UTC, to start thinking of 1am as midday, 7am as dinnertime, and 10pm as the start of the workday? How can you convince these people to reprint opening hours, schoolbooks, TV schedules? How can you convince these millions of people that this is "Simples", when most of them do not interact with people in other timezones at all? Why would they do this?


Do you know by chance a list of such time zone islands?


There is a nice discussion of Australian ones in Wikipedia [1] - including a train line with its own time zone.

[1] https://en.wikipedia.org/wiki/Australian_Time#Anomalies


There is a great video from google where somebody explains that time zones are UI related things, and everywhere else UTC is the only timezone that you have to be concerned about. This means humans like timezones, because it is easy to understand

So can I call my sister in Sydney? edt(utc) -> 4:00AM, nope I can't call her. You can use any other timezone as input as well to get the actual time in an other timezone.

I think it is pretty easy to conceptualize for humans, this is why we have timezones. The suggested system does not feel right to me.


If you remember back to the early 2000s Swatch as a marketing gimmick proposed something called 'Swatch Internet Time' which was "designed" (and I use that word in the loosest possible sense) to solve this very problem. I think a couple of Ericsson handsets had support for it.

Wikipedia has the gory details for those interesting in that bit of trivia. I've used it as a programming exercise when interviewing candidates a couple of times.


You may or may not be surprised to know that PHP supports it natively.

    php5 -r 'echo idate("B");'
http://php.net/manual/en/function.idate.php


Because of course it does.


The last time I lost way too much sleep and felt funky from day light saving time I created this concept of a continuous day light savings time. What you need to do is pass in a gps location and from there it comes up with the current time as if you were the center of a time zone. It sets high noon according to the actual celestial noon.

It kind of goes back to the reason that time zones were created in the first place -- when trains came into popularity it became very hard to coordinate times between trains arriving from different towns that calibrated themselves to their sun dials. The time zones helped standardize the time so that you knew when the train would arrive.

This introduced towns that would wake up earlier and later than noon and introduced big problems in season variations.

So here's a silly demo: http://mike.rs/highnoon/


I thought of this very idea once as well. It is actually pretty cool to think about. The nice thing about it is the consistency of the time relative to the sun. No matter where you go the "relative time" would match up with the suns position (per the season).

But I also realized that when scheduling events you have to consider the time adjustment depending on the distance east-west the event will be from you. Since the Earth is approx. 25,000 miles around in 24 hrs you have to adjust plus or minus 3 to 4 seconds per mile of travel (if I haven't miscalculated). Of course, when coordinating events more than a few dozen miles away, one would surely use "absolute time" instead.

I think watches that had both would be awesome. You could identify your exact longitude by the difference between your relative time and the absolute time.


How does this help people coordinate? If your time is different from mine, how do we get to the same place at the same time?


You'd technically pin the date to UTC, then have all future and past dates automatically adapted to your localk time zone, where possible. For non-electronic methods of relaying dates, you'd have to use UTC.

Nome of this is solid, just what I thought of while sitting here reading your comment. Seems like it may be problematic but with some possible upsides.


That's one way. You could use "New York time" too. So if I have a meeting with someone in Ohio sf and nyc we can use a micro time zone related to a major city. If you were going to create a massive number of time zones, it would make sense to peg them to major cities. So nyc could have a time zone, that's slightly off from Boston and same with sf and la


It would be a hard sell...


Anyone entertaining the idea to abolish timezones has to realize that it is even harder to do, than to make the US convert to the Metric system, and that conversion is actually a good idea.

But it won't happen, because in the latter case you need to convince a few hundred million people to change their language, cultural references, and worldview, and they just don't want to do that. (The former case, obviosuly, means you have to convince a few billion people to do the same. Good luck.)


As someone who has suffered as Daylight Savings Time has blasted his programming kingdom, I'd give DST the boot first.


They're the same problem. If you abolish DST then time no longer corresponds to what time people are awake/working/etc., in which case you might as well get the benefits of doing away with timezones too.


That's assuming DST is natural, which I can assure is not the case in large parts of the world. It's a business/GDP-centric measure and nothing else.


True, but you have a much smaller "margin of error" (+/- 1 hour) as opposed to the many more inconveniences described in the article.


In this thread we learned: computers are no longer powerful enough to deal with the complexity of timezones for us so we should make humans do extra work.


The answer will be that your device shows you at a glance whether your uncle likely wants a call.

It does this mostly through automatic indicators…

• whether uncle has his device in 'do-not-disturb' mode

• how many hours it's been since uncle has woken or arrived-at-work or seen the local sunrise

• whether he's in a loud or quiet place, stationary or moving

• a live video feed of his public-facing workspace, or perhaps the view from his window

But it might also have indicators from subtle learned cues: the device senses if he's in a meeting, or currently in a spirited f2f conversation with others, or in the time/place when he usually engages in phone calls. So if it thinks "now is not good", it can also predict and show the likely next workable time.

In fact, you might even just press a button (or vocalize a command) telling your device, "let me know when it's a good time to chat with uncle". It'll negotiate a time with uncle's device, only bothering uncle for decisions if/when it's locally appropriate. Then it will offer to connect the call only when it's almost certain to be welcomed by both ends.

All these technologies are likely to arrive, even without a single global time. But once they arrive it'll be easy for more people to prefer UTC. And it'll seem silly that anyone ever needed to write 2,200 words about how hard single-timezone coordination would be, because for people living with such assistive technology, that coordination will be effortless.

Sun-locked time-zones were a useful transitional coordination technology, but cheap computing and telecommunication will offer much better options.


I'd like to share my favourite timezone lookup tool: http://xkcd.com/now


Oh how I wish it was interactive.


How do you mean? It does stay in sync with the current time.


I'm not sure what this article is about. At a minimum it is misguided. I don't think anyone is seriously arguing for abolishing time zones. Also, while zoneinfo is great, it doesn't help with the more serious issues that arise from arcane/archaic timekeeping.

There are several major, common issues with timekeeping that cause huge engineering headaches:

- Frequent or arbitrary changes to time zone standards

- Daylight savings time

- Time zones with fractional hour offsets

- Frequent or arbitrary adjustments to UTC (universal time) in an effort to keep it more astronomically correct

The good things I can imagine happening to timekeeping are:

- Abolish daylight savings time

- Separate universal terrestrial time from "astronomically correct" time, with infrequent synchronizations fixed in time (e.g. once every decade or century)

- Abolish fractional hour offset time zones and encourage countries to consolidate time zones


Here is an article by a smart person on Vox arguing precisely for the abolishment of Time Zones: http://www.vox.com/2014/8/5/5970767/case-against-time-zones


EDIT: I see, you were providing an example of a smart person, not an example of a smart argument. I'll let my rant stand anyway. :-)

No, it's an idiotic argument, because it essentially boils down to forcing the billions of people who do not live in the GMT timezone to adjust their mental model of "what time is office hours", and you STILL haven't solved the original problem of answering the question "Hey, you are in a different timezone, is this a good time to schedule a meeting?" You STILL need to do a lookup!


I wonder why so many commenters seem to think that abolishing time zones requires a totalitarian government. The United States or Canada could do exactly the same thing, if they had the political will to do it.... afaik there is nothing in either constitutions enshrining time zones.


Yeah, it's really easy to do, you just tell everyone living on the west coast of the US that the sun rises around 11pm, work hours are now 1am to 9am, midday is at 4am, dinner is at 10am, bedtime is at 2pm and midnight is at 4pm.

If you think people are going to accept that voluntarily, you are insane.


Different meaning of 'time'.

Why can't we have two scales of 'time', one for marking what we'd now call 'absolute time' (an arbitrary clock that is the same all over, E.G. UTC) and another for 'time of day' (in a given location); they should also be displayed uniquely so that it's not possible to confuse one with the other.

Let's pick an arbitrary time. As a high tech society no longer ruled by sun dials we should instead make the start of the scale 'dawn' (on average). 'noon' would be at 25% of that scale, 'sunset' at 50%, midnight at 75% and 'dawn' again would be 100%/0% where it loops.

I think 00 Monday might be good shorthand for 'dawn' Monday, though most would still say 'dawn Monday', however 10 Monday would be a good marker for when you might want to meet someone for your daily dose of caffeine. 40 through 60 might be when you'd prefer to eat dinner.


Unless I'm misunderstanding you, your proposed time scale would be difficult to implement. Different areas have greatly varying amounts of daylight.

>As a high tech society no longer ruled by sun dials we should instead make the start of the scale 'dawn' (on average).

Aren't you basically proposing a (slightly modified) sundial?


Doesn't Julian time start at dawn or the day start then and end after sundown? It's something weird but I'm not sure even after looking it up.


Almost. Julian days start and end at noon GMT.


I often like to test new concepts by applying them to (realistic) extremes, and seeing how they play out.

So take a small town, high in the Arctic Circle, which experiences 30 minutes of darkness around the Summer Solstice. 00 Monday (June 21) would correlate roughly with 12.15am; 50 Monday (sunset) would be 11.45pm. 75 Monday would come 15 minutes after sunset; though 25 Monday ('noon') would be 12pm midday, 10 Monday would actually be 4.56am. In fairness, when I'm up then I do like my caffeine.

Fast forward to December 21 (Winter solstice). The town now has 4 hours of sunlight, roughly corresponding so 00 Monday is 10am and 50 Monday is 2pm. 75 Monday is midnight, but there are now 10 hours between 50 and 75 where there used to be 15 minutes. 10 Monday equates with 10.48am, a fine time for a coffee meeting (even though it's my third meeting of the day), but bearing little resemblence to 4.56am except they match the same proportional distance between dawn and noon.

Thise living in Singapore near the Equator can't understand the fuss, since barely nothing has changed.


Nit picking, I know; when above the Arctic circle, the sun is up for 24 hours at the summer solstice and never rises at the winter solstice.

I think you mean to say "just below the Arctic Circle". Some place like Umeå, Sweden, perhaps. (Even then, it's not really "dark" when the sun dips just below the northern horizon.) Yeah, a time system based on sunrise/set really isn't going to work in the high latitudes.

And if we're talking Sweden, "daily dose of caffeine"? Is that breakfast, morning coffee break, end of lunch coffee, afternoon coffee break or after work coffee? :)


Good addition. Yes, I avoided the 'midnight sun / sun never sets' extreme because it actually kills the 00 = Dawn in one fell swoop. But I didn't realise how low a latitude that came (I did experience 'midnight sun' holidaying in Bergen, Norway, but even then it actually set a little later in the night.)


> As a high tech society no longer ruled by sun dials we should instead make the start of the scale 'dawn' (on average).

The day starting at dawn wasn't an uncommon rule with sundials. Its kind of weird, I'd say, that your idea boils down to "we aren't ruled by sun dials, so we should govern time by the rising of the sun". What?


Waiting for the follow up, "So you want to abolish Daylight Savings Time".

Because that one I actually do.


Well.. here we have daylight saving time seven months a year, which really should mean that DST is 'normal time'.

That also mean that five months every year we have Daylight Losing Time. I want to abolish that.


Time zones make sense, daylight savings time does not.


I know a lot of people are opposed to daylight saving but I actually like them.

Waking up 2 or 3 hours before the sun rises in winter, or 2 or 3 hours after the sun rises in summer isn't very natural. Daylight savings reduce the gap by one hour.


True, but couldn't the same be accomplished by simply changing your alarm clock?


I think most people who say that we should eliminate time zones are confusing time zones with daylight saving time changes. Eliminating DST shifts is a fantastic idea. Eliminating time zones is pure stupidity.


We can't even ditch the terrible idea of leap seconds. Which in practice, are worse than DST, because you can just ignore them like you can ignore time zones.


I'm a technical person. I also hate time zones, month names and weekdays. Those are utterly meaningless. YYYYMMDDHHMMSS (UTC only) is the format I'm using with all projects. It's just compacted form of 2015-01-17T07:18:16Z aka ISO_8601. When working with globally distributed teams and comparing event logs etc. One single time indicator is the way to go. Alternate format is classical UNIX time which is nice and can be stored as INT if second resolution is enough. 1421479099 but isn't so well human readable. Yet it's trivial to convert of course.


Why compact it? I can see that it saves some bytes but it introduces ambiguity.


> Alternate format is classical UNIX time which is nice and can be stored as INT if second resolution is enough.

What do you do about leap seconds?


I feel like I'm losing my mind reading this comment thread. How can there be any people that think abolishing timezones is a good idea? The article covers the obvious problems pretty well.


I'm with you. It's mindbogglingly stupid.

Here's another example on why it doesn't help:

I live in San Francisco. I travel to South Korea for work, I check in at the hotel, I'm jetlagged, I fall asleep, and wake up some time later. It's dark outside. I'm very tired, but I can't be late for work. How do I know if I can go back to sleep, or have to get up?

In a world without timezones: My phone says it's 7pm. I don't know what that means in Korea, in SF it means it's time for lunch. I either have to look it up, or have it internalized that in Korea it means it's in the middle of the night.

In a world with timezones: My phone has already switched over to local time. It says it's 4am. I go back to sleep.


This is mostly nonsense.

The am pm thing isn't everywhere, in Europe they often say "19 o'clock" so that's just an accuracy issue (honestly it's a made up problem since am pm will still be used).

That Monday/Tuesday thing is also made up since there are offices, bars and clubs with overnight opening hours and they still separate the days, made up problem.

Lastly, the entire issue one has with not knowing if the offices are open in Australia is still happening today, and we mostly solve it via Google, like we can solve this one too


I'm in Europe (United Kingdom, not mainland Europe) and I've never heard anyone say 19 o'clock, that doesn't mean it doesn't happen, I've just never heard it and I'd find it quite strange if I did.

Schedules are usually made of context: Shall we grab a drink at eight? (Nobody drinks at eight in the morning, so it must mean the evening.) – Shall we have a meeting at two? (I'm usually asleep at two in the morning, so it must mean the afternoon.) – it's rare that I experience ambiguity in times.


Clearly the solution is to not call people. Synchronous communication sucks.

I do like the time shifting to maximize daylight in the high latitudes.


People use 24h where I live (Poland) and it's much better. I never remember if am in English is morning or evening and I have to check each time I need it (it's a few times a year so the next time I don't remember again).

Please, USA, quit making easy things difficult, switch to SI, put dates in correct order and move to 24h time.


I'm thinking that clocks to time would be more like GPS to maps. You could just simply ask what position is the sun in a particular location, at a fingers touch. So it would be too much of a big deal for most people. But it would make programmers and things that deal with logical time much easier.


> But it would make programmers and things that deal with logical time much easier.

But as the article points out, programmers and others who want to work with time without the hassles of timezones already have such a system (UTC).


Only sort of. The abolishment of Daylight Savings would help a great deal in this area, though (because it makes writing time-based code so much harder).


Why only sort of? What could be easier than than a time reference that's constant across the planet, doesn't observe daylight savings, has no leap years (but rare leap seconds, iirc), and can be trivially converted to a local time if necessary? I guess never having to convert to local times because there's no such thing in the world as time zones would be easier, but at the cost of all of the practical everyday uses of local times.


That was a very fun dissection of an idea I'd thought that nobody was foolish enough to actually suggest. Apparently I was wrong, so wrong that enough people had to ask the author about it enough times to provoke this wonderful essay. I wish being wrong was always like this.


Having the sun at its highest point as close to noon as possible is a human right.


I only read a bit of that I drifted off recalling when I first started using the Internet in the early 1990s and started communicating with people around the world.

I'm from a small town and never travel so it was a real eye opener for me. Of course I knew about the world, time zones and all that jazz I was excellent in geography and live maps, the world, cultures.

What I quickly realized is it's nearly impossible to have any relationship with a person who doesn't isn't in your time zone. Even e-mail is awkward since you know it's going off to someone who won't see it until they are awake, have time to read it, want to respond or may wait a few days.

It's like another world, I'd hate to see what happens when we discover intelligent life in a place where light takes years to reach.


This really isn't true. I know it sounds trite, but if you love somebody, you'll find a way to make it work, and it won't seem a chore. That isn't to say long-distance relationships aren't challenging, but small issues like time zone differences (even large ones) just melt away.


Can anecdotally confirm. I was once in a long-distance relationship crossing a few time zones, where we also had different work schedules. We basically lost contact because we could only sync when I got out of work and my partner would be about to go to bed.


If we could instead abolish leap seconds, DST, time zones that are a fraction of one hour and forbid any country from changing timezone we would be much better off than abolishing timezones.


Time zones are ok, I'd much rather abolish daylight savings time.


This entire problem seems to be solved by using asynchronous communication. Bam.


I feel like the author completely ignores the simplest answer to wether or not it's ok to call their uncle. Just use the exact same process they would use now. Instead of "It's that time for him" he would think "It's the time of day for him that corresponds to that time for me" but that's all that would change.


So... "use what are effectively time zones, just not labeled or standardized in any way"?

That's a pretty poor solution.


I think that's his point: if the solution is "a table of timezones to help answer my question", then we already have that.


Exactly. His point is that people who argue for abolishing time zones are playing a classic shell game. They're talking up the benefits of the new system while ignoring its problems (and the benefits of the old system).

This kind of shell game is pretty popular in software development, where new solutions are hyped because they lack some of the problems of the old solution, but also lack many of the benefits of the old solution.


If the only problem with abolishing time zones is having to look up a table to know when to call someone, then perhaps it's better to abolish them. Most of the uses we make of international times don't have anything to do with timing phone calls.

To make a decision on what's best obviously needs a comprehensive look at all the pros and cons, not just one word-gamey one like this.

Perhaps the current set of time zones are not optimal. Maybe the best solution turns out to be just rearranging them. I would hope making them wider but perhaps there are biological reasons to refine them even further and make daylight savings even more complicated. If that somehow extends out lives, makes us more productive, or saves enough power (money), it might be worth the added cost to programmers (money).


Obviously, that lookup table is not the only problem with abolishing time zones, the giant problem is forcing billions of people to change their notion of when "9 to 5" happens, as well as every other activity that is pinned to a specific time through custom, culture, and language.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: