My five-year experiment with UTC
timestripe.com
timestripe.com
Why? Because it is confusing if most people’s natural days are divided into different calendar days. If I wake up and the calendar says Tuesday, then I can be sure that I don’t need to worry about the dental appointment on Wednesday afternoon. But if Wednesday starts midday, then I have to spend extra brain cycles to work out which parts of my natural day belong to Tuesday and which parts are Wednesday. It’s a lot of hassle avoided by having days change over when I’m asleep.
If you want real life example of such confusions, look up “early morning flight confusion” on your favourite search engine. It turns out a lot of people are confused by flight times like “Wednesday 1:20am”, because if you are catching an 1:20am flight, you are heading to the airport on the previous day before midnight, and mentally the flight feels like a part of the previous “natural” day. What sometimes happens is people will book a flight on Wednesday 1:20am, head to the airport on Wednesday evening because they think “well the flight is on Wednesday right”, and find out that it is actually Thursday after midnight.
By the way, one successful attempt at addressing this problem is Japanese late night anime schedules. If an anime is aired on Wednesday 1:20am, the TV station will instead write “Tuesday 25:20” on the schedule. It makes no sense from a technical point of view, but feels right for the human because what is the early morning of Wednesday if not just Tuesday night prolonged?
[1] Even extreme cases like Spain are only a couple of hours out of sync, it’s not like the sun shines on Madrid during “midnight”.
[2] the article mentions remote workers and frequent travelers, which I am willing to wager are a tiny (but over represented in nerd spaces) fraction of the society.
It likely does not generalize at all.
They also haven't faced an issue where an event is booked by someone else at a DST-affecting time: as that time changes between two days, you now need an extra "if" in your brain (is it DST there? Europe and US are on different cycles there too).
I've tried booking all the cross-timezone meetings in UTC, but didn't get far with it: it was more work for everyone compared to somewhat reduced confusion and suddenly overlapping meetings.
If you use 24-hour ("military") time—00:00 to 23:59—this becomes less confusing. Perhaps an attempt to culturally shift from using 12-hour to using 24-hour time would be useful. (In the US it'd probably less of a challenge than metrification.)
[1] https://en.wikipedia.org/wiki/Date_and_time_representation_b...
[2] https://en.wikipedia.org/wiki/12-hour_clock#Use_by_country
Depending on the channel, the "new date cut-off" (more accurately: 日付変更) sometimes varies. Some channels start the day at 5am, others at 4am, and so on. Because of this, one gets to see wild things, such as a time-table scheduling a show at "29:30" whereas another channel would write "05:30".
In any case, your claim is correct. The people who stay up to watch late-night shows do not consider their day finished around those hours, and the times are intended to reflect that. However, there is still inconsistency around the hour of the new day.
A similar issue exists in video games. I know of one game that "starts the new day", so to speak, at 5am, and another game that does so at exactly 12am. To add further confusion, many Japanese websites and games are designed with the assumption that the user is always in Japanese Standard Time. (Most Japanese do not travel abroad, and Japan itself has exactly one time zone, thus Japanese programmers rarely ever think about time zones). Thus, when overseas, you may see the new day in a video game start at 1pm local time!
This is what really grinds my gears with the official definition of "night" and "morning".
Tuesday 21:00-23:59 is considered "night", and then Wednesday 00:00 is considered "morning", which is highway insanity to me. No one would say they were going out 'till "Wednesday morning" if they went to the pub until 02:30 Wednesday.
IMO it only starts being "morning" around 05:00, maybe even 06:00.
Even now, obtaining solar noon is comparatively easy. Then add half of an earth's rotation* (slightly affected by the earth's orbital period too), and that's midnight.
In the days of candle light, midnight wasn't a crazy time to draw that diving line. However these days even people with crazy early AM weekdays to avoid commutes sometimes cross over midnights on the weekend.
+
You are correct that the modern world really should just start counting the day at like 3 or 4 am, and continue counting past 2359 to almost 2700 or 2800.
I guess I am arguing that "morning" "midday" "evening" and "night" should hold no official definition, since they are much more about the vibe rather than an absolute position on the clock. However, if they do remain as official definitions, they should be heavily revised.
"I woke up very early in the morning at 4am to start work at 5."
For instance, we handle 24 hour clock (with 60 minute hours?! WTF?) fine, even if it's insane and we use the decimal system elsewhere. Because we got used to it when we grew up. Many even have the clock reset in the middle of the day, ending up with things like two hour activities spanning from eleven to one! (Don't get me started on shit like "half four" - it means different things depending on where you are, but it never means two!)
Heck, there's a whole, huge country that gets by on inches and ounces and pounds and miles and whatnot - a mind boggling amount of extra mental energy required, and with the exception of some recent lapse when selecting their boss dude they seem to be doing mostly as well as (and over the previous century arguably somewhat better than) the rest of us in the thinking department.
If we grew up with date/day changing in the afternoon, every kid in preschool would have it figured out already. "Early morning flight confusion" would seize to exist (except, perhaps, for the poor bastards who'd happen to live close to wherever we decided to place the prime meridian), because everybody would know what "1:20" meant, and not randomly guess at something based on when they happen to sleep.
If you're used to date changes happening "whenever", you'd stop assuming they happen at a fixed point during the day.
Any argument that starts with "Today, we're already confused about ... because we're used to date changes occurring while we sleep" is meaningless, since we wouldn't be used to date changes occurring while we sleep.
The whole terminology thing would be solved by having two words; one meaning "after next sleep cycle" (which would probably be more common when speaking to people in your own geographic region), and one meaning "after next date change" (which would probably be more common when speaking to people in other geographic regions). The appearance of this second word would likely happen naturally within weeks, if not days after abolishing time zones. As usually happens with language.
The argument against abolishing time zones always seems to assume that humanity would be incapable of adapting to a time zone free world. Like this would mysteriously be the first and only change in the history of mankind society couldn't adapt to.
But that's not the case. We're not all cosmopolitans traveling the world all the time. For anyone living in a timezone where 00:00 happens when most people are asleep, there'd only be one word for tomorrow, as it'd be synonymous for "after date change" and "after the next sleep cycle". There would be no other word, because there is no need for that. Except in circles where coordinating meetings with people across the globe happens on a regular basis. But we already got that figured out by explicitly naming the time zone or using UTC.
The vast majority of people don't have to deal with this, so having the date change at night is just more convenient, as it completely avoids this ambiguity in day to day life.
But obviously as a nerd you get hung up on that one aspect and go "but imagine we had only one timezone, we could just say 13:00 and there'd be no ambiguity, no chance for a misunderstanding, it would be so much more efficient and coherent!"
> The argument against abolishing time zones always seems to assume that humanity would be incapable of adapting to a time zone free world.
Again, you need to take a step back and try to look at this from the average person's pov, not as a scientist trying to find the optimal solution for a problem 99% of people don't have. And "people would be able to adapt to this" Is a pretty weak argument. You could probably also change the clock itself so that there's 63 seconds in a minute, 59 minutes in an hour, and 195 hours in a day. I'm sure people could adapt to this. Doesn't make it a good argument.
On the other hand, just imagine the sheer chaos this switchover would cause worldwide, for what, marginal gains in convenience for a miniscule amount of people.
It reminds me of "that guy" who constantly wants to refactor half the codebase because something would look a little cleaner then, and if you let them, they'll end up refactoring the same part of the code over and over again every six months because they had another idea how to make it even more streamlined and elegant. And then someone else comes in wanting to implement a simple feature, but is greeted with an incomprehensible construct, while the refactoring-guy is repeatedly and excitedly reassuring them that it's "really simple and intuitive once you get the hang of it".
Meanwhile, I know what the word desert means, even if I have never been in a country that has one. I know what a cyclone is, even if I've never seen one. We're affected by, and have knowledge of more than our local environment. We read books. We watch TV and movies. Heck, we're having this discussion in a language that that isn't spoken in the country I'm living in. Most people do, in fact, interact in some way with people in other time zones right now. Perhaps not on a daily basis, but at least a bit.
It just occurred to me; if we placed the meridian in the Pacific, very few people would have the date change during their sleep cycle. Sucks to be Niue, I suppose ;)
By the way - I'm not even for abolishing time zones. I'm just spectacularly tired of the straw man arguments against it. It's like saying "we couldn't build a city that's on both sides of the river, because people would drown when they go to work", as if building a bridge would be an impossibility.
There is a perfectly good argument against it btw, which is the one you propose at the end. The change itself wouldn't really be worth it. But that's simply because people are confused by changes. Not because the end result would be worse.
So which is it?
Anyway it seems to me this whole problem arises with 12 hour notation.
24hr time, folks. 24hr time.
This bit here, I don't understand what Thursday is supposed to do here. If my flight leaves Wednesday at 1:20 AM, I'm leaving for the airport Tuesday evening.
But if I'm interpreting OP correctly, people are booking flights for Wednesday 1:20 AM but leaving for the airport Wednesday evening? Why would you be confused about that?
So if you happen to think of it as "middle of the night, specifically 1am", and you fall into mental shortcuts a little bit too hard, oops you might end up there on the wrong day.
By the way, if you’re used to 12h time, you can get a similar experience to that described in the article simply by setting your watch/phone to 24h time. After a while your brain just starts to recognize 15:00 and 3pm as the same thing, and there’s no explicit conversion required for you.
Tuesday 25:20 has a much better chance of helping.
It also does not help for entry into a programmatic form or database field. That sort of notation is only good for output to a human who knows this custom.
But the important part is that 01:20 does not solve the problem.
There is no perfect solution, so I can't tell you what solving it would look like. But I'd say that putting "(Tuesday night)" next to where it says "01:20 Wednesday" would be helpful to many of the people making this mistake.
How would you define the problem? It's not people mixing up 1:20am and 1:20pm.
But that's subjective. Even my grandmother in the early-80s said differently. Everyone's computer-time and our smartphones and our digital clocks we've plugged into the wall, all have switched over to Wednesday at 00:00:00 midnight.
If you go to a bar they will tell you, "last call is at two in the morning." If someone heard a noise they say "I woke up at three in the morning to look around." My mother called them "the wee hours".
Sure, colloquially, 01:20 is very late on Tuesday night, because the Sun hasn't returned to announce Wednesday, and the cock hasn't crowed. So that is another way of looking at it. Time is analog, the orbits of Earth and Moon are analog, and we're cramming them into digital approximations and binary categories.
But you've misidentified the problem. The problem is that different humans interpret differently. The interpretation is influenced by culture and language. My scientific culture says that 01:20 is on Wednesday morning, and that is in agreement with my dearly departed grandmother, my smartphone, my calendars on Google and Outlook, and honestly there is no outlier in my life, except you, and Crunchyroll(?) who would insist that "01:20 is actually 25:20 on Tuesday night", and that writing it in this way would "help" solve a problem, because we're just explaining how damned confusing that is to rational Americans or users of either 12- or 24-hour clocks.
So "the problem" in my book is that a hypothetical human looked at our hypothetical "01:20 Wednesday" as rendered by a hypothetical computer, and that human wasn't thinking like a computer, or an American, or that human was Japanese, and that human erred with their interpretation of time notation. Now I don't have any idea of what time notation "01:20 Wednesday" would prompt a hypothetical trip to the airport 24 hours late, so it is not like this human was applying a different but correct interpretation to this problem.
Why don't we write "very early on Wednesday morning" or "before sunrise on Wednesday" or "just after midnight" (the latter is literally what "a.m." means...) and it won't look like another typo. Because if you write "1:20am on Wednesday is Tuesday night" then I'm going to call out a contradiction.
Computers can actually help with this, because if I've made a hypothetical calendar entry for my flight and I email the .ICS file to you, then the timezone and day-of-week is baked-in to that, so if everyone's settings are correct, bypass human interpretations. Many airlines have a way to email your itinerary directly to your friends. Just use the tools.
In the vast majority of cases, the most salient property of a timestamp, for humans, is what point it occurred in the 24-hour day/night cycle (and, secondarily, where it occurred in the 7-day week cycle).
Local time is an approximation of exactly that 24-hour cycle; the local time anywhere is approximately the elapsed time since the last solar midnight in that location. (Yes, it's not perfect, because in some places local time has a persistent offset from solar time, or even one that changes twice a year, but it's close enough).
UTC is an approximation of the elapsed time since the last solar midnight in England. Most people don't live in England, so this just isn't relevant most of the time.
If we all used UTC, sure, people living in one place would get used to the new correspondence to solar time; someone who spends all their time in Arizona would quickly get used to the fact that they now eat dinner at 12:00 UTC instead of 7pm. But it would make traveling more tedious: you'd have to re-learn the mapping every new place you went. It'd also make communicating with people abroad more difficult. If someone tells you "I went to bed at 08:00 last night", you have to know that in New Zealand that means they're an early riser or were sick; in New York it means they had a wild party and in Poland it means they worked the night shift.
There are a few cases where having a shared absolute time reference is useful; for example, scheduling meetings with people in many different countries, but in those cases people tend to spontaneously settle on a time standard and it doesn't cause many problems in practice. And even in many of those cases, the local solar time is still relevant (you wouldn't want to schedule a meeting for someone around their solar midnight), so you have to have some way of expressing it anyway.
The 7-day week originated in ancient Mesopotamia, particularly Babylon, around the 3rd millennium BCE. It was tied to the observation of seven celestial bodies visible to the naked eye: the Sun, Moon, Mercury, Venus, Mars, Jupiter, and Saturn. Each day was associated with one of these "planets," a system later adopted and refined by other cultures.
The Babylonians passed this concept to the Jews, who formalized it in their religious practices, linking it to the biblical creation story in Genesis, where God created the world in six days and rested on the seventh, the Sabbath. This Jewish tradition influenced the structure of the week in the Western world.
The Romans also adopted the 7-day cycle by the 1st century CE, replacing their earlier 8-day market cycle (nundinae). They named the days after their gods, aligned with the celestial bodies: dies Solis (Sun), dies Lunae (Moon), dies Martis (Mars), dies Mercurii (Mercury), dies Iovis (Jupiter), dies Veneris (Venus), and dies Saturni (Saturn). These names persist in Romance languages (e.g., French: lundi for Moon, mardi for Mars).
The 7-day week spread through the Roman Empire and was cemented by Christianity, especially after Emperor Constantine made Sunday a day of rest in 321 CE. Other cultures, like the ancient Egyptians and early Chinese, had different cycles, but the 7-day week became dominant globally due to Western influence, trade, and colonization.
It’s not perfectly aligned with natural cycles—365.25 days in a year don’t divide evenly by 7—but its cultural and religious roots have made it a near-universal standard.
I've known multiple people in my life whose working shifts do not correspond to the seven-day week.
I know of no illness from which they suffer.
No other primate cares a whit. No other animal cares a whit.
Citations would be very useful here, otherwise it sounds like another random thing from minds unknown.
Following local times won't spare you from having to learn local culture anyway, such as Spanish dinner times ranging around 21-23.
If you visit Spain as a tourist and eat at 19:00 local time every day, you’ll be slightly out of sync with the locals but you aren’t going to cause any huge drama. If you never even learned the fact that Spanish people eat later, you’d still basically be fine. So knowing that cultural fact about Spain isn’t absolutely essential the way it would be to learn the offset of each place in a world without time zones.
> As a programmer, I’ve always been annoyed by the concept of administrative time zones.
> Five years ago, I decided time zones should be abolished, and everyone should use one coordinated time.
If this person was vandalizing cultural treasures or something to pressure everyone to use UTC, then yes I agree his forced technocratic fix to a social issue is bad. Just writing about it and sharing an opinion is actually very constructive, even if you disagree with his take.
This wasn’t phrased like “I prefer to use UTC time and here’s why I like it,” it’s more like if this guy becomes king of the world we’ll literally be forced to used UTC time.
This is why the person you replied to said:
> it would be great if we could figure out a way to prevent it, and teach that in schools. Too many people like this in the world today.
Too many people think that constructively sharing a personal opinion and sharing a desire to impose restrictions on others are of equal merit.
I must have missed that quote, I seem to be unable to find that text in the post.
> This wasn’t phrased like “I prefer to use UTC time and here’s why I like it,”
I don't know, it kind of seems like it was:
> Here’s Adam’s story of how living by UTC transformed his productivity, and why it might work for you, too.
and
> Give it a try, you might find it as liberating as I do.
They're making it sound awfully optional if their intent was to put you in jail if you didn't do it.
“Five years ago, I decided time zones should be abolished, and everyone should use one coordinated time.”
That quote is also in the article.
> “it should be abolished and everyone should do what I say.”
The original and the (mis)quote are not the same thing at all.
The existing system is more inertial and political than practical.
"Noon is when the sun is at its zenith in your area" is significantly less useful in 2025 than being able to coordinate globally with ease. It's also just not true in many (most?) cases, as noted in the article, where places like China have a single timezone even though there is a 3-hour discrepancy in the sun's position from one side of China to the other.
3 hours is one thing, 12 hours is another (and "So you want to abolish timezones" notes that people in China commonly use lookup tables, creating implicit timezones).
When I coordinate across timezones, I say, "Hey does 9amPT/11amET work?" (Assuming those are the relevant timezones). I don't see the problem with that approach?
Let's say you abolish timezones. Okay. How do I know whether it's rude to call Uncle Steve in Melbourne? I'm going to consult some kind of table and implicitly reintroduce timezones, right?
Conversely learning when your Australian co-worker or uncle gets off work (in UTC) must only be done once.
e.g. You learn that Uncle Steve in Melbourne gets off work around 07:00 UTC. So a reasonable time to call him and chat about life is 08:00 UTC. You text him and say "hey, long time no chat, want to catch up at 08:00?". He doesn't have to look anything up because he is also using UTC (or whatever). He says "actually I have a dinner thing, what about 11:00 instead?". Know confusion can be created, because this is the same time frame of reference for you too.
People keeping different hours in different locales is inherent complexity. This is the crux of it. If you want to abolish timezones, you need to explain why it isn't a problem that people keep different hours in different locales. If you can't, then it boils down to timezones being annoying. They are annoying, but that's not sufficient reason to eliminate them.
Your solution doesn’t change that I have to figure out some “how many hours different is Uncle Steve”, just that somehow I’ll magically figure that out without using the simple UTC offsets we use to denote those differences now. It just papers over them.
Here is another example: I schedule a meeting with my co-worker in London. I say "how about 10am your time?" This is because I know he is X hours behind my timezone because I've learned that fact. He agrees and I show up an hour late because I forgot about DST and that he is only X-1 hours behind now.
If we both coordinated around UTC, he would've said "actually 11 is better for me" and I would've showed up an hour earlier.
It reads as tongue-in-cheek to me :)
> Give it a try, you might find it as liberating as I do.
Did you miss these quotes, or are they also part of the absolutism? Here are some ways I read it, which seem like fair interpretations and also don't make the author stand out as being particularly dogmatic or authoritarian:
Considering the quotes I provided, it seems likely that the author may have been suggesting that we stop using timezones as a society in order to mitigate the only two downsides that they listed, since those downsides only exist as a result of society still using timezones while they as an individual stopped using them.
Another way to read it could be that the author made a statement they believe is probably true ("timezones should be abolished") but they were not fully convinced. To gain confidence, they decided to gather further information: "I began with myself...To my surprise, it was easy" and then ended with two downsides of their experiment, suggesting that they realize it may not be a perfect solution, at least right now.
A third way one could read it is to consider that the author didn't propose fines, jail-time, or any other specific penalties for continued use of timezones in their hypothetical world. Without enforcing a penalty, what does it even mean to "abolish" a timezone? In the context of the whole post, it seems to be more of a thought experiment imagining what could be if society as a whole switched, rather than a prescription of what must be.
The unwillingness to engage with the author's thought experiment strikes me as more dogmatic than anything the author wrote, but perhaps I am not fully considering other possible interpretations.
I think folks are averse to "world time" (for reasons, largely inertial), so maybe the baby step is try 1 timezone per country (like China's done for .. 75 years ?).
I'd even argue Local Time has only ~4 useful times: dawn, daytime, dusk, & nighttime. Where I grew up, the parks closed at dusk every day. Nobody complained
If I take a walk at 7:05:35 PM every day, it seems very precise but doesn't indicate whether I need sunglasses or a flashlight. It's meaningless precision, like 0.6235 slices of pizza. If I'm coordinating a walk with you, I might as well use UTC: it still won't tell light from dark, but at least nobody'll be waiting for an hour due to DST. It'd make more sense to schedule our walk at `1 hour before dusk`, or "just" settle for UTC, IMO.
China did that for political reasons. The Republic of China used five, and Russia uses two timezones across those longitudes. Similarly, I also find it quite weird that France and Spain are in CET. France shifted when it was occupied in WWII, but maybe it's justified to remain in CET to reduce friction with the economies of its eastern neighbors. Whatever. Vive la weird. Countries do that to themselves.
A deviation of half an hour from noon is barely perceptible unless you use a sunclock. People can roughly estimate when the sun will set by just looking at local time and considering the current season. What throws a wrench in the works is daylight saving time. I fully agree that DST a huge annoyance and that its benefits were always rather situational.
So You Want To Abolish Time Zones - https://qntm.org/abolish
Being able to express where you are in the 24-hour cycle is useful. That's what local time is: an approximation of the time since the last solar midnight.
To the extent that the approximation is poor, it's an argument for improving it (Spain for example should really set their clocks back an hour), not for getting rid of it entirely.
Besides, Kelvin is the most true temperature scale, not Celsius.
I find this considerably more useful than the Fahrenheit equivalent.
That way you have the "human livable" range kinda between 0-100, which feels very intuitive. Anything above 100ºC becomes effectively unlivable. It also means that it is much easier to distinguish between certain "zones". e.g. saying "70s" or "80s" is easier and more clear than with celsius where you typically are staying within a sliding 10º range from day-to-day.
But why? Knowing the freezing point of water (at sea level, mind you) is only rarely useful to me in daily life. I think the eutectic point of salt and water (that is, the lowest point at which it's possible to cause ice on the roads to melt by adding salt to them) is marginally more useful; this happens to be about -6F or about -21 C. Perhaps it's a bit more important for farmers who need to know whether their crops will freeze overnight, but as a lifelong city/suburbs dweller who knows next to nothing about agriculture I couldn't really say. At any rate, most people are not farmers.
I'm asking this somewhat rhetorically: I don't really care about Celsius vs. Fahrenheit. The pros and cons between the two are so vastly outweighed by the utility of having a global standard that I really wish the US would switch. I'm just pointing out that I don't think that one particular physical property of one particular substance makes for a very strong argument either way.
Or: "below 0C my plants might die"
Or: "below 0C the pond will start to freeze over"
the 60's are warm in the sun and cool in the shade or breeze, the 70's are warm, the 80's are hot and the 90's are really hot, and over a hundred you're looking at death. The 50's you need a jacket, the 40's you need more jacket, the 30's are downright cold and you need an overcoat, and the 20's are frigid.
Celsius has approximately none of those properties. Celsius the 20's encompasses a vast change in comfort level.
https://xkcd.com/1982/ Had it entirely right.
- ex Celsius user
Unlike DST today, time zones do still have economic utility in the modern world since there are many more personal and industrial activities that exclusively happen during day or night in a single locale and far fewer that happen 24x7 and routinely traverse multiple locales.
https://en.wikipedia.org/wiki/Swatch_Internet_Time
Nothing is more rad than checking the dot-beat on my translucent plastic watch while waiting for dial-up to connect so I can update my ICQ status to let my global friends know I’m listening to “Barbie Girl” from a 128k MP3 in WinAMP. Pre-emptive multitasking baby.
For some reason no friends or family took me seriously when I brought it up at the time.
Yes, one universal time zone, used universally! Bravo, finally, indeed.
But with metric time:
• 1 metric hour = day/25
• 1 metric time minute = day/1000 = hour/40 = 1 mday
• 1 metric time second = metric minute/1000 = 1 µday
And while we are at it, since this is all about rotation, lets fix that too:
• 1 metric rotation degree = rotation/100 = 1 centi-rotation
• 1 metric rotation minute = rotation/1000 = 1 milli-rotation
• 1 metric rotation second = metric degree minute/1000 = 1 µrot
Notation: instead of '°', metric degrees use a filled circle '•'.
Now we have conceptual correspondence between 1 minute of time and 1 minute rotation of the Earth.
If you know your metric longitude, you know the local time shift, and vice versa.
Finally, a metric "milette" distance is defined as 1/1000 of the Earths circumference, or ~40km. The Earth's surface moves 1 millete distance -> in 1 minute of time -> over 1 minute of rotation
Ok, open for suggestion period.
I've also thought that setting time using longitude could make sense. Especially since I and many people tell time, schedule meetings, etc using a device with a GPS. This article [0] makes an interesting point about the effect that time shift based on longitude would have on computers in the same data center.
> At the equator, the position directly underneath the mean Sun travels west at about 463 metres per second. That means a standard rack unit is about one millisecond wide. ... So, strictly speaking, continuous time zones mean that clocks on machines in different parts of the same data centre — neighbouring racks, even — will need to be set to different times, depending on the exact positions of those racks.
It concludes that you would have to choose a single reference point to represent the time of a machine and that:
> We might even consider applying this consistency across all machines in any given data centre. This would simplify tasks such as e.g. collating accurately timestamped log entries from multiple machines. We would ignore the real longitudes of the various machines and set all of their clocks to the same local time. The interior of the facility would become an area of uniform time; a "time zone", as it were.
Have you seen the French revolutionary decimal time? It's closer to your proposal but with 10 hours/day instead of 25, for more consistency.
and the rotation speed isn't constant anyway...
Usual adjustments and shims apply, to maintain absolute accuracy, with some days having slight non-standard lengths.
> After an eight-hour flight, if you went to bed at 21:00 yesterday, you’ll do the same today.
The problem is, depending on where you landed, 21:00 could be typical start of the workday, and you'd be showing up to an empty office.
> If you had a call at 13:00 yesterday, it’s still at 13:00 today
This part sounds nice, until DST comes along to sabotage you. For only being twice a year though, it's probably still a net gain, one I hadn't really considered for someone who travels a lot.
That's how I have conky setup on my desktop. "Year : Julian Day : HH:MM:SS" in both local and UTC. With a fixed width font and the two right above each other it's exceptionally useful. It makes looking at logs captured in UTC much easier to parse.
If someone says they didn't get home until late, you know they probably mean 9 to some small hour of the morning etc.
And it's an idea that I see come up not too rarely on any post about time subtleties (time zones, leap seconds, etc.)
If you're scheduling something with someone non-local to yourself, you need not only know their time zone, but you also need also consider their latitude/longitude and what time of year it is so you account for the potential use of daylight saving time and their current sunrise/sunset times.
For example: if you schedule something at 7PM - is that daytime or nighttime? Well, if it's the first of October, and the person is in Kansas City, the answer is daytime. However, if they're in Phoenix, the answer is nighttime. And here's the kicker: the fact that Kansas City uses daylight saving time is irrelevant, as it would still be mostly light out even if you used standard time due to its higher latitude.
But if you try to schedule a meeting with someone they will immediately know if the time is reasonable for them and let you know.
Plus I don't see how it actually makes a difference, you just look up "what are business hours for $city" rather than "what time is it in $city".
Well said. Ill use this when explaining to my north american friends how 17:00 is 5pm in my head without doing any math.
https://web.archive.org/web/20211022024626/https://thehtime....
ZA:03 ZB:23 ZL:55 ZI:12 ZO:08
would work greatly for me
I like the concept of a global time that maps to local time, and while I am unconvinced that Latin letters are the best way to do that, I can’t think of a better way. It’s either letters of some sort, symbols, or names, and none of those sound good.
That said the whole conversion thing is not an immaterial burden unless you’re really up in the sky longer than walking on earth.
The whole point of UTC is to have a shared reference baseline to communicate with the world; however the author is obviously not concerned about that, I don’t think, as they have to do the conversion into the third party’s local time in their head if they want to have any friends, so that point is moot.
Moments that are fixed in time: store the event in UTC and record the rest in a string.
A description of a moment should instead by stored in the application's normalized and preferred format, often a string.
After an eight-hour flight, if you went to bed at 21:00 yesterday, you’ll do the same today.
You can complain and work on improving a tool that will covert badly, but for me it is kind of transparent with recent technology. Phone and watch update themselves automatically when you leave an airplane in a different time zone country and most tools look at your location to show you relevant time.
If you say that it will make it more convenient to work with people in other timezones, that's not even the case, because UTC will not tell you if a time is a good time or not regarding local time. So someone giving you his country/timezones is enough for you to select a right time.
Let's imagine I'm working with you that is located a quarter of the world away, how do I know in UTC what time is the work day morning proper time for a meeting? 11h UTC? 22h UTC? 4h UTC?
1) global time. i.e. UTC
2) local “sun” time. i.e. local timezone
When scheduling a meeting you are trying to find an intersection of global and local time that works.
It’s messy but also our physical reality. The conversion still needs to be done, just in the article the person is living in global time and having to convert to local time when needed. Vs living in local time and converting to another different local time.
1: https://en.wikipedia.org/wiki/Coordinated_Universal_Time#Ety...
However, I think that local time should be used as well as UTC, for different purposes, but that the local time should just use solar time, rather than using time zones and DST.
Or alternatively, each day would have to end at a different time, but then 12 o’clock wouldn’t be solar noon (or midnight).
[0] http://wordpress.mrreid.org/2014/10/19/rate-of-change-of-day...
The downside of solar time is that 1 minute of time is 15 arc minutes which is roughly 17 mi east west. Which means need to deal with time zones in same metro area.
The railroads invented time zones because dealing with local solar time was a pain. Minute accuracy is needed for time tables. We might be able to do it with computers, but it would require knowing the location of everything.
For local stuff that does not involve precise measurement of timing, you can use solar time (and you can use a sundial), and you can do by the sun light (even, in summer it is more light and in winter it is dark early, you can use that, e.g. to go outside when it is light and don't need to turn on light inside, and in night time is dark you can sleep), and by the moon light (less commonly, but sometimes it is also meaningful and useful).
Even for things that do use precise measurement of timing, UTC is not always appropriate; for some uses you will use TAI, or SI seconds instead of UTC seconds.
What's true is that the exact number of seconds between future UTC times isn't known, because leap seconds aren't known arbitrarily in advance. But that doesn’t usually affect measurements of time. It does affect adding (large) SI-second counts to the current time to compute a future time. The exact UTC designation of that future time may not be known in advance.
I don't really follow. I thought the whole concept was not having to do exactly this. Or maybe I just don't understand, I don't know.
If I'm investigating something and am interrogating logs, I'll mentally switch time bases to UTC, after which I'm natively thinking in UTC until I decide to switch back. I also have no issues converting to the handful of timezones I use often.
I'll never understand why people would want to use UTC as their native time. But then I'm on the other end of the absolutist spectrum, "often" looking up my local solar time, and shaking my fist at the sky about the local administrative time being oh so off from the "real local time" (that being, said local solar time).
These days, dealing with clients scattered across disparate timezones, no matter what, you either gotta deal with calendar differences or jet lag.
I would like to hear more about how that works out, you don't shift to sleep cycles at the place you are at, you just always go to bed at 21:00 UTC?
Thinking about the future, if humans as a species become interplanetary, keeping up with lunar time, asteroid belt time, martian time,etc.. is going to be a mess. UTC is the only way to go.
timezones are artifacts of olden days where people didn't communicate across great distances and worked fixed daylight hours. Much like being able to speak the same language, being able to use the same time across vast distances is a huge boost to productivity and progress. This is kind of like using the metric system, vs using local/traditional measurements (like the US does).
You can approximate UTC elsewhere in the universe, but it's going to be weird. You'll have the vary the length of your seconds as your velocity relative to Earth changes.
If we regularly have groups of people away from the earth, we'll probably want to follow a time standard that approximates a clock at the center of mass of the solar system.
Sunlight is different everywhere. But not everything timed is tied to sunlight. There is no reason that the time should be 12:00 at peak sunlight locally.
UTC is an issue near the international date line. I could see residents of Pacific islands wanting to not have the date change mid-day. Should think about adding hours 24, 25, 26, etc. to Wednesday that are equivalent to hours 0, 1, 2, etc. on Thursday.
(And as a quant who did a lot of time programming for financial markets, fuck Daylight Saving Time and the leap second.)
But IMHO "time of day" is less important because 1. People in the same time zone still have different schedules so it is an approximation anyways. 2. If you know what instant is being discussed it is very easy for each person to relate it to their personal schedule.
So I agree that switching costs aside using UTC would be better overall.
I find it easier to keep in mind the current time zone offset and add it mentally to the displayed value when that is necessary (i.e. for comparison with time values coming from external sources, e.g. event schedules), regardless if it is summer or winter or I am in a business trip in another time zone, than to change the displayed value to reflect the correct local time when needed, and then to account for that.
My wife has a friend who teaches classes to students in several countries and time zones. She does them by Arizona time which is among the most confusing time zones to folks from, say, India.
We tried the UTC experiment in our household once. Lasted about half a day.
The great advantage of UTC is you don’t have to deal with the abomination that is daylight saving time.
But yes I wish sometimes everyone had one standard time, trying to book a meeting with Chinese people AND west coast people while I am on the GMT line are a nightmare, as it isn't even the same day anymore!
The timezones are a formalization of existing practice, the way modern Metrication is a formalisation of practices which date back to early Weights And Measures laws.
In both cases people didn't wake up one day and from scratch invent the present sophisticated systems, they iterated, once upon a time the kilo was roughly "this much", by the 19th century it was a platinum iridium model object (the "international prototype kilogram"), today it's defined in terms of the measured constants of our universe.
Once upon a time midday was whenever the sun is directly overhead, people iterated, fast transport such as railways led to the use of standardized clocks and gradually there's a "standard" time agreed over whole regions or nations so that midday is whenever that standard says it is.
The timezones just codify and structure this existing practice. If you tell the people of Kyoto that midday ought to be eight minutes earlier than they've been having it so as to line up with a Japan-wide national time system, that's a minor annoyance. But if everybody in the world "standardized" on UTC that's eight hours different, I repeatedly tried to work out what happens, I kept getting muddled, I suspect residents experiencing this would fare little better. No.
I don't really care what time it is in England; I care about the local solar time where I live. I live in Tucson; the local solar time at my house is about 7.4 hours before UTC. I and the people who live near me would quickly start informally referring to times based on the position of the sun: "let's get lunch tomorrow an hour after solar midday".
(Aside: this isn't just hypothetical. It has happened in the real world. In the far west of China, people have settled on a different time zone for colloquial everyday use, which differs from the Beijing-imposed official standard time by two hours.)
So far so good, but someone on the far east end of Tucson is more than a minute later than me; if we want to be precise, we'll need one standard for the whole city, so maybe they'd build a big sundial at the University of Arizona and that would serve as the reference for the whole city.
But people drive around Arizona a fair amount; it'd be annoying to have to adjust my model of time by five minutes every time I go to Phoenix, and I wouldn't mind the few minutes of slop applied to my own life in Tucson, so I'd probably just start referring to everything in terms of offset from "Phoenix midday" and "Phoenix midnight" to make things easier. And if we're going that far, we might as well make it an even bigger area: big enough to include the whole Southwest, but small enough that the difference from local time doesn't bother anyone. Let's say we base it on Denver.
Bam, we just reinvented Mountain Standard Time.
Well, almost. Meanwhile, people on the East Coast would have (almost) reinvented Eastern Standard Time, presumably based on their most important city, which is New York. New York is about 1h55 ahead of Denver in solar time. Dealing with these almost-but-not-quite hour offsets would be annoying, so we'd probably just standardize on one system for the whole country, with nice, even, easy to reckon hour-multiple offsets for everyone.
Canada and Mexico, whose economies are highly dependent on trade with the US, would soon standardize on the US-imposed time zones. Even if doing so might hurt their patriotic sense of national pride, the practical value would be too great. But what about the rest of the world? Well, our historical closest friend and ally in a far-off longitude is the UK, so why not sync up with them and make all those easy to reckon hour-multiple offsets from their most important city, London, or more precisely, the point they've chosen for London's standard time: the Royal Observatory in Greenwich.
Now we really have reinvented Mountain Standard Time, exactly as it exists in the real world.
Why wouldn't we go further and just use Greenwich time for everything? Well, because we still do care about local solar time. The difference from my house's solar time to the newly invented MST is less than an hour, which I can live with, but dealing with a 7-hour difference would be annoying.
I have let through some timezone bugs, tho those bugs also didn't impact prod since prod machines are UTC
Done with both EDT & PDT,
I did have a mixup with catching flight: I read 2300 as being UTC (since I'm not used to 24h clock being localtime) & thought my flight was 7PM instead of 11PM. Would be nice if flight times were UTC, always annoying figuring how long flight actually is
Problem is though with distributed offices and remote workers. “The servers started crashing at 10:31 can you look at it?” Doesn’t mean anything without more context. If it’s 10:50 local does that mean it started 20 minutes ago or 7:20 ago? 3:20 ago because they’re on the east coast?
> Moscow time, my local context.
Ah, Russia, what would we do without you! Never change. https://gwern.net/note/note#russia
Because it was very easy to miss a little "PM" pip in the corner, when reading or setting the time; and setting an alarm depended on getting the pip correct, so eliminating the pip for me meant eliminating ambiguity in the digits.
It's been a fairly successful transition, although sometimes I look at "16:15" and think it's 6pm so the "mentally subtracting 12" needs practice. And there are still special snowflake UI widgets that ignore my i18n settings, and constrain how I enter or specify a time.
Seeing two analog clocks side by side is the fastest way I can decode any concerns around the time difference. Digits on 24h UTC clocks makes my brain switch to a wildly more complex mode of operation.
Isn't that sentence contradicting the entire thing? If you don't care, then use local time. Unless this is satire that I didn't get :P
Note I'm not saying I agree, just explaining.
That's because you committed one of the capital sins for a programmer: you tried working with time. The old mantra "use a library" still rules.
> I decided time zones should be abolished, and everyone should use one coordinated time. Why care if the sun rises at 8 or 21?
You seriously need to go out more often.
Pretty much that. Most mammals I can think of live by what the sun, one way or another.
Source: https://pmc.ncbi.nlm.nih.gov/articles/PMC4183310/#:~:text=Gl...
Some people just like to add complexity to their lives and keep convincing themselves that it's better.
Anecdotally, I worked on a military facility where all clocks were in "master timezone" of the country without daylight saving offset. 6 months a year we put "+1" sign made of electric tape to not forget about that it's summer and 0400 is actually 0400+TZ+1. I worked there for 5 years and never got used to it.
It's a little different when going to bed one day and waking up another doesn't happen anymore. And small numbers aren't morning, big numbers aren't night.
I hate this form of nerd clickbait. tech bro discovers UTC, thinks he’s cracked the code of the universe, and writes a navel-gazing manifesto about how changing his watch changed his life. of course it ends in a SaaS plug.
5 days a week I am working with people in Seattle, NYC, Munich, and often London or India. I have to coordinate meetings across time zones a lot. I travel a lot.
And yet at home or on the road I have to process things like "this store closes at 18:00" or "this train is at 16:45" infinitely more. Putting my devices in UTC requires me to think for all of those interactions which is way worse.
Why deal with this stupid human construct of your day starting when the sun comes up when the sun isn’t relevant to your local data center?
Heck why not just count up milliseconds since 1970?
Wow. This is amazing. Who knew this only worked with UTC!! /s
Missives like this just make me embarrassed for us “technical folk”.