Shameless plug-
Check out the Niles Canyon Railway (NCRy.org). It’s a volunteer organization that runs vintage trains (including steam engines) on the original connector route that was completed later in the year to finally connect the bay with Sacramento.
That does make more sense but my first guess was because Sacramento is the capital and close enough to San Francisco. I’m not sure that the Bay Area was particularly well defined yet either.
Interestingly while I was checking the incorporation dates of Oakland and Brooklyn (now part of Oakland) I saw that 1868 is the year that Central Pacific Railroad settled on Oakland as the West Coast terminus of the first transcontinental railroad and bought the San Francisco & Oakland Railroad (https://en.wikipedia.org/wiki/San_Francisco_and_Oakland_Rail...).
> my first guess was because Sacramento is the capital and close enough to San Francisco
That didn't stop them from independently listing Dayton, Cincinnati, and Columbus Ohio, all with different times (even though Dayton and Cincinnati are only 1 minute apart).
If what the parent said was correct, then it wasn't merely because Sacramento was the capital, so my first guess would be wrong under that criteria.
And it makes sense they would list so many Ohio cities if major railroad hubs was the criteria. Ohio was practically the hub of the nation with major railroads running through all the cities you listed plus riverboats. California at this point was comparatively speaking still pretty much a backwater, even post-Gold Rush, and wasn't politically nor industrially significant before at least the first transcontinental railroad. No one in Congress had even worked particularly hard for the California EC votes until the Chinese Exclusion Act of 1882.
I've been reading a book that was highly recommended here on HN a few months ago: "Nature's Metropolis", and just finished a chapter that explains this.
Back in 1868, there were essentially just two rail networks in the US - the eastern half and the western half. They met in Chicago. The operational model was very different, and the presence of Great Lakes shipping (with low summertime shipping rates) kept either side from trying to expand to be truly nationwide.
In essence, a timetable relative to Washington D.C. showed Sacramento just as an example. No railroad went from D.C. to anywhere west of Chicago. Not only did you change trains, you changed railroad companies.
I'm sure this gets recommended all the time, but "Time Lord: Sir Sandford Fleming and the Creation of Standard Time" is a fascinating (to me anyway) story about the creation of time zones; how and why they came about.
Himself a railroad engineer and present at the driving of the "Last Spike." Quite the individual. I really do like the charm of "cosmopolitan time:"
The zones were labelled A-Y, excluding J, and arbitrarily linked to the Greenwich meridian, which was designated G. All clocks within each zone would be set to the same time as the others, and between zones the alphabetic labels could be used as common notation. So for example cosmopolitan time G:45 would map to local time 14:45 in one zone and 15:45 in the next.
Depends on the time zone and changes to it. If it's a timezone without seasonal changes and its other properties don't change during the life of the organization, that's fine; maybe it's not ideal, but it's fine.
If it's a timezone that has its operational properties change, you have the potential for confusion where there is disagrement about when a given time happened or will happen; think about how fun it is when daylight savings rules change. Sometimes, although not really that often, a time zone's offset is simply changed; hopefully with warning, but maybe not.
If the timezone has daylight savings, you get to deal with all the fun of most days having 24 hours, but some having more or less.
If the organization moves to a new time zone, either because it physically moves or because the locality its in changes time zones for whatever reason, now you have a problem; it's not fun to deal with multiple time zones in server infrastructure, but now you have to pick if you want to move all the servers, move only some servers, or have your servers in a different time zone than your org. If the last option was fine, you might as well use a sensible timezone to begin with.
You can always interpret the timezone on the app/client/presentation side, I can't think of any scenario that is made better by the server itself being set on local time zone. It causes extremely messy issues IME.
1. If the time zone changes (like, as it's happened multiple times in my country, the date for switching to daylight savings changes) you have to make sure all your servers is updated with the change
2. Daylight savings mean that events an hour apart can end up having the same timestamp.
I've fortunately never experienced (1), but for (2), it helps to print the offset to UTC everywhere timestamps are printed, even for UTC. (I basically never trust an "unzoned" timestamp.)
I guess you’ve never had to correct the local time to UTC for handoff to a different system and the conversion was buggy in the wrong direction? Or haven’t had to figure out which 2:30am something happened? Or why something that usually takes a second took an hour?
I've had to do all of these things many times, but keeping server times on UTC doesn't really help with most of them:
Customers, partners, stakeholders all think in terms of local time zones, as do I; some amount of conversion is inevitable.
Ultimately, local time is a presentation layer concept, so why shouldn't I configure that presentation layer so that I have to do one less conversion in the most common case (of figuring out when something happened relative to my wristwatch or that of most of my customers)?
Are you saying you should configure the server to local time to save you one less configuration on the presentation layer? That doesn't make sense to me. I really can't think of any benefits to setting the server to local time that doesn't offset the huge load of negatives.
Here's something you may not be thinking about - your presentation layer may save a configuration step, but now every system that interacts with your server (who have all likely defaulted to UTC) have to know about this now. Think - logging infrastructure, schedulers, cloud infra, etc. So now you have to configure all those as well. Anything that touches the host will need to be reconfigured to save one very minor and commonly done thing on the presentation layer. Really doesn't make sense, as an infra guy.
> every system that interacts with your server (who have all likely defaulted to UTC) have to know about this now.
No, why? Almost all APIs/interfaces use zoned timestamps or seconds since the Unix epoch.
The argument of “servers must run in UTC” sounds a bit like “servers must run with an English locale” to me in that sense, as if the locale changed the name of API parameters or enum values instead of just human-readable strings.
When Saudi Arabia built the world's largest tower clock (43 meter dials) atop a hotel in Mecca, the original plan was to sync it to solar time at Mecca, as a political statement. Fortunately, sanity prevailed. It's currently set to UTC+3.
I never got what different time zones was supposed to accomplish. Why doesn’t everyone just use GMT? Why does the clock have to say 8 when you go to work? Wouldn’t it be better if we used the same clock everywhere?
Let’s say we do that, use UTC everywhere. You live in London. It’s 3 PM. For you, that’s mid-afternoon. It’s your brother’s birthday. You want to phone him and wish him a happy birthday. He lives in Sydney. In a world without timezones, it’s 3 PM for him too. So you give him a call and start singing loudly when he answers. He’s mad at you because actually it’s the middle of the night and he’s sleeping.
Or think about traveling. At home, you set your alarm for 6 AM. You really like waking up at this time even when you travel. You want time to hit the gym, and eat a nice breakfast. Now, in a world without timezones, you travel to SF. When your flight lands at 5 PM, what phase in the day is it? Local “morning”? Local “late evening”? Are you going to have a hard time catching a taxi? What time do people eat breakfast there? What time should you set your alarm for?
So you need a translation layer from your location to another location to know what phase in the day it is for people in the other location. Is it their “morning”? Their “afternoon”? Their “business hours”? When do they sleep?
That translation layer exists. That’s timezones.
Abolishing timezones doesn’t make coordination problems easier. In fact, it makes them harder. The time on the clock might be the same around the world, but when people do things (wake up, eat breakfast, conduct business, grab a cocktail) would vary around the world. (Yes, it’s true there’s regional variance to these things today, but for the most part you can rely on morning having a rough relationship with when people wake up and sunrise, for example.)
These are only obvious if you never move outside of a few select zones. Time on the wall still being somewhat related to the sun doesn't even apply to countries with strong summer/winter variations. That's part of why killing DST has gained much more popularity in recent years.
I think it's clinging to something that was an approximation in the best case scenario, and appears to be pure fantasy the more we dig into it and look around.
People think in different ways and have different levels of education. We would never learn if we weren't allowed to ask silly questions. It is very important not to discourage people from asking them.
Don't trivialize the other posters reasonable comment.
These things only become apparent with international travel or communication, which 99% of the world does not do on a regular basis. Even with the other comment making sense in it's points, I still disagree with it. So don't disparage that other user for having an idea with serious merit that only has some subjective drawbacks.
It's possible I'm being unreasonable, but I do not think it's a reasonable starting point for any discussion to open with "I never understood X, so why should we not delete X?" Like, is it not reasonable to educate yourself before asking strangers to educate you?
Timezones have existed for nearly 200 years, before international travel or distance communication was commonplace, so I implore you to explore even the assertion that those are pre-requesities for understanding why timezones might be useful.
> having an idea with serious merit that only has some subjective drawbacks.
Abolishing local solar-approximate time is an idiotic idea with multiple objective drawbacks, not to mention that it would be massively unpopular. IF you think it's a reasonable idea, you really haven't thought it through, and you absolutely do not understand how time has meaning to people, and how our entire societies are built on shared collective ideas of what different times of the day mean to us.
The level of stupidity of the idea is on par with "hurr durr, why don't we get rid of our nuclear waste by shooting it into the sun in a rocket?"
> Let’s say we do that, use UTC everywhere. You live in London. It’s 3 PM. For you, that’s mid-afternoon. It’s your brother’s birthday. You want to phone him and wish him a happy birthday. He lives in Sydney. In a world without timezones, it’s 3 PM for him too. So you give him a call and start singing loudly when he answers. He’s mad at you because actually it’s the middle of the night and he’s sleeping.
London to Sydney is a poor example because it takes a very stupid person to fail to realize their _brother_ who is on the other side of the world isn't up at midday.
But how about New York City and San Francisco. People do that _today_, because it's 8AM in New York and they forget it's 5AM in San Francisco -- or reversed, and it's 8PM and 11PM.
Why is that a problem? Because people have to keep track of the timezones and then do the math (most people are shockingly bad at basic arithmetic). That's a problem _today_ that you are ignoring.
> Abolishing timezones doesn’t make coordination problems easier. In fact, it makes them harder.
You claim that but you haven't demonstrated that. And the evidence I've seen points to the opposite.
Three brothers, Tom, Dick and Harry all live in the same city, in the same timezone. They know that Tom gets up early and goes to bed early, Dick gets up late and goes to bed late, and Harry travels so God knows when he's awake. They _know_ this and don't call Tom late at night, Dick early in the morning, or Harry without knowing where in the world he is.
Now spread those brothers around the world and the same thing is true: Tom gets up at 1000Z and is asleep by 0200Z, Dick gets up at 1700Z and is asleep around 0900Z, and God knows about Harry.
No timezones, a single clock and much simpler arithmetic. This also makes planning meetings easier.
Timezones are a 19th century solution to a 19th century problem. They _FAIL_ at our 21st century world.
The more interesting question though is why didn't timezones develop like this initially? In 1868 there was no telephone to call another continent. There was no airplane for fast travel over long distances.
In retrospect this makes sense but in the late 19th century, no one could have foreseen this use case.
This is the common rebuttal, but rarely mentions alternatives to timezones.
One might be a clock face with local solar time as light/dark patches portraying the "movement" of the sun. Perhaps digital UTC in the center. A button to switch cities, etc.
I've seen a few implementations of this with LCD clock faces. They don't typically show UTC as well, but could. Once implemened getting everyone to use it would be the hard part.
Yet it comes crashing down when you think that for people in Barcelona 3PM is mid afternoon. Or that nothing's running at 6AM in Tokyo. Or whatever you assume is happening at 6PM in the Shichuan area.
Those are all random assumptions that could be better served by a sunrise - zenith - sunset representation (which has not much to do with time of the day anymore), or heck, checking the typical day rythm of that place, instead of hypothesing in our heads. Not counting that what someone does in a day will be highly variable depending on their occupation and personality.
Knowing that it's currently 7PM in India also helps me in no way to decide wether it's a good time to phone a store. Checking the store website will help a lot more.
Same way if I'm traveling I want to align my wake hours with the stuff I actually plan to do. If that means waking ay 9PM for whatever reason, then 9PM it is.
We can do better than keeping heuristics that only match very small patterns, that basically shatter when we're talking about the other side of the world.
They match huge humanity-encompassing patterns that have been true for millenia, for as long as we've kept time. We kinda like organizing our time in waking periods that we already call days.
Yes, timezones are a rough translation layer that gives you an idea of what's going on in a society on the other side of the earth, and yes, it's not a perfect solution to the problem of cross-timezone communication and coordination. It is an arbitrary system, and it could be replaced with a different system.
But the biggest problem with abolishing timezones is that you're destroying the ability to keep track of days of the week for some large chunk of humanity. If it's midnight at the same time everywhere, then the day of the week switches at the same time, everywhere. Things like "open on Wednesdays" will cease to have meaning, because for billions of people, they day of the week will now switch in the middle of the working day.
"9 to 5" will only be true in a single former timezone, some people will now work from 8pm to 4am. When does their weekend start? Is that the waking period that now covers Friday/Saturday, or the one that now covers Saturday/Sunday?
Every place of business, every school, every store, every restaurant will have to print new opening hours depending on where in the world they are, because the local time has now changed for everyone.
Everyone will now have to learn their local translation table so that they know what normal working hours in their location is now, when schools open, when lunch hour is, etc. You're throwing away all of our collective knowledge and intution about time, in order to make it "easier" to schedule cross-timezone meetings.
Abolishing timezones would piss off about 7 billion people for absolutely no gain.
Yes, when Alice in London schedules a meeting with Bob in Sydney they will now make no mistakes about which point in time the meeting is at, but Alice still needs a translation table to figure out what the meeting time means for Bob.
Timezones are that translation table, it imbues times with meaning.
> Things like "open on Wednesdays" will cease to have meaning
They already don't have that much meaning and we still communicate fine enough. For instance a dance club open on Saturday nights probably closes on Sunday, but nobody is troubled by the imprecision. It might even actual open on Sunday at 00:30, but still advertise it as Saturday night 24:30. Same for restaurants that stay open late enough. Or convenience stores, gas stations, theaters, gym clubs, barbers, tv shows etc.
You're right that for centuries the notion of "a day" was structural to everyday life. Most countries are past that point.
> Alice still needs a translation table to figure out what the meeting time means for Bob.
Why doesn't she ask Bob ? Isn't he the one who understands when his kids are back from school, when does the grocery store close, or if he has a 2h slot right after the sun rises where he can focus on Alice's project, or he needs to be at the office that opens at 8AM.
In my view, knowing that "10 AM in Sidney is roughly a few hours before the zenith"
helps very little in practical matters. The information is way too vague and out of context to be usable.
How I'm reading this - the argument is: we need timezones because, above all, we all want time to be fairly inline with daylight hours everywhere or at least some approximation of a 9-5 workday?
China is one giant timezone, despite being 20% wider than the US. Imagine going to work at 6 am or coming home from work at 8 pm as the normal thing, depending on where you lived, not adjusted for local sun times.
we want to be able to say "in the morning" and have it be reliably interpreted, because most human activities are actually tied to the sun and not to the clock (the clock is incidental for precise timing, but it's just reflecting the sun position).
The part that's really weird to me is those lines are latitude based, so Finland and South Africa share the same timezone.
Sun wise, situations are a pretty different at any moment in the two countries. I wouldn't call the system "reliable" ( and those are not outliers, any zone with countries on opposite hemispheres has the same issue)
I'm kinda glad I tried to explain this, because I realize just how much I lack the proper words for it.
The earth rotation wobbles a bit, but the axis itself is also not tilted towards the sun (it has no specific reason to) nor tilted towards the inside of the revolution axis relative to the sun.
Basically the earth can be spinning on itself around any axis (let's call it A), while revolving around the sun on an unrelated axis (B). And neither A nor B has to be following the North/South axis. That's where the idea that a latitudinal timezone sees a similar solar cycle seems at odds to me.
Right they aren’t aligned. Sun’s axis is aligned perpendicular to the ecliptic. Earth is tilted 23.5 degrees from that. So North pole is tilted towards the sun on June 21 and South pole tilted towards it on December 21—why there are seasons. Doesn’t affect timezones to my knowledge.
The alternative is that we have a DNS-style lookup of every city on earth which contains what each city has chosen for their daylight hours. You’d run into many more boundary-centric issues than with our columnar approach.
> The alternative is that we have a DNS-style lookup of every city on earth which contains what each city has chosen for their daylight hours.
We already have to do that: every business and government office (including schools) has their own hours (and holidays); even in the same city, people are not waking (or going to sleep) at a uniform time.
Time zones are a 19th century solution to a 19th century problem.
A lot of people have already mentioned this, but the short answer is that you ultimately need a civil time that approximates local solar time. For example, you want to make sure that day changes happen at a time when there's little activity. Civil time doesn't need to be exact to solar time, and there's room for a fair amount of divergence, but when the divergence gets too great, people will spontaneously create their own civil time instead of using the official civil time (see, e.g., western China).
It's better to have a minimal set of official civil times that are good enough for local purposes than have a large set of unofficial and semiofficial civil times.
You need time zones to tell what's "early" or "late" in a place. If you're planning to call someone far away, picking a time "during work hours" becomes very difficult without time zones. You would have to consult some chart that describes when "work hours" are in each region which is basically just reinventing time zones.
> If you're planning to call someone far away, picking a time "during work hours" becomes very difficult without time zones.
Does everyone in your office work the same hours?
We _already_ have the problem of "What are Bob's work hours" because most office workers don't work the same hours. Some folks come in early, some work late -- some come in late and leave early.
That has been common for decades and has become even more pronounced with increased remote work, particularly across time zones.
We shouldn't build our society around a myth of 9-5 office workers.
I'll add that for serious scheduling, people use digital calendars extensively. Some sort of "work hours" would need to be marked on those, but everything else would be simpler.
I don't know what it is as a percentage, but globally there is in absolute terms significant amounts of East-West inter-timezone communication going on all of the time. Like literally, it does not stop happening and is happening right now.
Percentages are misleading when the population is several billions of people conversing and conducting commerce at near-light speed on a global scale. Just the three hour difference between the East and West coast of the United States is significant enough to be worthy of consideration on when to call people on the other coast (or just expect them to get back to you through asynchronous comms) because people on the West Coast are on average waking up three hours later as measured by GMT.
The same is true for the proposed global universal time. Without inter-timezone communication, everyone would just operate relative to their local solar noon, like they did pre-nineteenth century. The 'edge cases' of inter-timezone communication and travel are the only reason we need to care about any kind of coordinated time system.
A lot of us expect that "noon" has some connection to "sun at its highest point in the sky". Having that happen at 10 PM just feels wrong.
But in fact, the military does have one standard time - Zulu time.
Why have so many time zones? Well, they were originally introduced by railroads in the 1880s. Back then, "noon" was defined as "highest sun", so deviating very far from that was unreasonable. But also, railroads didn't run very fast back then. Each time zone was something like 700 miles across. That was enough to give enormous operational improvements. Removing the transitions (going to all one time zone) would have added only a little additional improvement.
"the military does have one standard time - Zulu time"
Nah they don't. Skipping over the (common on the internet) parochiality of assuming that US military practice is just 'what the military does'...
The US military certainly uses UTC/Zulu time for time stamping messages and coordinating operations globally, and timekeeping during extended missions (like submarine patrols, or air missions), but military bases and even surface vessels underway keep somewhat local time. They don't use Zulu time - that would mean the date changes in the middle of the day. You wouldn't be able to use words like 'next Tuesday' without having to disambiguate. It'd be crazy.
The justifications given above don't really make sense - "how would you know when to call Sydney" or "the sun being overhead at 10PM would feel strange". They are just rationalizations of a system that we're used to.
But everywhere around the world, people want the date to change at a time when most things are closed and most people are asleep - which coincides with night-time for obvious reasons. This is effectively more important than having the date change at the same time for all of us.
I'm still not sure the costs outweigh the benefits, but I agree that this is the only concrete requirement that can't be solved with time, slow cultural change, and adjustments to some software.
People used to set the town clock locally with noon when the sun is directly overhead. Also precision didn’t matter much when you weren’t trying to catch a train,
Time zones were invented for time agreement across longitudes. We could easily do away with them if we included longitude, and everybody used a true solar noon clock. It could be automatic for GPS-enabled devices to maintain accurate solar noon time. Actually GPS satellites send a very accurate time on the signal, so you wouldn’t necessarily need a time server, except to know which epoch you’re on.
