8 ms·

Why shouldn't we have one timezone? Why not have one worldwide timezone? The sun rises and sets at different times according to both latitude and longitude. Ti

Why shouldn't we have one timezone? Why not have one worldwide timezone? The sun rises and sets at different times according to both latitude and longitude. Timezones only handle longitudinal daylight adjustments. We already deal with plenty of incongruity around daylight. Why not embrace the variability and let go of the idea that we can predict daylight based on the hour alone? Let people localize the time to discover where the sun is.

chickenpotpie 3y ago
Won't people just end up coming up with their own pseudo time zones anyways? If I'm in New York and I'm calling someone in Seattle, aren't I gonna have to just mentally subtract 3 or so hours anyways?
- loudmax 3y ago
  Yes. There's no getting around the fact that the sun rises and sets around the world at different times. What we do instead is get rid of the assumption that everybody works 9am to 5pm local time. It does require a different way of thinking about time zones. It sounds complicated because it's unfamiliar, but if we ever make the transition, the system would be simpler, less ambiguous, and easier to reason about.
ethbr0 3y ago
> Why shouldn't we have one timezone? Why not have one worldwide timezone? Your watch hands read '20230710T142847Z', don't they?
- adolph 3y ago
  No. For one, I use the ordinal date option in ISO 8601 because the whole month and day thing is stupid. Today is 2023-191. Also my watch only reads 14:28 since the date and TZ are typically implied in watches. https://en.wikipedia.org/wiki/ISO_8601 https://en.wikipedia.org/wiki/ISO_8601 One thing for which I would advocate is removing zero based numbering from hours and minutes. Clocks reading 12:15 for 00:15 is tremendously confusing. Most human oriented counting systems are not zero indexed. One indexed series align cardinal and ordinal numbering which reflects the passage of time rather than each hour being nominally disconnected from a sequence. https://en.wikipedia.org/wiki/Zero-based_numbering https://en.wikipedia.org/wiki/Zero-based_numbering
  - Timon3 3y ago
    > No. For one, I use the ordinal date option in ISO 8601 because the whole month and day thing is stupid. Apart from this being very funny and an interesting idea I have a question: how do you deal with leap years? Many important dates throughout the year (e.g. holidays, birthdays, business quarters etc.) are bound to specific days of months, do you just have to remember that all such days after February are one off sometimes in an ordinal date system?
    - adolph 3y ago
      I appreciate your questions. I will honor them with the same dry humor by responding as if actually using ordinal dates were something many people did. Leap years: A sidereal year adjustment may occur on any day, there is not a requirement for it to occur on a specific day. Leap years merely have an additional day identifier in the series before returning to 001. Anniversary events: Again, there is no requirement that the memorialization of a person or event occur on a specific month and day. Many occurrences such as Mardi Gras and Eid using a lunar calendar or the USA's Thanksgiving using the day of week. If one wishes to commemorate an annual event at the same orbital position as when it occurred, then ordinal dates are more effective since the Gregorian calendar's fourth-year 02/29 throws off any anniversary after the first 1/6th of the year. NB: ISO 8601 "leap week" instances occur at a different rate than leap years (a more parallel construction would be "leap week" and "leap day"). "[The] 53 week years occur on all years that have Thursday as the 1st of January and on leap years that start on Wednesday the 1st." https://en.wikipedia.org/wiki/Leap_year https://en.wikipedia.org/wiki/Leap_year https://en.wikipedia.org/wiki/ISO_week_date https://en.wikipedia.org/wiki/ISO_week_date A higher order number than ordinal and cardinal is ratio. We can also express the current day as how far through the year we are. Since today is 192, it represents 52% of a year, the ratio of current day to days in the current year. Below is an example in bash. $ date '+%j' 192 $ echo "scale=2; $(date '+%j') / $(date '+%j' -d $(date '+%Y')'-12-31')" | bc .52
petesergeant 3y ago
> Why not have one worldwide timezone ... we already deal with plenty of incongruity around daylight ... why not embrace the variability and let go of the idea that we can predict daylight based on the hour alone? Because it's a useful heuristic in knowing whether someone will be at work or not without having to localize every time.
leetcrew 3y ago
timezones don't need to perfectly predict whether the sun is up to be useful. when I set a meeting with a colleague in India, I don't really care whether the sun is up. it will almost surely be down for at least one of us. I am just trying to pick a reasonable time where they don't have to end work too late, and I don't have to start too early. having everyone roughly align their work day to 9am-5pm in their local time works a lot better than asking every single person in my company for their working hours in UTC.
- laserlight 3y ago
  I don't get how the problem would be harder with a universal timezone. Currently, I have to look up which timezone India is in and find a good time that works for both of us. In a universal timezone, I would have to look up what time noon is in India and find a time that works for both of us.
  - leetcrew 3y ago
    so the high level problem here is "how can I infer a locale's working hours without asking every individual?" the current solution is to set an explicit expectation that people work 9-5ish in their local offset. is your proposal to do the same thing, but the expectation is that people work noon-4 to noon+4, or something like that? I don't see any major advantages to that, but I do see one major disadvantage: solar noon occurs at a different time every day. of all the arbitrary markers to align schedules with, this seems like a pretty bad one.
    - laserlight 3y ago
      > is your proposal to do the same thing, but the expectation is that people work noon-4 to noon+4, or something like that? Yes. Rather than using solar noon at one day, we can look up an average solar noon time. It should give a better idea how people would plan their day. I understand that schedules will be different at different locations, but then 9 to 5 also has that much variation from location to location.
  - makingstuffs 3y ago
ooterness 3y ago
Why even use time that's linked to the rotation of our home-planet? That's as geo-centric as it gets. Just count the number of seconds since the GPS epoch in a galaxy-centered inertial frame. This post was made at 1,373,034,914.
- shpx 3y ago
  A second is one 86400th of a rotation of our home-planet
  - rpearl 3y ago
    A second is defined to be exactly the duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium-133 atom. I don't see anything referencing any planets?
    - throwaway09223 3y ago
      Well, a second is more specifically defined as: "the duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium-133 atom taken as an average over 11 samples distributed throughout the earth in order to approximate the earth's inertial frame" Seconds as used in Coordinated Universal Time are in fact defined based on planets, and earth specifically - because they use earth's inertial frame.
      - rpearl 3y ago
        The SI definition of a second, available at https://www.bipm.org/documents/20126/41483022/SI-Brochure-9.pdf/fcf090b2-04e6-88cc-1149-c3e029ad8232 https://www.bipm.org/documents/20126/41483022/SI-Brochure-9...., section 2.3.1, does not reference our particular planet, nor any particular means of measurement, nor does it have to. In principle this frequency can be measured in any inertial frame. Seconds are not defined on Earth specifically.
watwut 3y ago
Because this way, when someone tells me "it was 5:00 or 16:00" I know whether it was their morning or evening or around lunch. I do not have to take out map to figure out where they were located to figure out any of that.
mjswensen 3y ago
I used to want this too, until I thought about what it would be like to have a date/weekday change during waking hours for most of the world. I suspect that would cause even more problems than the timezone headaches.
pseudalopex 3y ago
https://qntm.org/abolish https://qntm.org/abolish

chickenpotpie 3y ago

Won't people just end up coming up with their own pseudo time zones anyways? If I'm in New York and I'm calling someone in Seattle, aren't I gonna have to just mentally subtract 3 or so hours anyways?

loudmax 3y ago
Yes. There's no getting around the fact that the sun rises and sets around the world at different times. What we do instead is get rid of the assumption that everybody works 9am to 5pm local time. It does require a different way of thinking about time zones. It sounds complicated because it's unfamiliar, but if we ever make the transition, the system would be simpler, less ambiguous, and easier to reason about.

ethbr0 3y ago

> Why shouldn't we have one timezone? Why not have one worldwide timezone? Your watch hands read '20230710T142847Z', don't they?

adolph 3y ago
No. For one, I use the ordinal date option in ISO 8601 because the whole month and day thing is stupid. Today is 2023-191. Also my watch only reads 14:28 since the date and TZ are typically implied in watches. https://en.wikipedia.org/wiki/ISO_8601 https://en.wikipedia.org/wiki/ISO_8601 One thing for which I would advocate is removing zero based numbering from hours and minutes. Clocks reading 12:15 for 00:15 is tremendously confusing. Most human oriented counting systems are not zero indexed. One indexed series align cardinal and ordinal numbering which reflects the passage of time rather than each hour being nominally disconnected from a sequence. https://en.wikipedia.org/wiki/Zero-based_numbering https://en.wikipedia.org/wiki/Zero-based_numbering
- Timon3 3y ago
  > No. For one, I use the ordinal date option in ISO 8601 because the whole month and day thing is stupid. Apart from this being very funny and an interesting idea I have a question: how do you deal with leap years? Many important dates throughout the year (e.g. holidays, birthdays, business quarters etc.) are bound to specific days of months, do you just have to remember that all such days after February are one off sometimes in an ordinal date system?
  - adolph 3y ago
    I appreciate your questions. I will honor them with the same dry humor by responding as if actually using ordinal dates were something many people did. Leap years: A sidereal year adjustment may occur on any day, there is not a requirement for it to occur on a specific day. Leap years merely have an additional day identifier in the series before returning to 001. Anniversary events: Again, there is no requirement that the memorialization of a person or event occur on a specific month and day. Many occurrences such as Mardi Gras and Eid using a lunar calendar or the USA's Thanksgiving using the day of week. If one wishes to commemorate an annual event at the same orbital position as when it occurred, then ordinal dates are more effective since the Gregorian calendar's fourth-year 02/29 throws off any anniversary after the first 1/6th of the year. NB: ISO 8601 "leap week" instances occur at a different rate than leap years (a more parallel construction would be "leap week" and "leap day"). "[The] 53 week years occur on all years that have Thursday as the 1st of January and on leap years that start on Wednesday the 1st." https://en.wikipedia.org/wiki/Leap_year https://en.wikipedia.org/wiki/Leap_year https://en.wikipedia.org/wiki/ISO_week_date https://en.wikipedia.org/wiki/ISO_week_date A higher order number than ordinal and cardinal is ratio. We can also express the current day as how far through the year we are. Since today is 192, it represents 52% of a year, the ratio of current day to days in the current year. Below is an example in bash. $ date '+%j' 192 $ echo "scale=2; $(date '+%j') / $(date '+%j' -d $(date '+%Y')'-12-31')" | bc .52

petesergeant 3y ago

> Why not have one worldwide timezone ... we already deal with plenty of incongruity around daylight ... why not embrace the variability and let go of the idea that we can predict daylight based on the hour alone? Because it's a useful heuristic in knowing whether someone will be at work or not without having to localize every time.

leetcrew 3y ago

timezones don't need to perfectly predict whether the sun is up to be useful. when I set a meeting with a colleague in India, I don't really care whether the sun is up. it will almost surely be down for at least one of us. I am just trying to pick a reasonable time where they don't have to end work too late, and I don't have to start too early. having everyone roughly align their work day to 9am-5pm in their local time works a lot better than asking every single person in my company for their working hours in UTC.

laserlight 3y ago
I don't get how the problem would be harder with a universal timezone. Currently, I have to look up which timezone India is in and find a good time that works for both of us. In a universal timezone, I would have to look up what time noon is in India and find a time that works for both of us.
- leetcrew 3y ago
  so the high level problem here is "how can I infer a locale's working hours without asking every individual?" the current solution is to set an explicit expectation that people work 9-5ish in their local offset. is your proposal to do the same thing, but the expectation is that people work noon-4 to noon+4, or something like that? I don't see any major advantages to that, but I do see one major disadvantage: solar noon occurs at a different time every day. of all the arbitrary markers to align schedules with, this seems like a pretty bad one.
  - laserlight 3y ago
    > is your proposal to do the same thing, but the expectation is that people work noon-4 to noon+4, or something like that? Yes. Rather than using solar noon at one day, we can look up an average solar noon time. It should give a better idea how people would plan their day. I understand that schedules will be different at different locations, but then 9 to 5 also has that much variation from location to location.
- makingstuffs 3y ago

ooterness 3y ago

Why even use time that's linked to the rotation of our home-planet? That's as geo-centric as it gets. Just count the number of seconds since the GPS epoch in a galaxy-centered inertial frame. This post was made at 1,373,034,914.

shpx 3y ago
A second is one 86400th of a rotation of our home-planet
- rpearl 3y ago
  A second is defined to be exactly the duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium-133 atom. I don't see anything referencing any planets?
  - throwaway09223 3y ago
    Well, a second is more specifically defined as: "the duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium-133 atom taken as an average over 11 samples distributed throughout the earth in order to approximate the earth's inertial frame" Seconds as used in Coordinated Universal Time are in fact defined based on planets, and earth specifically - because they use earth's inertial frame.
    - rpearl 3y ago
      The SI definition of a second, available at https://www.bipm.org/documents/20126/41483022/SI-Brochure-9.pdf/fcf090b2-04e6-88cc-1149-c3e029ad8232 https://www.bipm.org/documents/20126/41483022/SI-Brochure-9...., section 2.3.1, does not reference our particular planet, nor any particular means of measurement, nor does it have to. In principle this frequency can be measured in any inertial frame. Seconds are not defined on Earth specifically.

watwut 3y ago

Because this way, when someone tells me "it was 5:00 or 16:00" I know whether it was their morning or evening or around lunch. I do not have to take out map to figure out where they were located to figure out any of that.

mjswensen 3y ago

I used to want this too, until I thought about what it would be like to have a date/weekday change during waking hours for most of the world. I suspect that would cause even more problems than the timezone headaches.

pseudalopex 3y ago

https://qntm.org/abolish https://qntm.org/abolish