Most are familiar with leap years, in which an extra day is added to February every four years to keep our human calendar on track with the Earth's orbit. A similar phenomenon also occurs on a smaller scale. Modern timekeeping systems have traditionally used "leap seconds" to keep us in sync but, as USC Dornsife physicist Vahe Peroomian explains, this system may change starting this month.
Replacing it with a leap hour is probably the worst possible solution.
But thankfully not something I'll have to worry about in my lifetime. Sorry to whoever is reading this trying to figure out what on earth possessed us, best of luck! And remember we had the y2k and the 2038 problem (assuming the last one didn't wipe out civilisation), so it's not like we pushed all our problems onto you. Just the big ones.
Both events (leap second vs leap hour) are discrete changes; it's just that there will be a lot fewer (roughly 1/3600) of them in the leap-hour model. Leap-hours are already worked into computers since we already do it routinely 2x a year; hell, they could time it such that you just don't "leap ahead" one year or whatever (for the nations that observe daylight savings). You seem to be concerned about the computer side of things, and it seems like this is much easier for that side. Also, this is from the guys who literally timekeep for a living; if they say it's easier/better, it probably is.
I agree that it would be frustrating for some folks when it does happen - it would 'reset' the sunrise after decades of a slow shift - but....alright?
If you want to guarantee that something is going to break computers you just need to make sure it doesn't ever happen more than once in someone's career so they won't write the code to take it into account.
Edit: And if you think timezones will help here, it's going to be very annoying to figure out how many hours are between 3038-09-10 12:30 GMT and 3039-09-10 12:30 GMT.
Computers and other network connected devices, yes. Not the half dozen clocks that are all manually set twice a year or more, to one of those network clocks.
TL;DR, General Conference on Weights and Measures is trying to pass a resolution to increase the maximum allowable difference between UTC (societal time) and UT1 (astronomical time based on Earth's rotation), which currently has to be under 0.9 seconds and requires a leap second every few years.
The proposed increase would allow about an hour of difference between the two time systems to accumulate over centuries, after which a leap hour would be needed to bring computer systems back in line with astronomical time.
But as I like the benefits of daylight savings but do not like the weeks around when we have to switch. I've always thought it would be cool, now that we are mostly all using digital clocks, to just adjust the time each day by the 1 hour divided by the number of days in the year. Or even just make sunrise the same time every day and the clocks just add/subtract the time at midnight to make that possible (or as the article is suggesting smearing the time across the day).
>Or even just make sunrise the same time every day
The problem with that is that the time of sunrise doesn't just depend on the time of year and time zone. It's also affected by the viewer's latitude, longitude, and altitude.
In this hypothetical situation where I'm king of the world and can force this, I still wouldn't make it an individual time for every point on earth. I'd still recognise time zones for the most part. It doesn't have to be exactly sunrise, just in the ballpark you know.
Ethiopian time is basically this, though. And yeah of course sunrise varies, that doesn't mean you can't say 00:00 is the time when whatever percent of the year whatever percent has experienced sunrise, it's just a matter of what thresholds you wanna set
I agree, I have mechanical watches and analogue clocks but I'd be happy to give them up for a smoother daylight savings. Elderly people would be a problem though they can be very stubborn :P
Part of me is surprised that we haven't more effectively de-coupled the time we use in our everyday lives (in which dawn-dusk is important) from back-end computer time (which elapsed or exact time is important, things like stock trades and satellites and whatever). Although I'm sure I'm just not appreciating all the differences and problems that would introduce.
If I'm understanding this, we switch back and forth in a year, so we take roughly half a year to change the clocks by an hour. That's 365/2=182.5 days for this gradual shift amounting to 60*60/182.5=19.726 seconds of change per day. So every day the clock moves forward or backward by about 20 seconds, depending on the season. Effectively, this would just mean winter days are 20s shorter and summer days are 20s longer. This change would cancel out at the end of the year.
I love this idea, but there might be an issue with calender dates in different places being different for a few seconds every night. Deadlines and expiration dates on business contracts would have to be timed for sometime other than midnight to avoid legal confusion between areas that accept the gradual approach and area that accept the DST approach (and also areas that don't use any time modification system).
[OP] USCDornsifeNews | a day ago
Most are familiar with leap years, in which an extra day is added to February every four years to keep our human calendar on track with the Earth's orbit. A similar phenomenon also occurs on a smaller scale. Modern timekeeping systems have traditionally used "leap seconds" to keep us in sync but, as USC Dornsife physicist Vahe Peroomian explains, this system may change starting this month.
XkF21WNJ | 14 hours ago
Replacing it with a leap hour is probably the worst possible solution.
But thankfully not something I'll have to worry about in my lifetime. Sorry to whoever is reading this trying to figure out what on earth possessed us, best of luck! And remember we had the y2k and the 2038 problem (assuming the last one didn't wipe out civilisation), so it's not like we pushed all our problems onto you. Just the big ones.
I_Tell_You_Wat | 11 hours ago
Why is it worse? Worse for who?
Both events (leap second vs leap hour) are discrete changes; it's just that there will be a lot fewer (roughly 1/3600) of them in the leap-hour model. Leap-hours are already worked into computers since we already do it routinely 2x a year; hell, they could time it such that you just don't "leap ahead" one year or whatever (for the nations that observe daylight savings). You seem to be concerned about the computer side of things, and it seems like this is much easier for that side. Also, this is from the guys who literally timekeep for a living; if they say it's easier/better, it probably is.
I agree that it would be frustrating for some folks when it does happen - it would 'reset' the sunrise after decades of a slow shift - but....alright?
XkF21WNJ | 11 hours ago
If you want to guarantee that something is going to break computers you just need to make sure it doesn't ever happen more than once in someone's career so they won't write the code to take it into account.
Edit: And if you think timezones will help here, it's going to be very annoying to figure out how many hours are between 3038-09-10 12:30 GMT and 3039-09-10 12:30 GMT.
classicsat | 5 hours ago
No timezone, just GMT, one year apart in September. No leap years even so no accounting for that.
My maths say 31,536,000 hours, presuming no leap hour or other corrections. Or what time is over 100 years from now.
classicsat | 5 hours ago
Computers and other network connected devices, yes. Not the half dozen clocks that are all manually set twice a year or more, to one of those network clocks.
yboy403 | 18 hours ago
TL;DR, General Conference on Weights and Measures is trying to pass a resolution to increase the maximum allowable difference between UTC (societal time) and UT1 (astronomical time based on Earth's rotation), which currently has to be under 0.9 seconds and requires a leap second every few years.
The proposed increase would allow about an hour of difference between the two time systems to accumulate over centuries, after which a leap hour would be needed to bring computer systems back in line with astronomical time.
Nickools | 19 hours ago
Not exactly relevant to the article.
But as I like the benefits of daylight savings but do not like the weeks around when we have to switch. I've always thought it would be cool, now that we are mostly all using digital clocks, to just adjust the time each day by the 1 hour divided by the number of days in the year. Or even just make sunrise the same time every day and the clocks just add/subtract the time at midnight to make that possible (or as the article is suggesting smearing the time across the day).
I know it will never happen but would be cool.
SoMuchMoreEagle | 13 hours ago
>Or even just make sunrise the same time every day
The problem with that is that the time of sunrise doesn't just depend on the time of year and time zone. It's also affected by the viewer's latitude, longitude, and altitude.
Nickools | 11 hours ago
In this hypothetical situation where I'm king of the world and can force this, I still wouldn't make it an individual time for every point on earth. I'd still recognise time zones for the most part. It doesn't have to be exactly sunrise, just in the ballpark you know.
TrekkiMonstr | 2 hours ago
Ethiopian time is basically this, though. And yeah of course sunrise varies, that doesn't mean you can't say 00:00 is the time when whatever percent of the year whatever percent has experienced sunrise, it's just a matter of what thresholds you wanna set
GroundbreakingBag164 | 11 hours ago
We are not yet "mostly using digital clocks" that's the first issue. What about watches? The elderly that still have non-digital clocks?
Nickools | 11 hours ago
I agree, I have mechanical watches and analogue clocks but I'd be happy to give them up for a smoother daylight savings. Elderly people would be a problem though they can be very stubborn :P
I_Tell_You_Wat | 11 hours ago
Part of me is surprised that we haven't more effectively de-coupled the time we use in our everyday lives (in which dawn-dusk is important) from back-end computer time (which elapsed or exact time is important, things like stock trades and satellites and whatever). Although I'm sure I'm just not appreciating all the differences and problems that would introduce.
mundaniacal | 11 hours ago
Interesting.
If I'm understanding this, we switch back and forth in a year, so we take roughly half a year to change the clocks by an hour. That's 365/2=182.5 days for this gradual shift amounting to 60*60/182.5=19.726 seconds of change per day. So every day the clock moves forward or backward by about 20 seconds, depending on the season. Effectively, this would just mean winter days are 20s shorter and summer days are 20s longer. This change would cancel out at the end of the year.
I love this idea, but there might be an issue with calender dates in different places being different for a few seconds every night. Deadlines and expiration dates on business contracts would have to be timed for sometime other than midnight to avoid legal confusion between areas that accept the gradual approach and area that accept the DST approach (and also areas that don't use any time modification system).
curien | 8 hours ago
> there might be an issue with calender dates in different places being different
This is already the case. As I write this, it's 2026-10-09 where I am, and 2026-10-10 in Sydney.
classicsat | 5 hours ago
We tried that, up until the 1870s. It mucked up transport and commerce, to some disaster they had to invent standard time.