The end of leap second?

Tech giants want to put an end to the leap second, their argument is:

The temporal tweak causes more problems – like internet outages – than benefits, they say. And dealing with leap seconds ultimately is futile, the group argues, since the Earth’s rotational speed hasn’t actually changed much historically. “We are predicting that if we just stick to the TAI without leap second observation, we should be good for at least 2,000 years,” research scientist Ahmad Byagowi of Facebook parent company Meta said via email. “Perhaps at that point we might need to consider a correction.”

What do you guys think about it? Read more here.

1 Like

Being a SI base unit, a second should always be static. Nobody should be allowed to correct the variation of earth rotation by fiddling with the second. Tech giants are exploiting physics to make their programmers happy…

Weren’t those the same companies that predicted the Y2K would be a catastrophe?
Also the same companies that never figured the software would ever pass the year 2000?
The same companies that said 640KB is enough for everybody?
The same companies that didn’t believe more then a handful computers was enough for the entire planet?
The same companies that destroyed competition, not on technology but simply sued them to death!

They are the last companies that I would ever listen too :grin:


I seem to have missed the part where the tech companies want to redefine the second; instead, I see they want the international earth rotation society thing (IERS) to stop introducing shim seconds to make the day line up.

There has not been one in five and a half years, and I’m happy with that.


They do have an excellent point though that at some point there’ll be a negative one, which has never happened before, and that will certainly cause some amount of disruption.

Then after that, it will occur again after gaps of multiple years, which gives ample time for new software to be written by people unaware of the issues from last time. We already have this problem with positive leap seconds. It will be much much worse for negative ones.

It is all very well saying, “well, the entire industry of software development will just have to raise its standards then”, but it is unclear how to actually achieve that across the entire industry.

It might really be easier and simpler for applications/sites that don’t care about smeared seconds during a leap second transition to just accept that.

I don’t think it’s entirely fair to keep going on about redefining the second while not acknowledging the real problems here. People who are willing to accept that during a leap second transition, a second may be smeared for 18 to 36 hours (hopefully there could be agreement on one standard) probably wouldn’t consider that to be “redefining the second” in general, only in particular. Those not happy to accept that do have other options for timekeeping.

Seems like a reasonable conversaton to have.

1 Like

As we already know, leap smearing is fiddling with the second, make it longer (or shorter, which was not happened before) in the smearing period. The definition of second is not modified de jure, but the actual implementation of second counting will make the second modified de facto in the smearing period, if leap smearing become universal.

On the other hand, ordinary leap second is fiddling with the definition of day: not 86400 seconds but 86401 (or 86399, not happened before). Since we already have leaps in the calendar level (leap year), leap second sounds more reasonable to me.

Leap years and leap seconds are nothing alike though. There are simple and stable rules for when a day is added to the calendar that have been largely unchanged for over 2,000 years. It also can never go backwards (remove a day)! Give 6 months of notice of a day being removed from the year and see how software copes! A lot of the concern about leap second handling is about the impact of a negative leap second.

The same argument of “this isn’t so different” could better be used to support leap smearing: Software is already used to seconds being fractionally longer or shorter because that is how time is corrected routinely by things like NTP. What software isn’t used to is a whole second being inserted (infrequent leap secnd) or entirely ommited (never yet seen but likely to happen negative leap second).


Now that the post has sparked an interesting discussion and coming on the heels of the leap second shift, what do you guys think of an artificial time subtraction?

It’s not like “The Catholic church announced the Gregorian Calendar to much rejoicing and all the problems went away.” It was not like that at all. There is also a thing where they might need to shim the year occasionally above the 97 leap days every four centuries Wikipedia - Gregorian calendar - Accuracy. Not to mention all the problems with getting governments to adopt it when they disowned it and later when it got re-adopted. (List of adoption dates of the Gregorian calendar by country - Wikipedia)

(/instantExpert timekeeping)

I massively favor making UTC a fixed offset to TAI over having leap seconds. If at some point the “accumulated” leap seconds add up to a minute or an hour society can choose to adjust via the regular (and in comparison very robust) time zone mechanism.