Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

So what's the alternative? The leap second is capturing an actual skew between uniformly-ticking clocks based off of atomic decay statistics and the rotation of planet Earth. Stop updating them, and we end up with the sun going down for equatorial latitudes at 3PM eventually.

This feels very can-kicky, even if the can can be kicked a thousand years down the road. I don't think more can-kicking is really the best solution.



Once the discrepancy exceeds 30 minutes you change the time zone offset.


So their complaint is that one-second skews happen infrequently enough that they are always a headache, and we want to "solve" the problem by... Replacing it with a less frequent time modification that introduces larger delta?

I've been on the front lines of adapting code for a novel timezone change. It wasn't pretty.


Globally, time zone data changes all the time (more often than leap seconds) and thus software is better at dealing with it.


It is not, ever. Doing it every year or two imposes just as much disruption every year or two. A disruption once a century is at least 50 times less disruption.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: