> Hospitals would just shut off all of their computers during this time to deal with it.
FWIW, there are many things that deal with leap seconds that way too. Too much risk of ending up in a difficult to fix or silently corrupt state, while coming up from a reboot is highly tested and known to work.
The cost of leapseconds is quite significant.
> but if you've ever written code that deeply interacts with time, you'd know how difficult it is to get right.
Good odds that even if someone has that they got it wrong and don't know-- especially when it comes to leapseconds as they're fairly hard to test esp. with distributed systems and infrequent enough that you may not realize the cause even when you've suffered from an issue.
FWIW, there are many things that deal with leap seconds that way too. Too much risk of ending up in a difficult to fix or silently corrupt state, while coming up from a reboot is highly tested and known to work.
The cost of leapseconds is quite significant.
> but if you've ever written code that deeply interacts with time, you'd know how difficult it is to get right.
Good odds that even if someone has that they got it wrong and don't know-- especially when it comes to leapseconds as they're fairly hard to test esp. with distributed systems and infrequent enough that you may not realize the cause even when you've suffered from an issue.