8 ms·

I don't think we will ever need a negative leap second. We can just wait longer until dispatching the next positive one. The only situation in which we'd need a

by kortex 4y ago

I don't think we will ever need a negative leap second. We can just wait longer until dispatching the next positive one. The only situation in which we'd need a negative leap second is if Earth's rotation were consistently speeding up over many years. But we can tolerate some wiggle room (as in, several seconds) between UTC and TAI (since it's not in lockstep anyway).

adrian_b 4y ago
The problem is that there are hardware devices and software applications, which compute UT1 (the angle of the mean Sun) from UTC and from dUT1 (which are transmitted both on various communication channels, e.g. by radio stations) and all those expect that dUT1 is a number between 0 and 1 seconds. If dUT1 can become either negative or larger than 1, all such hardware and software must be reviewed and possibly upgraded, to no longer make assumptions about dUT1. Communication protocols may have to be changed, if there is no way to encode a negative dUT1.
- kortex 4y ago
  > If dUT1 can become either negative or larger than 1, Negative values should already be accounted for, afaict. > UTC is maintained via leap seconds, such that DUT1 remains within the range −0.9 s < DUT1 < +0.9 s. I can't imagine why it would ever be constrained to abs(dut) < 1, as that does not appear to be a hard spec anywhere, it's just the goal. But I could see some really stupid implementations, so maybe. https://en.m.wikipedia.org/wiki/DUT1 https://en.m.wikipedia.org/wiki/DUT1
- ncmncm 4y ago
  No. Literally no one uses UTC or leap seconds for that. The only people who need to pay attention to sun angle are astronomers, and they use their own system based on TAI. So you are arguing to continue rejiggering a complex and error-prone system that benefits exactly nobody.
dexwiz 4y ago
UTC and TAI are already over 30 seconds apart. So not using a negative leap second to keep them aligned isn't really a valid argument. On a geological scale, Earth's rotation is slowing down, but on a decade scale, it's still pretty chaotic. We could just as easily be having the reverse conversation. When we need a negative leap second, we may need a few in a row. So just waiting for a positive one to cancel them out doesn't really work. Using the logic of it will all eventually be a wash and we shouldn't use them at all is kind of the argument the article is making.

by kortex 4y ago

adrian_b 4y ago

The problem is that there are hardware devices and software applications, which compute UT1 (the angle of the mean Sun) from UTC and from dUT1 (which are transmitted both on various communication channels, e.g. by radio stations) and all those expect that dUT1 is a number between 0 and 1 seconds. If dUT1 can become either negative or larger than 1, all such hardware and software must be reviewed and possibly upgraded, to no longer make assumptions about dUT1. Communication protocols may have to be changed, if there is no way to encode a negative dUT1.

kortex 4y ago
> If dUT1 can become either negative or larger than 1, Negative values should already be accounted for, afaict. > UTC is maintained via leap seconds, such that DUT1 remains within the range −0.9 s < DUT1 < +0.9 s. I can't imagine why it would ever be constrained to abs(dut) < 1, as that does not appear to be a hard spec anywhere, it's just the goal. But I could see some really stupid implementations, so maybe. https://en.m.wikipedia.org/wiki/DUT1 https://en.m.wikipedia.org/wiki/DUT1
ncmncm 4y ago
No. Literally no one uses UTC or leap seconds for that. The only people who need to pay attention to sun angle are astronomers, and they use their own system based on TAI. So you are arguing to continue rejiggering a complex and error-prone system that benefits exactly nobody.

dexwiz 4y ago

UTC and TAI are already over 30 seconds apart. So not using a negative leap second to keep them aligned isn't really a valid argument. On a geological scale, Earth's rotation is slowing down, but on a decade scale, it's still pretty chaotic. We could just as easily be having the reverse conversation. When we need a negative leap second, we may need a few in a row. So just waiting for a positive one to cancel them out doesn't really work. Using the logic of it will all eventually be a wash and we shouldn't use them at all is kind of the argument the article is making.