From Newsgroup: comp.misc
There has been some argument for a while about whether itrCOs worth
continuing to maintain leap-second adjustments to keep our atomic
clocks in sync (roughly) with the EarthrCOs rotation. Every time a
leap-second is added as the EarthrCOs rotation slows down, itrCOs a bit of
a headache for various organizations to babysit their computer systems
through the change, and fix any glitches that arise. And the rate of
addition of such adjustments is only going to increase over the coming centuries.
Now, for the first time ever, the EarthrCOs rotation is (temporarily)
speeding up for a change, so a leap-second will need to be
*subtracted* this time. When a leap-second is *added*, that day has
86401 seconds, so the clocks tick over from 23:59:58, 23:59:59, to
23:59:60, and then to 0:00:00 for the new day. When it is
*subtracted*, the day will only have 86399 seconds, so the clocks go
from 23:59:58 straight to 0:00:00, without stopping off at 23:59:59
along the way.
Nobody is ready for this, and it seems likely to cause even more
problems that have not been encountered before.
So a vote will be taken in October to put a stop to further
leap-second adjustments; if this goes ahead, UTC (the atomic-clock
time reference) and UT1 (referenced to the EarthrCOs rotation) will be
allowed to diverge beyond their current maximum difference of 0.9
seconds in future, until the discrepancy is (according to the article)
as much as one hour.
What happens after that? Presumably somebody will come up with a
bright idea for what to do next, in the centuries between now and then
...
<
https://www.computerworld.com/article/4220034/leap-second-proposal-will-keep-software-stacks-in-sync-2.html>
--- Synchronet 3.22a-Linux NewsLink 1.2