Post by @sundogplanets@mastodon.social
Just had a 5:30pm surprise CBC interview about leap seconds: https://www.cnn.com/2026/10/13/science/leap-second-elimination-time-standard
I didn't really have much of an opinion on leap seconds until I talked through it with the journalist. But wow, leap seconds are such a bad idea! What a silly way to get our timekeeping to work with the imperfect, constantly changing orbital system we live in.
@sundogplanets I've always thought they are a nice compromise. We have TAI for precise timekeeping and UT1 for keeping precisely in step with the day/night cycle. Then UTC gives civilian time the best of both worlds, in a way. The people who object seem to be stock market trader types who think that civil timekeeping should revolve around their obscure needs.
@sundogplanets Programmers hate this one weird trick! (When they are working with time manipulation.)
@sundogplanets I don't much like the leap hour solution, though. That's a whole time zone's worth of discrepancy! I'd have preferred a leap minute once or twice per century or whatever it works out as. That should be infrequent enough to avoid any negative ones, while still keeping midday close to midday.
@sundogplanets I used to work on astrodynamics software and leap seconds, the TT/UT delta, and the empirical earth orientation parameter tables were the bane of my existence. It always felt like I was trying to use astronomical time and space coordinate systems that were invented in the renaissance to fly spacecraft in the 20th and 21st centuries. (I never worked for SpaceX or so called "commercial" space :) )
@eleanor My pulsar-measuring colleagues also passionately hate leap seconds! Leap seconds are not good for astronomy.
@sundogplanets Back in my days, I worked on SWIFT system (international financial transations). The system ran local time and transactions timestamped to local time (aka bank's business day).
Spring forward was not a problem. One transaction at 01:59 and the next at 03:00. But fall backaward was a problem because you could have transaction at 01:59 and the next one at 01:00 so time stamped before the previous one. Solution: shutdown for one hour so next transaction would arrive at 02:00.
@jfmezei@cosocial.ca@cosocial.ca @sundogplanets@mastodon.social@mastodon.social #JustCurious was there a good reason to use local time instead of #UTC?
One of our customers had distributed systems that were running local time, were crossing borders and timezones. Web apps for the service UI had their own timezone settings — and of course, that should all make sense. Eventually we got there. For the next generation system we all agreed to use UTC internally.
@docwhozee older computer systems uses local time. And for a bank due to « business hours », local tine is important. If end of day is defined as 17:00 and a payment arrives 17:01 then the payment is processed the next business day. So when you send a payment to Chinese firm via Chinese bank, you need to know when their local end of day is to ensure payment arrives before their end of day.
@sundogplanets Systems connected to NTP have to be able to adjust their time by a certain amount in either direction when their own clock diverges from NTP time. So any software should be engineered to support time changes in either direction.
The BIPM' draft resolution C:
https://www.bipm.org/documents/d/guest/cgpm-2026-draft-resolution-c-en
Note, they are also entertaining the redefinition of length of a second (correctly defined as time needed for 9 192 631 770 Cesium 133 oscillations. Much, including the metre depends on "second".
@sundogplanets When astronomer discovered the blumb line at Greenwhich used to align their telescope was not "perfectly" vertical, instead of changing time to match the correction to alignement of stars relative to Greenwhich, they moved the prime meridian by ~102m or 5.3 seconds longitude east. (Instead of changing all clocks by 0.35333 seconds).
@sundogplanets Very early view on the topic after too short a consideration? We have clocks like TAI, UT1 besides UTC and a full set of solutions in metrology and software engineering. This kind of complaint is a bit like "Why should I consider EMC when dabbling in electronics, or toxicology in making toys?"
@sundogplanets I didn't have an opinion on this, and going into the article I couldn't imagine that I would come out of it with particular thoughts in either direction. but the arguments against are just so bad! "Keeps humans in time with the natural world*... after a thousand years, The difference would still be one that no human would ever detect. maybe if we jumped a thousand years in a day. " punting the problem for future generations"? No! they would have the problem anyway!
I like the quote they chose for me at the end of this article: https://www.cbc.ca/news/world/leap-second-decision-9.7381049
"This is a real human problem that we've been struggling with forever, as long as we've been watching the stars move across the sky and tracking when the sun rises and sets, and trying to get the patterns right," Lawler said.
"It's always going to be a little bit wrong, and I think we just have to accept that."
That's right, everybody. No clocks. I want chaos! Hahaha I'm going to get some weird emails...
I don't mind if they miss a few leap seconds. I'm getting too old for leaping, even for a second.
@sundogplanets let's get rid of calendars too! We can switch to a system that is based on "when I feel like it" lol
That's quite an article. Crazy to think a one second adjustment can cause things to crash. It makes sense.. but also makes systems appear so fragile
@sundogplanets words to live by. "It's always going to be a little bit wrong, and I think we just have to accept that."
@sundogplanets Simple solution, everyone switch to UTC, disambiguate time from the which side of the earth is facing the sun. No more timezones!
@sundogplanets software helped cause negative leap seconds (with a little help from gas turbines), so software can accept negative leap second consequences
I loved this; thank you! I'll offer:
"All models are wrong, but some are useful” -- attributed to George E. P. Box
Clocks are just a model of the earth's rotation around its axis and its sun.
Sundials are accurate: the gnomon's shadow indicates the exact position of the sun; on any given day the minutes and hours are relatively squishy or stretchy.
Clocks, on the other hand, are merely precise -- unless they have stopped, in which case they are accurate twice a day.
🌞 ⏰ 🤪
@sundogplanets See? This is what happens when time gets woke! No rules. People can choose what time it is. Don't like the time? Change it. "I'm not late for work! You're all early!" Even the units are up for grabs. Instead of hours, I'm going to call them blorps. In five blorps, I can go home.
@sundogplanets So, it looks like you're supporting the idea of a BIG clock change once in a generation instead of multiple fiddly little changes. Just like Y2K caused generational stress for tech folk.
If you're quick, get out there and NAME this approach in interviews.
#goatzilla
@sundogplanets If the clocks and earth are drifting apart, can't we just get anchors to keep them together? Or just glue the clocks to the ground?
...and I now have a live radio interview about timekeeping scheduled tomorrow at 12:30 on 880 CHED in Edmonton and QR 770 in Calgary. Hahaha!
