A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
From: Frank Reed
Date: 2022 Nov 20, 10:46 -0800
Tony Oz, you wrote:
"I do not see the point in the idea to abandon leap seconds."
The expression you're looking for is "I don't get it..." :). But why not flip it around: why do we have leap seconds in the first place? Who do they serve? What good are they? How much do they cost governments and businesses to implement? I am sure that you can see that a system without leap seconds is less expensive.
"Why all the fuzz? We do survive the "summer|winter time" jumps, what's wrong with the UTC then?!"
TWO key contrasts that you may be missing here:
Deterministic versus non-deterministic: Leap seconds are added based on the observed variations in the Earth's rotation, essentially due to the "weather", both in the Earth's atmosphere and in its mantle. That generalized "weather" is the cause of the variation of the Earth's rotation and it makes leap seconds unpredictable and arbitrary. If I ask you today, how much time will elapse between 23:00:00 UTC on 20 November 2023 and 23:00:00 UTC on 20 November 2048, you would not be able to tell me. We don't have a table of leap seconds for the next 25 years. The interval would be "about" 789004800 seconds (that's 25 years of 365 days each plus 7 Feb 29's multiplied by 24·3600 which is entirely deterministic). The exact value might be 789004805 or 789004809. We won't know until we get there as long as leap seconds continue to be inserted into the second.
Global standard versus local rules derived from the global standard: Around the globe, our civil times are defined in terms of offsets from UTC (many countries, especially those in the Commonwealth or with historical ties to Britain, call it GMT but in application they mean UTC). UTC, the global standard has a leap second problem. This is real. If local civil time authorities wish to adjust their offset from UTC, they can do so, they have done so, and they will do so, including by creating "daylight saving" rules. But that won't fix the global standard that is the basis of all civil times. Occasionally in discussions of this issue (by discussions, I mean the online "coffee klatsch" jabbering again), someone will say, "Hey why don't all those countries just switch their time basis to TAI (international atomic time)? That already exists, and it's got no leap seconds?? Why should us astronomers have to change our ways just to suit the rest of human civilization, huh?!" Doing away with leap seconds solves the problem with global time, easily and quickly. Civil times can still be legislated to offset from UTC as desired.
"(other than someone wanting to leech to the government money to "fix" the timekeeping software and/or hardware)"
Yes, you should always be skeptical that someone is just out to make a buck or collect some government pork. But there is almost nothing like that here. There are no lucrative contracts to be won apart from a few minor deals to update code for astronomical observatories. In fact, the effort required to work with leap seconds costs money as it stands and requires the attention of well-paid experts. They'll be able to put their time to other uses if leap seconds are eliminated.
"Why not to abandon the 29ths of Februaries, - the calender would become 'more simple and straightforward'..."
You got it, and that's the classic "coffee klatsch" reductio ad absurdum argument. That's why I posted a little parody of Elon Musk announcing that February would be removed from the calendar and its days annexed by an even longer "Mars month" ('cause, see, Elon is gaga about Mars, and lately he considers himself the world's leading expert on everything). March is "Mars month" so we can make it really long, maybe rename it to plain old "Mars" to get folks jazzed about colonizing Mars, and it will be exactly 59 days long forever. And why not?! Why not do that? Simply because there is no problem to solve. The rules for adding the leap day, February 29, are well-established and good enough for a few thousand years. The mathematical tools to calculate days between dates in the modern calendar, simply and quickly without reference to a bunch of conditional statements, are among the very first bits of code that everyone learned while setting up applications involving calendar dates decades ago. Do we need the leap-year cycle in the modern world? Do we need to keep the historical alignment between Easter and Passover and therefore the first Full Moon after the Spring equinox? Probably not, and it certainly isn't any pressing global need. But the system is well-established and "not broken".
If you're still puzzled, try reading the article from the "Facebook Engineering" team that I posted last week here (conveniently includes my Musk parody for your amusement):