God forbid somebody suggest a solution to a problem.
> "As an industry, we bump into problems whenever a leap second is introduced."
Their suggested solution is:
> "As engineers at Meta, we are supporting a larger community push to stop the future introduction of leap seconds and remain at the current level of 27, which we believe will be enough for the next millennium."
Most folks are rightly pointing out that there are many other solutions that we could introduce, which wouldn't have the downsides of UTC drifting away from our wall-clock time. Facebook didn't even discuss those solutions though.
An example possible other solution: programmers (especially programmers at Facebook) should stop using UTC and should instead use TAI (which is literally UTC but without leap seconds). Indeed, using UTC time as anything other than a wallclock time should trend towards the norm. Even though this is a clear tru-ism, seeing it adopted would be way harder (due to language inertia and habits) compared with just changing the standard (nevermind that changing the standard would break a ton of logic built around expectations like "midnight UTC falls on an integer multiple of 86400").
The purpose here is for things to get better, not worse.
We do lots of things that are hard because doing them right is often more important than doing them easily. Switching from ASCII from UTF-8 was a pain, but we did it. Software upgrades are a pain. Security infrastructure is a pain. Timezones are OMG such a pain. But in all those cases, we collectively said “welp, guess we’ve gotta do it”.
And what Facebook notably didn’t propose was a way to actually make this happen. Who’s going to project manage the global coordinated effort to migrate the planet to Facebook Time? That sounds like much more work than them just fixing their time handling.
I agree that we don't just need to do something because fb wants to do it, but it might actually be worth having a discussion over.
But I don’t feel like this rose to the level of an actual proposal. It was very short to assert a claim with such wide-reaching implications as “we should start ignoring leap seconds”. As such, I don’t think it calls for an in-depth rebuttal.
Consider:
Proposer: “It’s time to leave Unicode in the past. It requires us to update every part of our system to deal with UTF-8 strings instead of much simpler ASCII, and we’re spending a lot of resources. Because it’s so hard and expensive, we should all use well-tested ASCII code. People who want to interoperate with our system can just rename themselves to use the Latin-1 alphabet.”
Everyone else: “No.”
Yes, it is hard, and people have spent a lot of, ahem, time and money to figure out how to manage this at scale. But there are real-world-tested approaches to dealing with the issue, and I firmly believe it’s better to work out and coordinate on the remaining rough edges than to throw the whole thing away to make a handful of engineers’ jobs easier.
"A small number of countries however, including Russia and the United Kingdom, want to keep the leap second."
UK used to decide on which date to switch to and from "summer time" every year. "The admiralty will take up the question in coming weeks."
Source: I managed to make a transition from UTC to TAI in a broker trader in 2009 when I was fresh out university.
"Let's just kick the can down the road" isn't solving anything.
The world is messy, life is messy, and so is everything else. If that’s too much for poor programmers, they need to find another job.
I obviously meant “re-engineering”, but “de-engineering” works as well, in this case…