2099?

[Aitch]

Some years back, I supported a Vax cluster. Before we (OK, I ) got around to synchronizing all the clocks, there were a couple of occasions when someone tried to edit a file on another machine and got a message along the lines of Cannot edit file - creation date is in the future.

 

[DavidS]

Why would they stop making their calendars on 2099?

While a better reason has already been posted for that particular "they", another reasonable basis is that 2100, like 1900 but unlike 2000, will *not* be a leap year. That's such a long time off that it may not be worth the trouble and overhead to implement special century-year handling (leap year only if divisible by 400) for every date calculation. The few applications that care about dates 90 years ahead can take care of it on their own.

 

[Ysidro]

And here I was hoping this would be about the Marvel 2099 comics. I really liked Spider-man 2099 too.

 

[Magic 9-Ball]

Well, the Hebrew calendar is 5768, so they must have solved that problem. Maybe buy MS Windows in Israel?


But if a new messiah came, would they reset the calendar to zero again?

 

[drkitten]

While a better reason has already been posted for that particular "they", another reasonable basis is that 2100, like 1900 but unlike 2000, will *not* be a leap year.

Won't it? I know that's what the strict reading of the Gregorian calendar rules would say, but I see no reason to belive that those rule will actually be followed; it would not surprise me to see that the Federal Government decides in 2066 that it will be cheaper to make 2100 a leap year than it would be to rewrite all the calendar-planning software in use throughout industry.

Case in point -- a lot of industry wonks have decided it will be easier/cheaper to abandon the practice of astronomically-motivated leap seconds and just let the sky drift out of sync with computer clocks.

 

[Nick Bogaerts]

Case in point -- a lot of industry wonks have decided it will be easier/cheaper to abandon the practice of astronomically-motivated leap seconds and just let the sky drift out of sync with computer clocks.

Absolutely. As I hinted at above, the current choice of the POSIX and NTP standards of using UTC over TAI is completely idiotic. As a result of taking leap seconds into account, POSIX timestamps are both discontinuous (some timestamps are not valid dates) and ambiguous (some timestamps refer to multiple dates). And to make matters worse, these discrepancies aren't even deterministic.

As a result, you can't calculate the time elapsed between now and a timestamp in the future, because you can't know how many leap seconds will have occured and be silently discarded.

Ironically, this decision was taken precisely to make it easier for computers, at a time when computing a human-readable date would have been an expensive operation. This is no longer the case today. All you need is a catalogue of all leap seconds inserted and deleted (not difficult to update in this internet age), and you can calculate a human readable UTC date from a TAI timestamp. And no more ambiguities.

 

[Ron_Tomkins]

2099?

No.

 

[Whiplash]

If only Harry Seldon was here, he could calculate exactly when the end of times would be, as well as a great many other things!

 

[Jon_Stripe]

To thread starter.

I don't think so. They probably figure you aren't going to be making any plans for March 2, 3000.

BTW. I like your avy.