- cross-posted to:
- memes@sopuli.xyz
- memes@lemmy.world
- cross-posted to:
- memes@sopuli.xyz
- memes@lemmy.world
You must log in or register to comment.
LibreOffice calc has
- 0 = 1899-12-30,
- 1= 1899-12-31,
- 2=1900-01-01
I wonder if they were trying to one-up excel.
So, are these the so-called 0-days everybody talks about on windows and linux?
(This is a joke for everybody not getting it)
I mean… They’re all kind of arbitrary in their own way, but one could argue that the excel time being invalid is actually useful because it serves as a null value - you know it’s incorrect, whereas on the other systems it could be the default date or it could be the actual date.
If you think that’s bad, look at what “1” means.
(And, honestly, at least windows’ “last big calendar change” and excel’s “start of the century when we wrote it” are reasonable points. The unix “let’s make it recent so we can fit an absurdly small unit as an integer!” Epoch is just… Weird.)
“start of the century when we wrote it”
This meme wouldn’t exist had they done that. But they chose the zeroeth, not the first, of January for Excel.
Yep. “1” is 12:00am on 1-Jan-1900
Numbers less than zero just give a weird error. Between zero and less than one give a nonsense date-formatted non-date.
Fucking Microsoft.
Now do 03:14:08 UTC on 19 January 2038!
I think we’re mostly using 64 bit machines now. Even loads of embedded stuff is running on 64 bit processors now.
There will still be a lot of old software and hardware that needs updating before the 32 bit Unix time overflow
Hopefully it’s another Y2K nothing burger. (Which was largely because a lot of people prepared for it)
Yeah, I got my first office job thanks to Y2K. An enormous amount was spent fixing it, with some of the fixes needed years before 2000-01-01, for example systems that projected into the future
Biggest problem I saw was a program that stored 1999 as 99 and displayed “19”.year
So when set to January 2000 it showed 19100. Its calculations were fine, just its display and reports were wrong
Agreed, most of the actual problems seemed to be in reporting. I saw some cobol stuff that went to 1900. There were a few things where 00 wasn’t an option, But mostly it was just really heinously written stuff that wasn’t expected to be in service even in the '90s.
I’d never noticed that in this order, it’s almost π!
Off by 0.001 and some change; that coincidence is going to haunt me.
the actual last 32 bit micro happens in 3:14:07 but that’s the wrong way :)
Easy. Just be 64-bit. (Or unsigned, but that’s not as good.)
I work in the accounting team of my company, doing python development, and a lot of the sources I get came from the accountants that only use excel and every once in a while the date columns came completely fucked with that integral number excel uses sometimes and is a pain in the ass to translate to datetime.
🧐: 0 is the origin of time, the big bang (if you believe in that kind if thing)
The problem then is figuring out when earth (and then human) time starts, but we can just add some arbitrary offsets that feel right and everyone agrees on.
Time is a pain in the ass and unfortunately not this simple, since it’s affected by energy and doesn’t pass at constant rates. The amount of time that’s passed since the Big Bang is different on earth than it is on a neutron star, for example, so one fundamental reference time isn’t possible the way it is for other units.
Listen. Let’s just do it simple. Everyone sync their clocks with mine right now. It says it’s 10:57 at the moment, but we’ll just subtract that in the future going forward. Ready? 3, 2, 1, Go! Now we know where 0 is, whew. Problem solved, scientists!
Edit: wait, now it says 10:58…
January zeroth, 1900. A red letter date.
On Linux, 0 often works out to 12/31/1969 after local time zone conversion…
0 means now 🙂
