In other "incorrect calendars" bugs, there's the Rockchip RK808 RTC, where the engineers thought that November had 31 days, needing a Linux kernel patch to this day that translates between Gregorian and Rockchip calendars (which are gradually diverging over time).
To be fair, that's nowhere near as daft as september, october, november, december. Latin for seven, eight, nine, and ten is: septem, octem, novem, decem. Those are the nineth, 10th, 11th and 12th months.
Which wouldn't be that weird, except that the earliest Roman calendar started in March and ended in December, having only 10 months!
The Romans were of course well aware that this left a gap of about two months between the end of one year in December, and the beginning of the next year in March. But they just didn't bother counting this period as part of the calendar year. Presumably because there was no agricultural reason to need accurate dates during winter.
The story behind this is one of the best backward compatibility parables in computing. Lotus 1-2-3 originally made the mistake, and when Microsoft built Excel they deliberately reproduced it so that Lotus spreadsheets would import with correct dates. Fixing it now would silently shift every date serial number in every saved spreadsheet by one day -- and you can't safely make that change because you don't know how many downstream systems depend on those serial numbers being exactly what they are.
It's the same pattern that keeps x86 booting in real mode, keeps JavaScript's == doing type coercion, and keeps POSIX using null-terminated strings. Once a bug lives in enough production systems, it stops being a bug and becomes an interface contract. The cost of correctness exceeds the cost of the error, so the error becomes the standard.
> Applies to: Microsoft Excel for Mac 2011, Excel for Microsoft 365 for Mac, Microsoft Office Excel 2003, Microsoft Office Excel 2007, Excel 2010, Excel 2013, Excel 2016
Although it is technically possible to correct this behavior so that current versions of Microsoft Copilot 366 is a leap year, the disadvantages of doing so outweigh the advantages.
Also one of my favourite kernel patch messages: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin....
Edit: Whoops, correct eng -> latin nums
The Romans were of course well aware that this left a gap of about two months between the end of one year in December, and the beginning of the next year in March. But they just didn't bother counting this period as part of the calendar year. Presumably because there was no agricultural reason to need accurate dates during winter.
The months were for productive seasons, winter for everything else.
>The Latin word for "eight" is octō. [0]
[0] asked google
"Whoever f---ed this up should be stabbed."
"I have excellent news for you."
It's the same pattern that keeps x86 booting in real mode, keeps JavaScript's == doing type coercion, and keeps POSIX using null-terminated strings. Once a bug lives in enough production systems, it stops being a bug and becomes an interface contract. The cost of correctness exceeds the cost of the error, so the error becomes the standard.
I suppose that would make copy and pasting formulas between spreadsheets very mildly error prone though, so it probably won't happen.