The COBOL days since 1601 epoch makes sense. Now that I think about it, if those functions don't work with negative numbers, then I imagine even the YYYY strings couldn't have years before 1601.
> Do you know which ones specifically? Would be interested to know this
> The COBOL days since 1601 epoch makes sense. Now that I think about it, if those functions don't work with negative numbers, then I imagine even the YYYY strings couldn't have years before 1601.
IBM mainframe COBOL has two operation modes, ANSI (where day 1 is 1601-01-01) and “Lilian” (where day 1 is 1582-10-14, the day on which the Gregorian calendar was first introduced). ANSI mode complies with ANSI/ISO standards for COBOL, Lilian is an IBM-only legacy standard. So, in Lilian mode, the earliest value for YYYY is 1582. I don’t believe negative numbers are accepted.
> Do you know which ones specifically? Would be interested to know this
This is the only one I've found:
https://docs.ogc.org/is/18-010r7/18-010r7.html#100
Temporal datum with Calendar ... and with TimeOrigin omitted so should be assumed to be 1875-05-20.