Show HN: Daystamp – like timestamp but for storing days (of birth, etc.) | Hacker News Reader

Show HN: Daystamp – like timestamp but for storing days (of birth, etc.)

github.com

2 pointsEGreg9 comments

this is using epoch "year=0, month=0, day=1", which is pretty unusual for day-based timestamps.

Turbo Pascal and its descendants use "days since dec 30, 1899": https://docwiki.embarcadero.com/Libraries/Sydney/en/System.T...

Microsoft Excel uses "days elapsed since January 1, 1900", except on Mac where it uses 1904 as a base (because Microsoft!): https://support.microsoft.com/en-us/office/date-systems-in-e...

Erlang uses this epoch

There is already Julian Days, though they are off by 0.5 since they've got day start at noon (Sun in the zenith).

This is for the usual Gregorian Calendar

There's nothing usual about extending Gregorian calendar backwards: you likely mean proleptic Gregorian calendar if you use 0000-01-01 as the start date.

Gregorian calendar had "holes" in it, and where those holes are, depends on where and when the date was noted.

To uniquely and usefully identify a daystamp, you need more than proleptic Gregorian calendar, since calculations will lead to unfamiliar historical dates from the past.

please elaborate

Da Vinci was born on 15 April 1452 of what was the incumbent Julian calendar.

How do you represent that date with your daystamp? In proleptic Gregorian calendar, it will be a different date, unfamiliar to people who know the birthdate of Da Vinci.

If a particular area changed from Julian calendar to Gregorian calendar in 1599 on July 1st, there will be no July 2nd-July 11th in recorded dates.

And so forth.

Calendars are hard because there is a lot of circumstantial data. Approach like you take with daystamp is useful for when they have been largely stable worldwide: since 1920 or so. There is no sense in starting the epoch at year 0 (or should it be year 1?), without being aware of the limits.

Sure, even proleptic Georgian calendar (extrapolated into the past) has its uses, which is why it has a name. But it helps to clarify where is it applicable.

Not to mention that various historic dates in particular regions are highly problematic wrt Year.

Today, now, here in Australia it's January 1st 2023 ..

but in countries and at times when the New Year didn't come about until some part of March or coinciding with Easter changes.. then it would be

January 1st 2022

.. as the New Year hasn't yet arrived.

I just realized you may not know of all the different Gregorian calendar adoption dates:

https://en.m.wikipedia.org/wiki/List_of_adoption_dates_of_th...

Reply on news.ycombinator.com