Andrew Brown <atatat@atatdot.net> writes:
here's a list of points i think we could probably agree on.
(1) things changed. (2) having a list of said changes would be good. (3) assembling such a list will be difficult.
actually, the third point would probably be about as hard as it was to bootstrap the tz data, maybe simpler,
I'm not sure that I agree. The tz data was bootstrapped by ignoring most problems that occurred before 1970, which is the vast majority of the problems. People interested in historical use of calendars have no such luxury.
and we also don't have lots of "crazy" governments running around saying "gregorian! julian! reformed julian! gregorian!", so the churn ought not to be as high.
That problem is relatively limited now, as most people know and use the Gregorian calendar even if they are in a location (Iran, say?) where another calendar is official. However, the point of coming up with such a list will be to go back into history, where the problem was very real. Anyone who seriously wants to date events in the Middle Ages will know what I'm talking about: at times it seems that every author used a different calendrical method. And it's not just the Middle Ages: for example, three different dating methods were used in the logs of Captain Cook's voyages, depending on whether one thought the day started at 00:00, 12:00, or 24:00. It would take a lot of work to come up with a reliable history of calendrical usage.