Some popular genealogy programs calculate durations (the time elapsed between two dates) as a convenience. For example, informing you of a calculated length of life, and/or the age of an individual on all that individual's dated events. This seems problematic...
Take for example the dates March 24 1598 and March 25 1599, and consider the duration between them.
If the dates were recorded in Spain, then the duration is a year and a day. If the dates were recorded in Scotland then the duration is only two days. This is because in Scotland, the day of the new year increment was March 25th. It's apparent that we needed to know that a the year incremented between those two dates. I propose that more generally, in order to correctly calculate the duration between two dates, one needs to know whether the day of the new-year-increment for all years between the two dates (inclusive) changed. When a day of new-year-increment was changed, it created a short year. For example, in the year 1600 Scotland changed the day of new year increment to January 1st, therefore the year 1599 in Scotland only lasted 283 days.
Consider next an individual that was born in Nova Scotia on Jan 1 1700 and died on Jan 1 1760. One might be disposed to subtract 11 days to account for the switch to the Gregorian calendar system, but the true duration was exactly 60 years excluding the end day because this part of Canada reverted from Gregorian back to Julian then back again to Gregorian, within that lifetime, the added and subtracted days cancel out. From this I also propose that in order to accurately calculate the duration between two dates, one needs to know about any Gregorian/Julian calendar system changes between the two dates.
As a final example, consider two dates from Sweden: Feb 29 1712 and Mar 1 1712. The duration is up to 3 days. This is because Sweden decided to have a February 30th that year. From this I also propose that in order to accurately calculate the duration between two dates, one must completely understand any exceptions in the calendar system we are calculating.
These three examples seem to show that calculating durations, and by extension, sorting events by date, is very impractical to do accurately without asking for confirmation of the necessary requirements described above. I believe that attempting to confirm these requirements is impractical, as it's impossible to know if any particular village or hamlet got the memo, and immediately followed the rules.
Is my conclusion correct?