https://bugs.documentfoundation.org/show_bug.cgi?id=144699
Eyal Rozenberg <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords| |needsUXEval --- Comment #15 from Eyal Rozenberg <[email protected]> --- (In reply to Mike Kaganski from comment #14) > This is related to your "If the user enters a date from a period before > 1582, how should we interpret that date?" question. Why did you choose 1582? > I *guess*, because in some small part of the world, that year was the year > of adoption of the calendar. > Meaning, that in that small part of the world, > dates from manuscripts created in 1583 were *somewhat likely* based on > Gregorian calendar, while dates in manuscripts from 1581 were Julian. No no... 1583 is just the earliest adoption. That means that _everyone_ - in all places where Gregorian is used - will have problems with historical dates from before 1583, while with later dates, the problem depends on whether the country where the event occurred had already adopted Gregorian or not. > OK, > for that small part of the world, that October 1582 is a useful boundary, > and a French user working on French historic documents could take the dates > "as written", and have the correct chronology (including e.g. correct date > difference). For the greater part, and probably all, of the world - any date specified in Christian Enumeration, that is before 1583, is problematic, since it has multiple possible interpretations, so one needs to explain what kind of calendar is being used, or decide what the fallback is. Actually, even if it were just a "small part of the world", that would be a problem enough for us as a project, because LO, in principle, doesn't consider small-parts-of-the-world negligible. But like I said - it's everybody's problem, whenever someone uses early dates in LibreOffice. > But what good is that October 1582 for a user working on e.g. Bulgarian > historic documents? There, Julian was used till 1916; taking dates from > documents "as is" will give an error. For most of the world, October 1582 > means nothing in this regard, except for the need to do a double conversion > for some "arbitrary" time frame in respective region. So no, that is not the case. All early dates are problematic for everyone - to the extent that they use early dates at all. So, suppose you are working on some paper for school on some historical period, and your sources mention dates in years, or even years and months; and you make a chronology table in Calc. You may insert dates as their appear in your sources, which are very likely to be Julian, and Calc will interpret them as different points in time. > > And finally - do we want to discuss this at a UI/UX design meeting? > > Possibly; but if, then please call me. Of course, we would definitely not do this without notifying the relevant people. For now, maybe I'll add a keyword here. -- You are receiving this mail because: You are the assignee for the bug.
