Harvey wrote: > Actually, it's rather easy--it all depends on the number 4. If you're > dealing with century years (100 years, ending with 00), then it's a > leap year if the century year is divisible by 400 (100x4); otherwise > it's not a leap year. If you're dealing with a non-century year, then > it's a leap year if it's divisible by 4; otherwise it's not a leap > year. (Always do the century test first--then the rule is simple.) > > Technically, the Gregorian "fix" very slightly over-corrects the > calculation, and so some have suggested the following preliminary step > (neither you nor I will be around when this might become necessary): > if you're dealing with millenial years (1000 years, ending with 000), > then if the millennial year is divisible by 4000 (1000x4), it is NOT a > leap year. But even that needs correction about every 20,000 years!
From my perspective, the above is awfully complicated vs a decimal notation of days! Finding the time elapsed between to dates is simple; for example, the elapsed time between the day 2458274 (today) and the day 1446496 is 2458274 - 1446496 1011778 days, that is, about 1011778 % 365.24218967 2770.15643 tropical years. Let us do it the other way around; pick two dates with a similar timespan using the BC/AD world standard. How many years, months and days have passed between them? Does the question even make sense? Forget about it. Let us try another one: how many days have elapsed? Can you show us how to perform the calculations in J? On Mon, Jun 4, 2018 at 11:59 PM, PR PackRat <[email protected]> wrote: > On 6/4/18, Jose Mario Quintana <[email protected]> wrote: > > Harvey wrote: > >> It's just the way it is. > > I merely intended to inject some reality. Despite all of the > discussions and arguments pro and con for various perspectives, > nothing is going to change current civilization regarding dates and > times. Everybody in these discussions can view it the way they want, > accommodate their programming to the realities of the world, and life > will go on. > > > ... Then, it gets even worse due to Greek and > > Roman influence: one year of 12 months, numbered from 1 to 12 (when > > the months are not explicitly named), and months of different numbers of > > days, numbered from 1 to the last day, > > If you want all the gory details, go to pages 987-1003 of volume 4 of > the 11th edition of the Encyclopaedia Britannica > <https://books.google.com/books?id=mP4tAAAAIAAJ&printsec= > frontcover&dq=editions:XzxQJyk6QDAC&hl=en&sa=X&ved= > 0ahUKEwil-YTBqLvbAhUrzoMKHRxCDaAQuwUIRDAF#v=onepage&q&f=false>. > > In brief, the old Roman calendar (as revised by Julius Caesar) started > in what is now March. It had months with the number of days > alternating between 31 and 30 (except the last month, February, had > 29). (Another way of looking at it is that the odd-numbered months > had the odd number of days, 31, and the even-numbered months had the > even number of days, 30, except for the last month, February.) The > months had numbers as their names. (We still have the names September > through December from the original 7th through 10th months. Using our > way of naming them, the 5th and 6th months would have been Quintember > and Sextember.) After Julius Caesar's death, Quintember's name was > changed to July to honor him. Later, Augustus Caesar had such a high > estimation of himself that he renamed Sextember to August and, in > order to also have 31 days in that month like Julius Caesar, he > "stole" a day from the last month, February, leaving it with only 28 > days. However, this created three 31-day months in a row, and so the > next several months were adjusted in terms of 31 and 30 days, so that > there would be no more than 2 months back-to-back with 31 days. At > this time the equinox occurred on March 25, but, because there was no > adjustment for the slight difference between the calendar year (even > with leap years) and the true solar year, the equinox had precessed to > March 21 by the time of the Council of Nicaea in 325 AD (when the date > of Easter was determined). At some point in history, the year > switched to start on January 1 rather than March 25 (which was the > original equinox date and the date of the Annunciation, 9 months > before Christmas). However, that was not the case in England, which > retained New Year's Day on March 25 and did not drop days at the time > of the Gregorian calendar reform. (One of the purposes of the 10-day > reform was to get the equinox back on the date it was at the Council > of Nicaea, March 21--by 1582 it was on March 11.) Finally, in 1750 > England passed a law that, beginning in 1752, 11 days (an extra day > had accumulated by that time) would be dropped from the calendar to > synchronize with the rest of Europe and that the civil year would > begin on January 1 rather than March 25. This created an interesting > situation because anyone who had been born between January 1 and March > 24 had not only a change in date but also a change in their birth > year. For example, George Washington was originally born February 11, > 1731 (O.S. = Old Style), but, with the calendar reform, that changed > to February 22, 1732 (N.S. = New Style)--the date we know nowadays. > It took a long period of time for people in both England and the > American colonies to become accustomed to this change and to change > all the various records and references to dates. > > > and the number of days of one > > particular month depends on awfully complicated rules! > > Actually, it's rather easy--it all depends on the number 4. If you're > dealing with century years (100 years, ending with 00), then it's a > leap year if the century year is divisible by 400 (100x4); otherwise > it's not a leap year. If you're dealing with a non-century year, then > it's a leap year if it's divisible by 4; otherwise it's not a leap > year. (Always do the century test first--then the rule is simple.) > > Technically, the Gregorian "fix" very slightly over-corrects the > calculation, and so some have suggested the following preliminary step > (neither you nor I will be around when this might become necessary): > if you're dealing with millenial years (1000 years, ending with 000), > then if the millennial year is divisible by 4000 (1000x4), it is NOT a > leap year. But even that needs correction about every 20,000 years! > > I won't be here, but with all this talk about time, I think it will be > very interesting to see what happens (or what workarounds will be > developed) when the Unix time counter (2^31) turns over to zero in > 2038--it's Y2K all over again! But that's also right around the same > time that Social Security is forecast to run out of money. Hmmm..... > > Harvey > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
