3

There already are algorithms that can be used to concert a date from different calendars into the Julian Day Number (JDN) and a JDN into the Hebrew calendar.

But how can one calculate the conversion from JDN to the Hebrew calendar in the old fashioned way, for all practical purposes only with a pen and paper? What would be the computational steps necessary, when one can only use +, -, *, /, modulo, and rounding?

If the JDN to Hebrew calendar conversion is too complex, is there a different calendar that could be used as a source which would allow an easier calculation?

Martin Handrlica
  • 147
  • 1
  • 1
  • 2
  • Why the downvote? – joriki Apr 06 '16 at 10:41
  • I don't think this is a question about mathematics; it is more about calendar systems. – Zoran Loncarevic Apr 06 '16 at 10:42
  • This is a complex subject with plenty of material available elsewhere eg http://aa.quae.nl/en/reken/juliaansedag.html (at section 8) – almagest Apr 06 '16 at 10:43
  • @almagest: I truly wish you had shown a different website, I tried to follow that one but ended confused and getting incorrect results. – Martin Handrlica Apr 06 '16 at 10:47
  • Please elaborate on the exact structure of each calendar. Users here are expected to know Math, not general facts. Also, please elaborate on what you've tried so far. Thanks. – barak manos Apr 06 '16 at 10:49
  • @joriki The downvote is because this question does not belong to this site. – 5xum Apr 06 '16 at 10:52
  • @5xum: But you have enough reputation to vote to close as off-topic. Why downvote instead of voting to close? – joriki Apr 06 '16 at 10:55
  • 1
    @joriki I didn't downvote or vote to close, actually. Just offering an explanation. – 5xum Apr 06 '16 at 10:57
  • 1
    @joriki: My guess is that the reasons for down-votes were more or less due to the fact that OP hasn't provided any mathematical aspect of the problem, so it's a little hard to refer to it as such. For example, what is the exact structure of each calendar? In addition, OP hasn't described any thoughts or attempts to solve the problem on his/her own. I actually think that the question does belong in this site, provided that all the above is properly described. – barak manos Apr 06 '16 at 10:59
  • May be interesting: http://www.amazon.com/Calendrical-Calculations-Nachum-Dershowitz/dp/0521702380 Both authors have a website with additional material. – Jean-Claude Arbaut Apr 06 '16 at 11:00
  • @barak manos: I believed have thought that [tag:calendar-computations] would focus only those who already understand calendars, also there are the linked algorithms. I also did not believe thy my own 8 pages of calculations would help the issue in any positive way. – Martin Handrlica Apr 06 '16 at 11:22
  • @Jean-Claude Arbaut: I have the book, it was what started me on the quest for an algebraic solution as it uses algorithms too. – Martin Handrlica Apr 06 '16 at 11:22
  • 1
    @all: I apologise if this question is bad or in a wrong place. But I planed it not only as a way to solve where I went wrong with aa.quae.nl/en/reken/juliaansedag.html, but also as a stepping stone to convert some other calendric systems that are problematic to me. If you cannot give any help then I shall delete the question, but I have to ask where is the right place to ask the question as I have not been able to find any major handling of the issue and just must ask how were the conversions done before the 1960s? – Martin Handrlica Apr 06 '16 at 11:22
  • Users here typically do not appreciate linked information. They do, however, appreciate to see your effort on the problem. It kinda makes them feel like their effort is worth it I suppose. – barak manos Apr 06 '16 at 11:26

1 Answers1

3

In order to construct an algorithm for converting a calendar date to a Julian Date we must first understand the calendar's structure. There are resources on the Internet for understanding the structure of Hebrew calendar e.g. Mathematics of the Jewish Calendar and Hebrew Calendar Science and Myths.

The main points are summarised as follows:

  • The calendar is a lunisolar calendar, with occasional leap years. The leap year cycle has 7 leap years in 19 years, spread as smoothly as possible (see also this answer). Leap years have 13 months and non-leap years have 12 months.
  • The start of the year is determined based on the molad, which represents the time of the mean new moon. The molad increments by 29 days, 12 + 793/1080 hours for each lunar month.
  • The start of the year may be postponed so that specific holidays do not fall on certain weekdays.
  • The length of most months are fixed, with a few exceptions. The lengths of all months can be determined given the number of days in the year.
  • The year number changes at the start of Tishri, which is the seventh month.

Thus, we can write down what we need for a conversion algorithm as follows:

  • Start of year
    • start of year requires molad
      • molad requires number of months elapsed from the epoch to the start of year
    • start of year requires postponements
      • postponements depend on the weekday and time of day of the molad
      • postponements also depend on where we are in the leap cycle
  • Lengths of months
    • lengths of some months require lengths of current year
      • length of current year requires start of next year
    • whether there is a leap month

Once we have all of these, we take the Julian Date of the start of the year, add the lengths of each elapsed month and then add the remaining number of days to reach our desired calendar date.

We can see these components in the algorithm linked in the comments and discussed further down on the same page:

  1. $c_0$ is a correction for the year number depending on whether we're before or after the start of Tishri.
  2. $c_1$ is the number of elapsed months up to the start of the year. We take 12 months per year and add the number of leap months (as seen in equation 630).
  3. $\mu$ is the molad corresponding to the start of the year, for which we take the molad epoch and add $\left(29 + \frac{1}{24} \times 12\frac{793}{1080}\right)c_1$ days (equation 634). $\nu_0$ is the JDN corresponding to the molad.
  4. $c_2$ is the date of the start of the year, and $\nu_1$ through $\nu_4$ correspond to the results of the postponement rules. Tabular methods would probably be easier for pen-and-paper calculation to work out the postponements as opposed to the formulae used on the page.
  5. $c_3$ is the number of days from the start of the year to the start of the desired month. Again, tabular methods would be easier for this step.
  6. $z_3$ is the number of days from the start of the month to the desired date.

Similarly, looking in the code linked in the original post allows us to pick out these same components.

Arcorann
  • 186