The Origin of the Archetypes Calendar

The Archetypes Calendar was devised by Peter Meyer during February 21-23, 2010, and its description was first published on this website on February 24, 2010. He revised the calendar during March 6-10.

The selection of the initial version of rules (i) and (ii) was influenced by the rules of the Meyer-Palmen Solilunar Calendar and by Karl Palmen's Rules for YLM Calendars. (The Archetypes Calendar is similar to a YLM calendar but is not a YLM calendar because all years in a YLM calendar have either 354, 384 or 385 days, whereas some years in the Archetypes Calendar have 355 days.)

Determination of the accuracy of the calendar (that is, how close dark moons are to month starts) relied on a FORTRAN function (ported into C) for the calculation of the moments of the lunar phases which was written by R. H. van Gent based on algorithms by J. Meeus.

When first published, the definition of the calendar used the following rules (where the range of y is 0-1802):
(i) A year with position y is a long year if and only if (y*L1) mod Y < L1
(ii) A year with position y is a leap year if and only if (y*L2) mod Y < L2
where Y = 3606, L1 = 1328 and L2 = 700. Following its announcement on the CALNDR-L mailing list on 2010-02-24 CE Karl Palmen pointed out that Y = 1803, L1 = 664 and L2 = 350 produce a calendar with the same structure (and that he had previously mentioned on CALNDR-L the 1803-year cycle of the same number of months and days as one instance of a class of calendars, though without specifying any internal structure). So, although L2 = 700 allows easier multiplication of y, the original numbers were replaced by the smaller ones in the interest of mathematical simplicity. The modified definition was published in a revised version of this web page on 2010-02-25 CE.

On 2010-03-04 Karl Palmen stated (in a message to CALNDR-L) that if the following rules were used (where the range of y is 1-1803):
(i) A year with position y is a long year if and only if ((y*L1) + (Y-1)/2) mod Y < L1
(ii) A year with position y is a leap year if and only if ((y*L2) + (Y-1)/2) mod Y < L2
then an ARC period of 1803 years would be a Helios cycle (so called after a contributor to CALNDR-L named "Helios" who, with Karl Palmen, first studied these). Since a calendar with a Helios cycle of years is more elegant than a calendar without, Peter Meyer decided to adopt the change suggested by Karl (replacing year with the more exact position-in-cycle).

Peter also decided to change the range of new year's days from December-January to a range centered on a day halfway between the winter solstice and the vernal equinox (as occurs with the Chinese Calendar), which is a better time to celebrate new year (not too early and not too late). This entailed a search for a new Julian day number which would produce this result. After much computation a number was found which produced a calendar (a) whose mean dark moon was close to midnight at the start of the month and (b) whose range of new year's days was not only the same as that of the Chinese Calendar but also such that new year's days in the two calendars mostly coincided (as noted above), a result as pleasing as it was unexpected. The revised Archetypes Calendar was published here on 2010-03-10 CE. On the following day the names of the first two months were changed from "Sol" and "Luna" to "Apollo" and "Diana", the Greek and Roman deities associated with the Sun and the Moon respectively.

In July 2011 Peter Meyer made the following changes to the Archetypes Calendar:

• The term "week" (of 9 or 10 days) was replaced by "tweek" (short for ten-day-week) to distinguish it from the usual 7-day week.
• The names of the 7th, 12th and 13th months were changed respectively from "Kronos", "Isis" and "Hades" to "Chronos", "Demeter" and "Persephone".
• The leap month was changed from the 6th month to the 10th month.