stdlib: Calendar library

[statementreply]

The numeric convention is tentative and open to discussion. It is somewhat arbitrary.

• month() is 1-12 for regular months, (N + 0.5) for leap month
• day() is 1-30
• largeHour() is 1-12 (子, 丑, ... 亥)
• smallHour() is 0-23 (子正, 丑初, 丑正, ... 亥初, 亥正, 子初)
• quarter() is 0-3 (4 reserved for old system)
• minute() and second() are 0-59 (60-99 reserved for old system)
• sexagenarianYear() and sexagenarianDay() are 1-60, 1 = 甲子
• Total year (60 N + 1) is 甲子 year
• Total day (60 N + 1) is 甲子 day
• Total year is close to the convention(s) used by some overseas Chinese communities
• Total day 0 is close to the beginning of total year 0 and total month 0 if we apply the current calendar rules backwards in time.

For example, under the tentative convention, 2020-01-25 05:42:00 庚子年正月初一日卯初二刻一十二分 would be:

Function	Value
彼年何年	2020
彼年何干支	37
彼年積何年	4717
彼月何月	1
彼月積何月	58341
彼日何日	1
彼日何干支	4
彼日積何日	1722844
彼時何時	4
彼時何小時	5
彼刻何刻	2
彼分何分	12
彼秒何秒	0

[LingDong-]

Looking forward to it! 👍

[statementreply]

The calendar data is generated with low accuracy astronomical computation of the time of new moons and solar terms. Due to the low accuracy and/or uncertainty of future leap seconds, calendar near the following days might be off by one day or one month.

1722401 (2018-11-08): maybe wrong 1st day of month (new moon at 1722401.0023)
1736606 (2057-09-29): maybe wrong 1st day of month (new moon at 1736606.0011)
1748269 (2089-09-04): maybe wrong 1st day of month (new moon at 1748270.0003)
1751164 (2097-08-08): maybe wrong 1st day of month (new moon at 1751164.0019)
1757572 (2115-02-24): maybe wrong 1st day of month (new moon at 1757573.0000)
1758015 (2116-05-12): maybe wrong 1st day of month (new moon at 1758016.0002)
1687201 (1922-06-25): maybe wrong leap month (solar term at 1687230.010, new moon at 1687230.8663)
1709998 (1984-11-23): maybe wrong leap month (solar term at 1710027.015, new moon at 1710027.8248)
1710973 (1987-07-26): maybe wrong leap month (solar term at 1711002.009, new moon at 1711002.8329)

Computation methods used for generating the data:

• Solar motion: formula from https://ssd.jpl.nasa.gov/?planet_pos
• Lunar motion: ELP/MPP02 (There happens to be an implementation of it on my computer :D)
• Light time and aberration correction: approximate formula
• Other relativistic corrections: ignored
• Precession: IERS conventions 2010
• Nutation: only the largest 18.6-year-period term
• Delta T: historical value before 2020, extrapolation for 2020-2028, fixed extrapolated 2028 value after 2028 (assume that leap seconds were to be abolished in 2028)

The resulting new moon time is accurate to ~1.5 minutes (uncertainty of delta T excluded), solar term time to ~20 minutes.

[LingDong-]

Awesome, thank you so very much! 👍