The angular calendar.

A year of 360 days, where the date is simply an angle – and the five or six days left over are held out of time.

The Gregorian calendar is a mess. The months are uneven, the weeks divide evenly into nothing, and working out what day of the week some future date lands on takes more arithmetic than it should. A clock at least has the excuse of an ancient pedigree. The calendar is just accreted habit.

Here is a cleaner idea. Drop the seven-day week for a ten-day one, and a date becomes a single running day-count from which both the week and the weekday fall out by inspection. Today is day °70: that is week 7, day 0. Define the regular year as exactly 360 days and start it at the spring equinox, and the date stops being a number and becomes an angle – which is why it earns the degree symbol. °90 is a quarter turn round the year, the top of the circle, midsummer.

The days out of time

The solar year is not 360 days but a little under 365¼. The angular calendar keeps the clean 360 and tucks the remaining five days – six in a leap year – at the very end of winter, treated as if they fall outside the calendar proper: a short end-of-year holiday before the year turns over at the equinox. On the wheel they sit just past the °360/°0 seam, detached from the ring, belonging to no week.

Leap years by observation

There is no leap rule to memorise – no “every fourth year, except centuries, except every four-hundredth.” Because the new year is pinned to the spring equinox, the calendar simply takes however many days fall between one equinox and the next. Most years that is 365, leaving five days out of time; roughly one year in four it is 366, leaving six.

So leap years arrive on an irregular schedule – usually every fourth year, but not always, and never on a fixed cycle you could write down in advance for all time. That irregularity is the point: the pattern is dictated by observation of the heavens rather than by decree. Today we would derive it from the same astronomical models that predict the equinox to the minute, but in principle a single careful observation each year is enough. The calendar can never drift against the seasons, because the seasons are what define it.

How far ahead could we compute the schedule? The astronomical part is surprisingly cheap. The exact instant of an equinox can be pinned to within a minute, across several thousand years on either side of today, by a single cubic formula plus a short table of two-dozen-odd periodic corrections – the whole thing fits on a page.^[1] Bring in the fuller planetary theory and the Earth’s orbit itself stays computable for tens of millions of years, until the slow chaos of the solar system – neighbouring orbits diverge on a timescale of about five million years – finally makes prediction meaningless.^[2]

The catch is turning that instant into a calendar day. For that you need ΔT, the widening gap between clock time and the real Sun caused by the gradual tidal braking of the Earth’s spin – a drift that grows with the square of the centuries elapsed.^[3] It can be modelled but not derived from first principles, because the slowing is shot through with irregular wobbles that only measurement can catch. Its sheer size passes a full day within about five thousand years; even the uncertainty in it reaches hours within a few millennia and a full day within roughly ten thousand. So the leap pattern is exactly predictable for centuries, fuzzy over a few thousand years, and genuinely unknowable beyond ten thousand or so – defeated not by the heavens but by the Earth’s own rotation. Which is precisely why the rule is best left to observation.

Counting the years

A calendar also needs a year count, and the Common Era has two problems: it is bound to one religion, and it cuts human history in two, with everything before the year 1 counted backwards. The Holocene, or Human Era – proposed by the geologist Cesare Emiliani in 1993 – fixes the second by adding 10,000 to every CE year, so that the whole sweep of civilisation since the last ice age carries a positive number.^[4] But it leaves the religious anchor in place.

A better epoch is already in every computer: the Unix epoch of 1970, arbitrary but quietly symbolic of the machine age that reshaped how we keep time. Count from there, then add the 10,000 of the Human Era, and the present year becomes 10,056. Since almost every date we use sits within the last few decades or in the future, we drop the leading ten-thousand and mark the omission with an apostrophe: ’56. For any year after 2000 the conversion is just “take the last two digits and add thirty” – 2026 becomes ’56 – with the one caveat that, because the year now turns at the equinox, the shortcut is off by one for the quarter of the year before then.

The degree symbol does double duty. After a number it marks a year, like a quiet CE: ’56°. Before a number it marks a date within the year: °70. Written together they share the symbol between them – ’56°70 – a complete date in five characters. Read aloud, the degree becomes “orbit”: fifty-six orbit, orbit seventy. Where the symbol is awkward to type, an asterisk stands in: ’56*70.

An ancient revival

This might read as cold and over-rationalised, but it is very nearly the calendar the ancient Egyptians kept: ten-day weeks (the decans, each keyed to the rising of a star), a regular year of twelve 30-day months, and five days at the end held outside the year proper – mythologised as the birthdays of the gods, and considered unlucky. Their one mistake was leaving out the leap day, so their calendar – first tied to the flooding of the Nile – slid one day every four years and wandered through the seasons in a slow 1,460-year circuit. The fix was even legislated, in the Canopus Decree of 238 BCE, and then simply ignored by the priests for two centuries until Augustus imposed it; the repaired calendar survives to this day as the Coptic and Ethiopian calendars, the oldest calendar system still in use.

The 360-day skeleton is older and stranger still. Mesopotamian scribes ran a schematic year of twelve 30-day months alongside their real lunar calendar for two thousand years – for rations, interest, and astronomy – and when Babylonian astronomers divided the ecliptic into twelve signs of thirty parts each, the 360 degrees of every circle since were born from it.^[5] One ideal day, one degree of the Sun’s annual journey: the angular calendar is not imposing geometry on the year; it is restoring the year to the geometry it created. The Maya made the same move independently – their Long Count turns on a 360-day tun, and their civil year set five unlucky Wayebʼ days outside the months, an ocean away from Egypt.

The revolution that tried it

The same design has also been tried in modern times – and it ran a great nation for twelve years. The French Republican calendar of 1793 had twelve 30-day months in ten-day décades, five or six sansculottides held outside the year as festivals, and a year that began at an equinox – the autumn one, as actually observed at the Paris Observatory.^[6] It even had this proposal’s leap rule, and its controversy: the founding decree promised both a leap year every fourth year and a year pinned to the observed equinox, which is a contradiction – the equinox does not keep a four-year lattice. The astronomer Delambre objected that observational leap years made future dates uncomputable; Gilbert Romme, the calendar’s author, proposed an arithmetic fix in May 1795 – and was condemned in the Prairial purges within weeks and died on the courthouse stairs. The fix died with him, and the calendar stayed observational until Napoleon abolished it from 1806.

The calendars that still watch the sky

Leap years by observation have the longest pedigree of all, because they never went away. It is how Iran keeps its calendar today. The Solar Hijri year begins at Nowruz, the vernal equinox, by a rule descending from Omar Khayyām’s Jalālī reform of 1079: if the true equinox falls before noon at the Tehran meridian, that day is New Year’s Day; if after, the next day is.^[7] That single sentence makes the calendar fully deterministic given an ephemeris – answering Delambre’s objection – and because it tracks the equinox itself, it can never drift against the seasons; its emergent 33-year leap pattern is more faithful to the Sun than the Gregorian rule it predates by five centuries.

And the switch can still be made deliberately, in the modern world: in 2014 the Bahá’í community moved its worldwide calendar from a fixed March 21 to the astronomically computed equinox at the Tehran meridian, effective at Naw-Rúz 2015 – its four or five intercalary days now varying year by year exactly as the angular calendar’s five or six would.^[8] Between Egypt’s skeleton, the Republic’s structure, and Iran’s living leap rule, nothing in the angular calendar is untested. What is new is only the synthesis – and the angle.

Criticisms and responses

A reform with this much history also inherits the objections, and they deserve straight answers.

“The ten-day week was tried, and hated.” It was. The Republic’s décade was resented from the first décadi to the last, and the resentment had a number: one rest day in ten where the old week gave one in seven. But that is an objection to a rest schedule, not to a week. The décade of 1793 was an austerity measure that happened to be decimal; nothing about a ten-day week dictates how many of its days are free. Give the décade two rest days and it is more generous than the week it replaces, not less. The lesson of 1806 is real, but it is a lesson about rest – the reform must be felt as a gift, not a confiscation.

“Days outside the week break the seven-day chain.” This is the objection that actually killed modern calendar reform. Auguste Comte’s Positivist calendar (1849) and Moses Cotsworth’s International Fixed Calendar (1902) put thirteen 28-day months around a “blank” day outside the week; George Eastman ran Kodak’s internal accounting on the thirteen-month calendar from 1928 to 1989, and the League of Nations seriously examined it alongside Elisabeth Achelis’s World Calendar.^[9] All of it died on the same argument, pressed from the League in 1931 to the United Nations in 1955: a blank day sets the true Sabbath wandering through the civil week, and no religious calendar could absorb that. The response is that the civil week need not claim to be anyone’s liturgical week. Religious communities already keep their own unbroken counts against civil calendars – the Hebrew calendar against the Gregorian, the Friday prayer against every workweek on Earth – and a seven-day liturgical cycle can run uninterrupted alongside a civil décade in exactly the same way. That answer was available in 1955 too, and it lost; this remains the political mountain, and it is honest to say so.

“Tampering with the week tears social life.” The strongest version of this comes from the Soviet nepreryvka of 1929, which kept the Gregorian months but staggered rest across five shifts so the factories never stopped – and found that spouses no longer shared a day off, machines no longer got maintenance, and by 1940 the seven-day week was back. But that failure indicts desynchronised rest, not reformed weeks: the angular calendar keeps everyone’s rest days the same, merely on a different period. The Soviet experiment is the control case that isolates the real variable – rest must stay shared.

What the record shows, in the end, is that the astronomy was never the contested part. Egypt’s structure ran for three millennia; Iran’s equinox rule runs now. It is the week where reform meets the world, and so far the world has said no. It may go on saying no – this is, admittedly, a quixotic proposal. But the ground is shifting under the old objections: nearly every clock and calendar is a screen now, and software – increasingly, software that can rewrite software – could carry out in a season the vast downstream adjustment that doomed every earlier attempt. The change is nearly hopeless, and worth hoping for anyway. It is a better system, and times do change.

References