Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Did you ever wonder why December has 31 days? (1997) (csail.mit.edu)
120 points by soegaard on April 15, 2014 | hide | past | favorite | 64 comments


And it's wrong. The email is from 1997, but now you can look it up on wikipedia: http://en.wikipedia.org/wiki/Julian_calendar

Short version - Julius didn't name the month after himself, the roman Senate did. There's good indications that July and August had 31 days before the reform, and February has less days because it was the last month of the year.


Not only that, but the Roman calendar was a mess from whole end of the Republic going on for the last ~100 years. The dates for some of the events listed in the Roman civil wars of Caesar had drifted from solar time pretty badly at the end (40'sBCE) and were off by month from the solar year. (The Romans has used a lunar calendar previously) In fact, members of the priesthood who determined the festival days that would bring the lunar calendar in sync with the actual solar time were often bribed to extend the year which would allow the consuls longer times in office. One of the first things Caesar did when he became the sole ruler of Rome was to reorganize the calendar around the solar year. He did so with the help of a Greek Astronomer from Alexandria. I believe the Greek astronomer made some faulty calculations regarding leap years, which is why the Gregorian calendar was instituted to fix the mistake in the 1500's.


I have thought February would make more sense as the last month of the year. But I guess that's because I live in the Northern Hemisphere and having the year start right after Winter starts seems like a bad combination.

EDIT: Apparently, 90% of the world's population lives in the Northern Hemisphere. Source: http://bigthink.com/strange-maps/563-pop-by-lat-and-pop-by-l...


You can also kind of tell March used to be the first month by how september-december are named (7 8 9 10)


This is a cute story, but it lacks all sources, and it contradicts Wikipedia [1] [2]:

> Commonly repeated lore has it that August has 31 days because Augustus wanted his month to match the length of Julius Caesar's July, but this is an invention of the 13th century scholar Johannes de Sacrobosco. Sextilis in fact had 31 days before it was renamed, and it was not chosen for its length (see Julian calendar). According to a senatus consultum quoted by Macrobius, Sextilis was renamed to honor Augustus because several of the most significant events in his rise to power, culminating in the fall of Alexandria, fell in that month.

The fact that February has the fewest days makes sense considering it was the last month of the Julian year (March was the first, hence July and August were called quintilis and sextilis, for fifth and sixth, respectively). It doesn't quite explain why the 31-day months are distributed unevenly over the year, but if you are going to take days off, it makes sense to take them off the last month of the year.

1. https://en.wikipedia.org/wiki/Julian_calendar 2. https://en.wikipedia.org/wiki/Augustus#Month_of_August


> March was the first, hence July and August were called quintilis and sextilis, for fifth and sixth, respectively

An even better example is the actual numbered months:

- Septum (7)

- Octo (8)

- Novem (9)

- Decem (10)


Contradicting Wikipedia does not equal a lack of truth.


No, but information on Wikipedia tends to be accurate and based on reliable sources -- when a person contradicts Wikipedia and fails to provide any sources of their own, they are usually wrong. Not necessarily, of course, but often enough to say the burden of proof lies with the person making the unsubstantiated claim.


Uh. There was a source provided: 1911 Encyclopedia Britannica.

I wouldn't call that more reliable, but it's still as valid a source as Wikipedia.


The 1911 Britannica is not the source provided. The source is a hypothetical entry in an encyclopedia that most people don't have. That is not as valuable a source as Wikipedia because I cannot easily verify that the 1911 Britannica corroborates the facts in this email.



Just because you are lazy and unwilling to verify the source doesn't mean it's not a source.



You're right -- my mistake. I missed that somehow.


"if you are going to take days off, it makes sense to take them off the last month of the year."

It's sort of understandable to take them off the last month of the year, but I think it makes the most sense to take them off the longer months first.


Backwards compatibility.


I don't follow. If you're already resizing months, what backwards compatibility is preserved by moving days from the last month rather than elsewhere?


Nature abhors a vacuum.

Internet abhors a claim contradicting Wikipedia.


To me, the most mathematically logical way to split up the calendar would have just been something like:

  01 02 03 04 05 06 07 08 09 10 11 12
  31 30 31 30 31 30 30 31 30 31 30 30 -- common year
  31 30 31 30 31 30 30 31 30 31 30 31 -- leap year
- All months 30 or 31.

- Leap month alternates between 30 and 31.

- Never have more than two consecutive months with the same day.

- Only time two consecutive months have the same number of days are Jun/Jul (30) and (Jan/Dec (31) or Nov/Dec (30)).

- Puts leap month at the end to avoid offsetting only part of the year.

- Any given 3-month span consists of either 91 or 92 days.

- Any given 6-month span consists of either 182 or 183 days, except for spans involving leap month and January, which have 184.

- First half (Jan-Jun) always consists of 183. Second half consists of 183 (leap) or 182 (common).


This is close to what was done, February is the last month of the year in the Julian calendar. Only two consecutive months have the same amount of days (July/June) and the last month varies by leap year (adding an extra day at the end of the year)

It's only weird that the last month got 28 due to doubling of the 31 day months instead of 30 day months


I have always felt that the 31-30-31-30-31 pattern from March through July, Repeating August through December, and then starting again in January was significant. I once wrote an extremely short assembly language routine (PDP-11 I think) for calculating the ordinal date from the calendar date that used simple integer arithmetic (multiply by 153 and divide by 5, I think) to calculate the days in the whole months starting in March.

30.6 rounds to 31

61.2 rounds to 61

91.8 rounds to 92

122.4 rounds to 122

153.0 rounds to 153 etc.

See http://www.cs.utsa.edu/~cs1063/projects/Spring2011/Project1/...


Even easier. 13 months of 28 days with one or two "leap day" per year make up 365-366.


Which is like to a Lunar Calendar with 1 or 2 holidays at the end. So, you have 13 months, each month is 28 days. At the end of the year you get 1 or 2 days off free!


That is a bad idea because number 13 is unlucky. Still, we could fix that by skipping the 13th month and saying the last month is 14.


Surely it would also make more sense to start with 30 and alternate, rather than having months 6 and 7 both be 30? Then it almost fits an alternating pattern, with only one weird month to worry about.


Problem is you would end up w/ three months of 30 days during common years.


"Did you ever wonder why December has 31 days?"

Honestly? No.

I did wonder why each of the 12 months got their specific amount of days, sure. I even wondered it about the special one, February, individually. But wondered it about December individually? No.


I vote for 30 days for all months, followed by a 5 or 6 day festival (depending) at the end of the year. ;)


I approve of this with one change: the festival would last to the second exactly as long as it would take to synchronize the calendar to our orbit around the sun (no leap years or anything).

Could you imagine a little more than five days with no representation on the calendar? Truly a "let everything go" period of time. Society would fall into chaos, no appointments could be made for those days because there is no way to address an exact point in time before the calendar reestablishment; truly an environment for a giant worldwide party.

In reality tough we would just call it the 13th month, or the minimonth and everything would go business as usual :/


This is the Alexandrian calendar, proposed in 238 BC and adopted in 25 BC. (Previous Egyptian calendars had a fixed 5 day intercalary festival.) It is still used by the Coptic church and parts of Egypt.


Not sure if you know this, but this is exactly how the Republican Calendar worked during the French Revolution. The bonus days were known as the sans-culottides (meaning days of the "sans-culottes," the self-chosen name for the Revolutionaries reinforcing the difference between them and the nobility) and celbrated Republican values like Labor and Virtue.


I thought it was because of the knuckles.

But if you want to discuss Romans and calendars, you've got to mention that Mars was there first month (because it's such a great month to start a fresh military campaign). So February was the last, and got whatever remains.

Why July and August both have 31 days, I have no idea. Augustus' jealousy sounds plausible, but apparently it's not true. So what's the real reason for this?


Because it falls on my middle finger's knuckle.


ring finger in the version of the mnemonic algorithm I know


I like the universal calendar with 13 months, all with 28 days in four weeks.

One extra day at the end of the year. Another extra day for leap year.


How about no months, 52 weeks per year with a 53rd 5-day week on leap years which would also correct leap seconds and such.

I have always organized myself weekly, so does everyone around me. I think it's much more natural than months, specially because you have week-ends which mark the end of each week, but not a month-end, you just drift from one month to the other, nothing happened (except bills and paychecks and whatnot, which can also be made weekly)


Then why not extend the week to a ten day period?

The decimal system seems to work better.


Because then we would have to alter the moon's orbit :D

Interestingly, contrary to the seasons, the moon's phases bears no effect on our lives since the artificial illumination of the cities. The seven-day week could be a ten day week; on the other hand I would be weary of ten-day week because of possible connections with my (and humans in general) metabolism and internal clock. Perhaps five days worked per two days rested is our optimal. There is no reason why ten would be the magical number, but seven seems to be working; further studies are necessary.


To see how/right wrong it is, you can actually look at the primary source material and see for yourself.

Here for example is an image of a pre Julian calendar http://en.wikipedia.org/wiki/File:Roman-calendar.png

Among other things, you can also see another bizarre truth: the pre Julian calendar did not have a leap day, it had a leap month (more or less randomly inserted sometimes to delay elections). It's labeled INTER on the image.

The terms NON and EIDVS are the Nones and Ides of each month the latter of which I imagine Julius would have liked to reform out of the calendar all together. Hindsight.

Which reminds me of a not so useful poem:

In March, July, October, May,

The Ides are on the fifteenth day,

The Nones the seventh; but all besides

Have two days less for Nones and Ides.


According to wikipedia, the julian calendar already had 31 days for august so the explanation of the article doesn't seem to be correct.

https://en.wikipedia.org/wiki/Julian_calendar


This is a un-true myth according to Wikipedia https://en.wikipedia.org/wiki/Julian_calendar#Sacrobosco.27s...


For centuries in the western world, the Christian church, which considered this to be in its purview, kept trying different calendar schemes to try to make a set of months, preferably equal, fit neatly into the year.

There have been may solutions tried or proposed, and the history of this has a certain puzzle-like fascination, but it's not clear to me why people cared so much about this. Astronomical periods are OK, apparently, for years and days, yet weeks and calendar months are arbitrarily imposed.


> For centuries in the western world, the Christian church, which considered this to be in its purview, kept trying different calendar schemes to try to make a set of months, preferably equal, fit neatly into the year.

That's not really a fair description of adopting a pre-existing system and issuing a single patch after more than a millenium. Sure, lots of things have been discussed, but not much has actually been tried.


It's my understanding that the original basis for months was the lunar cycle, which is astronomical for common definitions of the word. Part of what makes picking a calendar difficult is that the ~29.5 day lunar month doesn't fit neatly into the ~365.25 day solar year. So we've ended up with succession of adjustments and fudge factors.


There are a number of lunar calendars that have had long use. The Jewish, Islamic, and Chinese calendars, for example. They are a bit complicated for that reason. The Jewish calendar has the equivalent of "leap months", for example, while the Islamic calendar has a "year" that is ~10 days shorter than an astronomical year.


13 * 28 = 364, pretty close, you are left with 1.25 as a remainder, so every 4 years have a 5 day leap week.


Based on the comments in this thread, it seems the post is cute but wrong, so I'm burying it.


The more interesting question to me has always been: Why are the months numbered wrong? December is month 12, november is month 11, october is month 10. There's an offset of 2 between the month name and the number it references.


March or Martius* was originally the first month and the year was only 10 months long

http://www.webexhibits.org/calendars/calendar-roman.html

https://en.wikipedia.org/wiki/Martius_(month)


Because the calendar used to start in March. Feb was truncated because it was the last month of the year, thus where to put all the fudge factor to make the year work right.


It's also worth noting that 'winter' used to be monthless in the Roman calendar, so there were 10 months, ending in December (so all of the number names make sense) and then a big blob of 'winter' before the new year started.


True or not, it's finally given me the means to remember which months have how many days.


Interesting, I never thought of this problem before. In Thai, all month names are suffix with a word indicating number of days in a month. For example, "-khom" in "Mokkarakhom" (January) indicates 31 days, where "-yon" (e.g. Mesayon, "April") indicates 30 days. Benefit of being late-adopter allow us to name a month in a convenient way, I guess.


Interesting. What about February and leap years?

Does khom and yon have meanings?

The rest of the month's name is based on what?


February is using a special suffix, "-phan".

Both "-khom" and "-yon" means "come" or "arrival". Since our month names are basically a Zodiac in Hindu astrology, "Mokkarakhom" could be translated to "the arrival of Capricorn" (Mokkara/Makara) or April "Mesayon" is "the arrival of Aries" (Mesa).


Make your hand into a fist

Starting at the left, each knuckle and the space in between counts as one month.

Count all the way across and when you get to the end of your fist, start over (or carry over to the next fist)

Knuckle months have 31 days, months in between knuckles have 30 days (except for February)



Anyone else share a craving for Girl Scout cookies after reading this?


Who is Orvitz?


Same question here. The closest i could find was Ovitz: http://en.wikipedia.org/wiki/Michael_Ovitz

Ousted from Disney by Eisner in 1997, but was given a $100 million in stock severance - Roughly around the same time this mail was written.


No.


er...don't make sense


If it were true, that would mean Augustus messed up our calendar in addition to his biggest sin - sending us into the dark ages by making Christianity a viable religion. We are still in the shadows.


Why was this down voted?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: