Wikipedia talk:Manual of Style (dates and numbers)/Archive 1

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Archive 1 Archive 2 Archive 3 Archive 5

I notice that 10 January is available as a redirect to January 10, which is really really convenient for articles on military topics, but it seems to be the exception rather than the rule. Is there any reason not to create all the others? Stan Shebs 04:32 Feb 25, 2003 (UTC)

The only problem I can see is it would be likely to decrease the conformity of dates across Wikipedia, which might be a problem in, for example, biographies. Anyone using 13 February at the moment gets an empty link and can see immediately that that's not the standard way of formatting the date. But this is a small thing, and conformity in date formats might not be important enough to negate the convenience of having the redirects. -- sannse 07:29 Feb 25, 2003 (UTC)

My vote? I'd prefer all the dates to be uniform throughout the Wikipedia. -- Anon.

I don't understand -- why do military topics need the "10 January" format? --Eloquence

Accounts of military doings tend to have lots of dates, and they standardized on "10 January 1952" long ago; it looks pretty old-fashioned to read "January 10, 1952". Thinking to find an explanation for the difference, I consulted my Bible, 13th edition, and it says (8.36) "The University of Chicago press prefers that in all text, including notes and bibliographies, exact dates be written in the sequence day-month-year, without internal punctuation", which goes a long way towards explaining why most (though not all) of my recently-published history books have been using the "military" style for all their dates. :-) Stan Shebs 11:44 Feb 25, 2003 (UTC)

Fowler also says "10 January". I had assumed "January 10" was the US style, but if Fowler & Chicago agree I think we have to consider changing our own Manual of Style policy -- Tarquin 12:23 Feb 25, 2003 (UTC)
While not requiring dates in the "10 January" form The Elements of Style (Strunk & White) states:
The form is an excellent way to write a date; the figures are separated by a word and are, for that reason, quickly grasped.
If we were to change it, there are thousands and thousands of dates in the other form, every biographical article, for instance. Of course, our own software uses the other form, as shown in the signatures here. Ortolan88 15:52 Feb 25, 2003 (UTC)
I think it should suffice to allow both forms, add the redirects, and perhaps express a preference. If all dates are habitually wikified, it becomes trivial to convert to a single standard later. (Chicago just expresses a preference, not a rule; my guess is that day-month is today more of an academic/professional thing, and month-day is more "normal". Imagine it in an article about a grunge band, seems positively stuffy.) Stan Shebs 13:05 Feb 25, 2003 (UTC)
Hm, no, I don't think so - from my perspective here in the UK "month day comma" seems a strange way to express dates. I have a feeling that it might have been used 40 years ago, but nowadays day-month-year seems normal. Arwel 19:10 Feb 25, 2003 (UTC)
On a grunge band's CD sleeve maybe, but an encyclopedia article about a grunge band should be just as serious as any other. It would take a VERY long time to convert all mentions of dates; so even with a change of standard there'd be a certain amount of cohabitation -- Tarquin 15:00 Feb 25, 2003 (UTC)

Wouldn't it be easy for a programmer to create a bot to search all Wikipedia articles for [[month day]], formatted links and convert them to [[day month]]? I've already suggested that we should standardize on the international format. mav

Despite having suggested the use of day-month-year, I can't say that I'm 100% keen to try to change to it and/or enforce it everywhere. To take an opposite example, the authors and editors of the 1911EB had very definite ideas about "proper" subjects and "proper" ways to talk about them - and at the distance of a century, it all looks like a curious relic of a vanished age. I think that's why Chicago hesitates to require day-month-year of their authors - I just recently bought a 2000-published book Broadsides: the Age of Fighting Sail by Nathan Miller, solid history, but aimed at a less-technical audience, and it uses month-day-year. Wikipedia aspires to be solid content, but an encyclopedia with dozens of entries on Tolkien characters, comic strips, and video games is really aiming for the broadest possible readership rather than to be some sort of highbrow 21st-century EB. If day-month-year becomes popular, it will be obvious in the new articles that are being added. Stan Shebs 20:30 Feb 25, 2003 (UTC)

The esteemed Shebs omits what was the latest thing in dates for the 1911EB, which was the form 19th of March 1821. Ortolan88
Actually, many of us are working towards a highbrow 21st century EB. The Tolkien and Simpsons stuff is just along for the ride ;) It'll be first up against the wall when the Larry sifter project comes... -- Tarquin 20:44 Feb 25, 2003 (UTC)
Drifting off-topic, but I do see the range of subjects as a Wikipedia strength rather than a weakness. Classical scholars today would give their right pinkies just to know what the ancient Greeks really thought about homosexuality, all because at the time nobody thought it important enough to document as carefully as Alcibiades' personal failings. In fact, to use the Simpsons as an example, there are important archived software design discussions today that could not be understood 2000 years from now without an explanation of what "Doh!" is all about. Stan Shebs 21:21 Feb 25, 2003 (UTC)

Just to throw a simian wrench into the mill (being fresh out of wooden shoes), the reallio truelio military format is yyyyMMMdd; e.g. 2003FEB25. It sorts good, too. --the Epopt



VOTE HERE well, maybe not. Better try the top of the page then!


I thought a vote block might be handy - it's getting confusing keeping track of what everyone thinks... Approval Voting - pick one or more positions :)

Oh please, please do day/month/year. It is pain in the neck doing month/day/year on Wiki when here in Ireland nobody uses that form and everyone writes d/m/y. I keep having to remind myself on non-Wiki stuff to write d/m/y, then on Wiki to use m/d/y. According to my eMac settings

d/m/y is used by

  • Australia
  • Austria
  • Belgium
  • Brazil
  • Canada
  • Catalonia
  • Denmark
  • Finland
  • France
  • Germany
  • Ireland
  • Italy
  • Netherlands
  • Norway
  • Spain
  • Switzerland
  • UK

m/d/y is used by

  • US

y/m/d is used by

  • Portugal
  • Sweden
  • Japan
  • Korea

So if the Mac is right (and I'm going on what the setting say), of the countries listed, month/day/year is purely an American version. Elsewhere, most people use day/month/year. JTD 22:55 Feb 25, 2003 (UTC)

True, though my understanding is that the ISO standard date format is yyyy-mm-dd Arwel 01:41 Feb 26, 2003 (UTC)

It looks like nobody wants the status quo so I say we need to expand the exposure of this discussion. Hopefully a bot-writer will offer his/her services too. --mav

JTD, we are talking about dates with a spelled-out month here. Nobody wants to use the stupid 04/29/03 numerical format. There are times when it flows better and reads more clearly to say "On the 10th September the king said ..." and times when it reads better some other way. If you want to link to a date, sure, you find out what the Wiki convention is and do it that way so that the link works, but trying to specify writing style in this level of detail as a global policy is a draconian and quite daft idea. Tannin

That's what I'd much prefer to use. The point I was making is that most people outside the US use the format of d/m/y anyway, whether writing it out or in the numerical form. And the computer allows you to have the date in numeral or spelt out month form. It is the natural format whether written or numerical, in most places outside the US.

It isn't natural, it is only conventional. Let's convert February 14 to Day 45 (February 29 can be Day 59-L). You non-United-States-of Americans are so non-United-States-of-Americocentric.  :=) BTW - I have used the 10 January form since I first read about in Strunk & White 40 years ago and so do millions of others in this country without a name.Ortolan88

I don't see much benefits to change the format of dates. My position is always not to much concern about the format because the format cannot be consistent forever. -- Taku 01:16 Feb 26, 2003 (UTC)

Besides, it is possible to get new version of wikipedia software have some kind of rendering engine. In Japanese wikipedia, currently there is a discussion about the standard timezone. In USA history, Peral harbor happended December 10, 1941. But in Japan, it did in December 9, 1941. I am using the format like August 15, 2003 JST in Gyokuon housou because the date varies in country. The smart render should recognize some standard format including US style, EU style and display the date in uniform format. -- Taku 01:22 Feb 26, 2003 (UTC)

Actually the Attack on Pearl Harbor happened on December 7, 1941 Hawaii time/date. This issue has come-up before many times - it is best to indicate the date and time of the event where it happened. So the Pearl harbor attack happened on the morning of Dec 7, not on the date in Japan, Dec 8. --mav
Just so, Mav. The date to use is the one that appled in the place of significance to the event. So, for most purposes, Pearl Harbor was the 7th of December. But when considering palace politics in Japan (as in Hirohito), the correct date to use is the 8th. Tannin
Then you are talking about a different event that happened in the palace. So same logic applies: The article whose subject is the event in Hirohito's palace would say the 8th and the article whose subject is the attack would say the 7th. --mav
First forgive my ignorance about history event date (I can't believe I didn't when Pearl Harbor happended as Japanese!), what about the declaration of war? In Japan, it is December 7, 2003. In the USA, it is Dec 8, 2003. I guess. (Please please overlook historcal accuracy here). But if you say, this is a really rare case so MoS needn't cover, I can agree. Anyway, sorry this is off-topic.

Some years ago an American student I knew at university went into a panic when she received a note informing her that her exams were scheduled to start on 03.05.91. She went into a severe weekend of cramming (studying for 18+ hours per day), until she was almost hospitalised from exhaustion. Only then did she realise that her exams didn't start on the 5th of March - two weeks from the date she got the notice - but the 3rd of May. The problem was a couple of her American friends initially presumed that the date mentioned was read as dd/mm/yy, but when the saw her reaction, then presumed she must be right and began studying like hell also. And her moment of doubt disappeared when she saw them studying like mad, presuming that if they are studying she must be correct in reading the date as mm/dd/yy. The end result was five extremely exhausted and highly embarrassed American students, who were never let forget their mistake by all other students, who teased them unmercifully afterwards. JTD 01:28 Feb 26, 2003 (UTC)

This is a typical example about confusing of dates format. How can anyone be sure about which day is 01/02/03? -- Taku 02:31 Feb 26, 2003 (UTC)
Or perhaps five students who, for the one and only time in their lives, started studying for an exam at a sensible time instead of leaving it till the day before to start cramming? - Tannin (Who never started studying for any exam until 24 hours before!)

Now a really dumb question: What dates are we talking about here? Is this:

  • (a) all dates anywhere in the text of any entry?
  • (b) Only dates that are links to a date entry?
  • (c) Just the names of the date entry pages themselves?
  • (d) Dates only in particular circumstances, e.g., birth and death dates in a biographical entry?

Tannin 02:18 Feb 26, 2003 (UTC)

Yes, I'd like to know the answers to Tannin's questions too. Catherine
Personally, I'm interested in what to do with all dates in the text of each article. (That's why I'm not so keen to force a change; I'm sure there are bunches that want to be left alone (such as quotes from a book), and nobody's head is going to explode if they see both 10 January and February 28, even mixed in the same article.) Stan 04:39 Feb 26, 2003 (UTC)
To quote from the main Wikipedia:Manual of Style (and almost every other publications style book):
A Manual of Style has the simple purpose of making things look alike - it is a style guide. The following rules don't claim to be the last word. One way is often as good as another, but if everyone does it the same way, the Wikipedia will be easier to read and easier to use, not to mention easier to write and easier to edit.
Since there are many thousands of entries in the form January 10, I believe we should just leave it alone. The MOS has many such compromises in it. Consistency is the goal. Ortolan88
Hence my idea for a date bot. --mav

What dates are we talking about here?

Excuse me repeating myself, but without an answer to this question, discussion of anything else is impossible. Tannin 08:36 Feb 26, 2003 (UTC)

People who are voting for MoS policy are probably voting for (a),(b),(c),(d), except where in quotes, etc. People who are voting for encouragement are probably voting for changing the names of the date pages, and slight encouragement one way or the other on the dates.


Years in music

I've noticed that several pages relating to music have links of the form [[1991 in music|1991]], which looks a bit dodgy to me, since when we click on what looks like a year, we generally expect to end up at the ordinary year page. Has any policy been decided about whether or not this is the thing to do? (I don't like it myself, and I've just removed some, but I thought I'd better ask, to see what other people think.) -- Oliver P. 11:45 Feb 26, 2003 (UTC)

Seems okay to me. The 1991 in music page does link to the general 1991 page and someone following that link is probably interested in the surrounding music. Personally I'm more annoyed when links take you way off the subject, such as the icon reference in [MTV]] taking you to a page about religious icons. Bagpuss
This issue has been decided at Wikipedia:Wikiproject Music standards -- I think the discussion is in the talk archive now. Tuf-Kat
Thanks for pointing me to that page, Mr. Kat. (Don't use piped links to the "year in music" articles, in case anyone else is reading this and can't be bothered to follow the link...) -- Oliver P. 23:16 Feb 26, 2003 (UTC)

Date markup

I am not sure about what kind of date markup people here are talking about. But if they think like MathML or Tex, it is not neccesarry. If you know little about Perl or regular expression, you might know that it is a snap to convert various format of date into certain date format. It is not important to make uniform in raw text but the diplaied date should be consistent for the sake of beauty. Again, it is quite possible to make rendering engine render serverl dates format into only one like "01/02/03", "01-02-03", "Jan 2, 03", "January 2, 2003" into 02-01-03 (day-month-year) in display. Remember the representation of data should be separated from data itself. It is a basic lession from software enginnering. -- Taku 20:57 Feb 26, 2003 (UTC)


I have written a Perl routine that
  • accepts a string (probably submitted to the routine because it was found surrounded by "<date> - </date>")
  • parses the string into a day, a month, and a year for a wide variety of formats
  • returns one of four things:
    • [[month day]], [[year]] if a flag was set to "American"
    • [[day month]] [[year]] if a flag was set to "European"
    • [[ambiguous date|original string]] if the date was similar to "1/2/2003"
    • [[unparseable date|original string]] if it couldn't figure it out at all

The format flag is to allow users to set a preference indicating how they want to see dates formatted.

Does anyone care? --the Epopt

Sounds good. To make your life more complicated (hehe) can you get it to make links to December 2002, etc, if the date is after the start of 2001?
Also, does it handle BC dates as well?

I have given up on trying to partipate in this discussion since no-one seems to want to bother trying to define just what dates it is we are talking about. If we can't even spell out WTF it is we are supposed to be discussing, how on earth is anyone supposed to be able to make up their minds about it? My thanks to Martin for his courtesy in at least having a guess at this fundamental question. But guesswork is not good enough. The first step in any discussion must always be to define the question clearly. Tannin 09:12 Feb 28, 2003 (UTC)

On the one hand, I think at the moment there's plenty of room for different styles, as Stan said. We've got plenty of other things to worry about besides a draconian date consistency.
On the other hand, lack of consistency can end up looking unprofessional; in that sense, I think any decision would have to apply to ALL dates, linked or not.
I've gone back and forth on the date markup idea. Originally my thought was that it would allow a date entered in ANY format to link consistently with the proper date page -- something that would automagically add links to [[December 2002|December]] [[December 17|17]], [[2002]] OR [[February 28|28]] [[February 2003|February]], [[2003]], or whatever. This remains the most compelling reason to use it, for me. It would also make it easy to change things to a consistent format sometime in the future, if we decide it's needed, or to customize the display in user preferences. I do think that if it were implemented, we'd also need a <nodate> tag to prevent changes to book quotes or similar uses.
However, as it's been pointed out on the mailing list, it's one more markup for a user (esp. a newbie) to remember; it's unlikely to be used consistently by every user, meaning a lot more editorial cleanup to make it work; and in the end, dates are fairly easy to find and replace programmatically even without markup (with some human oversight). Are the benefits worth the cost? I'm not sure -- it depends on what decision is reached about consistency, or user preferences (which, by the way, will not affect what the average drive-by google searcher is going to see -- only those close enough to the project to log in.)
Excellent effort on the code, Epopt. I'm sure it will come in handy one way or another.
Personally, (even as an American), I prefer the "28 February" format -- I find it easier to read, type, and edit. I'll continue to watch the discussion, though, and abide by whatever decision is finally reached. -- Catherine 18:50 Feb 28, 2003 (UTC)
I won't. I'm dammed if I'll abide by decision to specify such a minute and piddling detail of what I (and other contributors) are supposed to write (which does not and cannot take account of context) when no-one can even be bothered stating what the "decision" is supposed to apply to. Tannin

There is no ISO standard for writing dates. That's a great misunderstanding. The standard has to do with representing dates as numbers only. Says Russ rowlett, in A Dictionary of Units of Measurement:

"The international standard covers numeric representations only; it doesn't prohibit traditional statements of the date, using words, in whatever the local style is. Thus "April 5, 2001" remains perfectly acceptable. However, non-standard numeric representations, such as 04/05/01 [should be 2001-04-05], should not be used, because these representations are interpreted differently in different countries."

Arthur 00:34 Mar 1, 2003 (UTC)

Whatever about 04/05/01 (mm/dd/yy) I would absolutely be opposed to yy/mm/dd. It is rarely used and when it is is usually regarded as absurd. Most people use dd/mm/yy, some (mainly in the US) use mm/dd/yy, but where on the planet uses yy/mm/dd? dd/mm/yy is the obvious one to use. mm/dd/yy those who don't use it can tolerate, but yy/mm/dd? Use that and you'll confuse and annoy everyone JTD 00:45 Mar 1, 2003 (UTC)

Actually, YY/MM/DD is common, JTD. It's used as the standard in Japan, several other places too (I think), and has become the standard numerical form for multi-national cooperative ventures (web discussion pages, for example). It's common as mud, and is the only unambiguous numerical method of writing a date. However, because Wikipedia has (very sensibly) chosen to go with spelled-out dates for the most part, it is not something we need consider using here. Tannin 00:54 Mar 1, 2003 (UTC)
The reason it would be used is that it's easily sortable, JTD; 'twas not a consideration in previous eras, but it certainly is now. Not necessarily easy to parse for those of us raised with a different "obvious" system, but I think I could get used to it pretty quickly. Just not in the text of 'pedia articles.  :) Catherine

It may well be used in Japan and for multi-national corporative ventures but it isn't used generally in academic source books, is not used generally in Europe, Africa or much of Asia. And in ordinary discourse outside places Japan it is almost completely unknown. Its principal use would be as an unambiguous numerical method that would not be understood by the vast majority of users and potential users. Thankfully we do spell out dates so the problems that would arise here wouldn't occur, but if we ever want to totally confuse the vast majority of users, this is the system to use. JTD 01:09 Mar 1, 2003 (UTC)

The point is moot: we ain't using numeric dates now and we ain't planning to. Printed books don't use it, JTD, because (a) they mostly use spelled-out dates, and (b) if in doubt, you only have to look inside the flyleaf to see if it's American (in which case you have to put up with the absurd 05/29/79 format) or anywhere else (in which case it's 29/05/79). So for most printed things, there is no risk of ambiguity. The better quality electronic publications, where a spelled-out date is not appropriate, use 79/05/29 as routine. No-one is in the least confused by it. Tannin

This entire section is silly -- NEVER USE TWO-DIGIT YEARS, unless discussing the years from 10 through 99. --Brion 04:41 Mar 1, 2003 (UTC)

The only time anybody should be thinking about copyediting dates is when they're ambiguous and as long as we spell out the month, we won't have ambiguous dates. If it's decided that we have to stick to one date style for reasons of consistency, it can only be a matter of time before it's decided that we have to stick to one spelling style for reasons of consistency. That way lie extremely silly and pointless disputes between people who know what the "correct" style is and people who don't. And consistency is not necessary in the way that NPOV and unambiguity is. It just makes it more difficult for new contributors to add "acceptable" material without a lengthy "training" process and the "training" process is long enough when it just involves NPOV, never mind all the other standards which have grown like weeds round here. We've learned to avoid a lot of pointless arguments over spelling by allowing any widely recognised spelling of a word to be used. Let's learn from that and apply it to dates -- Derek Ross 01:40 Mar 1, 2003 (UTC)

Well said. No unambiguous form should be disallowed. I'm embarrassed that I didn't see relation to the already solved spelling issue. Arthur
I like standards. I think if it were made policy that dates can be used in any way people want, visitors to the Wikipedia would find the inconsistency off-putting. When I first arrived, I was horrified by the American-style dates, and tried to avoid using them. But soon I was swayed by the fact that even if it was an unpleasant convention, it gave the place a nice uniform feel, and so I gave in and started writing dates that way. I accepted that following the convention would make the Wikipedia more coherent, and thus more professional-looking. Of course, my actual contributions may not be very professional-looking, but I do try... :) -- Oliver P. 03:58 Mar 1, 2003 (UTC)

We are all talking a lot about this issue, but what is the decision? When will it be made? Or is this an open-ended vote where nothing will be decided because the vote never will actually end? We can decide to make a decision. Or decide not to make a decision. But can we at least decide if we are actually going to take a decision here!

Or we could even decide what it is we are making a decision about. (As asked no less than four times already without result.) But it doesn't seem likely. Tannin

Ok - time for answers: JtdIrL 04:36 Mar 1, 2003 (UTC)

Question 1: Which dates are we deciding should be subject to these dates?

  1. All dates
  2. Dates at the start of Articles

Question 2: When are we taking the decision?'

  1. In the next twenty-four hours?
  2. In one week?
  3. Never?