Wikipedia talk:Manual of Style/Dates and numbers/Archive 10

This is an archive of past discussions about Wikipedia:Manual of Style. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 5

←

Archive 8

→

Spaces in date ranges

It is obvious that unspaced en-dashes need to be used in unspaced date ranges such as "1978–2004". However, what about spaced dates like "1 September 1978–1 August 2004"? Doesn't the dash look like it is joining the "1978" and the "1" rather than the first whole date and the second whole date? I think "1 September 1978 – 1 August 2004" is better and should at the very least be permitted. — Chameleon Main/Talk/Images 21:44, 1 Aug 2004 (UTC)

This seems sensible to me, certainly.

James F. (talk) 23:36, 1 Aug 2004 (UTC)

Spaced en-dashes are certainly permitted. Perhaps the word "to" would be better then[sic] either? Gdr 14:58, 2004 Aug 3 (UTC)

The thing is that I think they should only be permitted in this case for clarity. I don't see why they should be allowed in other cases. As you said, usage guides say no to spaces. — Chameleon Main/Talk/Images 15:41, 3 Aug 2004 (UTC)

I have edited the article to reflect the de facto Wikipedia standard. See stats and comments above. chocolateboy 16:05, 8 Aug 2004 (UTC)

Why? The style guide is about how Wikipedia should be, not how it is. It's like being on an examination board and disagreeing with the other professors because you think students should get higher marks not for putting the de jure correct answers, but the de facto ones,[sic] i.e. the answers most frequently given by other students... on the grounds it is more democratic or something.

Any research such as that carried out by you above is flawed because it does not take into account one thing: it is harder to insert dashes than hyphens because they do not feature on keyboards. That will always skew results. The same thing occurs with the correct use of special characters in foreign words. Most English speakers will enter them incorrectly, and the rest of use[sic] have to come along afterwards and tidy up, as we do with punctuation. The de facto standard on Wikipedia is also stubs, lists, semi-correct info, badly-written prose and huge US bias. Should we also encourage this so as not to be nasty, pedantic, ivory-tower prescriptivists? Does anyone agree with you? — Chameleon Main/Talk/Images 18:01, 8 Aug 2004 (UTC)

Hola.

Please don't reformat my comments.

Why?

For the reasons given above. Please answer facts with facts rather than op ed.

Any research such as that carried out by you above is flawed

I usually find that's the case when someone disagrees with me.

because it does not take into account one thing: it is harder to insert dashes than hyphens because they do not feature on keyboards.

It's "harder" to enter <em>foo</em> than ''foo''. That's why the latter is preferred. The flaw is in your contention that "difficult to enter" should be the preferred solution in wikitext.

The de facto standard on Wikipedia is also stubs, lists, semi-correct info, badly-written prose and huge US bias.

I took considerable pains to support my position. I hope you will corroborate your indictment of Wikipedia by citing your sources or providing evidence rather than impugning the competence and cluefulness of the majority of Wikipedians.

Does anyone agree with you?

The vast majority of Wikipedians agreee with me as the statistics cited above amply demonstrate.

chocolateboy 19:06, 8 Aug 2004 (UTC)

You assume that your opinions are statements of facts and that mine are mere opinions. All individuals are probably guilty of this, but you don't have to be so open about it. You have not supported your position at all; you have merely shown that there is much poor punctuation on Wikipedia. Only in your opinion does that support your position. In my opinion that supports a drive to improve punctuation in Wikipedia.

There is a huge logical leap between seeing that there are a great deal of hyphen hacks on Wikipedia and concluding that everyone actually agrees with you. The vast majority of people have never thought about it, don't care either way or have no idea how to enter a character not on their keyboard. — Chameleon Main/Talk/Images 20:07, 8 Aug 2004 (UTC)

You assume that your opinions are statements of facts and that mine are mere opinions.

I have repeatedly cited Wikipedia policy and expounded an argument supported by verifiable statistics. You've done neither. If your opinion has the weight of consensus behind it, it should be easy to demonstrate. It merely requires a complete rewrite of a raft of policy documents, and a rigorous demonstration that the statistics I provided are inaccurate.

All individuals are probably guilty of this, but you don't have to be so open about it.

Why? Are we all required to dissemble ("don't have to be so open") and argue from opinions rather than facts?

You have not supported your position at all; you have merely shown that there is much poor punctuation on Wikipedia.

Please answer the points made in my comments above before wading further into the realm of patronising anti-Wikipedia invective.

The vast majority of people have never thought about it, don't care either way or have no idea how to enter a character not on their keyboard.

Unfortunately for your argument, the vast majority of Wikipedians decide Wikipedia policy. It is not decided by a militant minority hell-bent on propagating the illusion that Wikipedia is paper.

chocolateboy 22:46, 8 Aug 2004 (UTC)

Who is this militant minority propagating an illusion that Wikipedia is paper?

Wikipedia has been printed on paper and the plan is that it will be printed on paper again, many times. After all, anyone can print it. But I don't think about that when I edit. I think about the screen display that I see. And I think dashes look better on the screen, that they give an article a more professional air. That did seem to be consensus, though of course people do not especially like the entities necessary to accomplish this. But, one works with the tools at hand. I would also like to simply press CTRL-B or something similar and have bolding turned on in my editing display without any visible coding.

I have background in editing and writing and doing layout on paper. I am a programmer by profession. There are desireable differences between print on paper and screen display. Most of those differences can be accomplished by twiddling a CSS while using the same text. I don't see that use of dashes falls into differences that are for some reason desireable on screen but not on paper or vice versa. Hyphens for dashes were never really asethetically[sic] desireable on typewriters. It was simply that for reasons of cost and mechanical practicality that the number of characters had to be kept down and using "--" or " - " for dashes and hyphen for both en-dash in ranges and for minus was a reasonable way to do it, especially when characters were, for the first time, all of a single width. People got used to the typewriter set of characters, and it made sense at the time to mostly transfer it to computers as it existed. But new technology throws away its limitations. To turn your strange accusation back at you, are you propagating the illusion that Wikipedia and the web should continue to accept old typewriter limitations when the reaon behind them is gone? You don't ask that Wikipedia articles be displayed in a fixed-width font at 80 characters to a line, which also used to be computer "standard". You accept italics rather than underscores and bolding rather than asterisks. I presume you accept superscript and subscript and so forth. Why are you so set on something which has been "standard" mostly because people could do no better?

Jallan 04:37, 10 Aug 2004 (UTC)

Unified date feature proposal

I propose a more general approach to the Date/Time handling, which I filed as a Wikimedia feature request [1]

Please comment (at sourceforge). If you feel that it's better to continue the discussion here than there (I'm pretty new in Wikipedia so I'm sorry if it goes without saying that date/time issues are only discussed here and not there[.]), then please put a comment at the sourceforge pointing here and continue the discussion here. BACbKA 22:12, 2 Aug 2004 (UTC)

There are a lot of difficult issues in your proposal that need to be thought about (and argued over), so I think it would be a better idea to work out the details here before presenting a more worked-out proposal. I can't add comments to the SourceForge feature request, so I hope someone else will add a note pointing here. Gdr 15:45, 2004 Aug 3 (UTC)

BACbKA's proposal

At the present moment, whenever a date/time period/epoch is referenced within a Wikipedia article, various notations are used, and they are also unportable with respect to language and calendar systems.

It would be nice to have a special tag for a date specified to a given degree of accuracy ([I].e., referring either to a particular date or even date+time, or to a particular week/month/century/millenium[sic]/epoch or some interval, like XIV-XVII AD[.]).

(I am purposefully omitting any concrete implementation details and markup details at this stage, preferring to reach full agreement on the desired functionality first. If you want it in XML, I can easily prototype a DTD encompassing all the features once the agreement is reached[.])

The markup should allow specifying the date according multiple calendar notations (such as Julian/Gregorian/Russian Orthodox, or Arabic, Jewish, Mayan or whatever else—the system should be designed in such a way that adding another notation should not affect the existing dates specified in the notations already supported[.]).

Aside of the way the date is *specified*, there should be a possibility to attach an attribute to the date that would signify the native calendar system for the date (rationale[,] a national/religious recurring holiday is celebrated according to the given calendar).

Now when a Wikipedia article is rendered, depending on the language it is in, a given date is translated as follows: 1) it is coerced from the originally specified notation to the native notation for the article language (as long as the attribute forcing a particular notation is not set) 2) the current language/locale is used to express the particular notation (also maybe as per the current user's preferences environment) 3) in case of the mismatch about the native notation for the article context and the native notation for the specified event/period, the date might be represented twice (this behavior might also be affected by the current user preferences)[;] [e].g., in an English Wikipedia article about an event native to the Hebrew calendar: 26/Jul/2004 (9th of Av, 5764)

This would simplify trans-article date formatting decisions within a language-specific Wikipedia greatly, and would also facilitate easier translation of Wikipedia article portions.

Discussion

My initial thoughts are

It looks like a good idea and something like this will work for many cases, especially modern dates.
But there are lots of reasons why a date should not be translated, or should be translated with care:

Sometimes it is wrong to translate: for example, an article discussing dating systems themselves.
Date conversions between systems that have different starts to the day might be tricky. For example, days in the Hebrew calendar begin at sunset, so 2004-07-26 isn't necessarily 9th of Av, 5764.
Calendars like the Islamic calendar rely on an actual sighting of the new moon, so when translating to the Gregorian calendar there are a couple of days of uncertainty.
When translating between the Julian calendar and the (proleptic) Gregorian calendar you can't be accurate to the day because of the doubt over which years between 44 BC and AD 8 were leap years.
With some calendars there is disagreement over how they relate; for example there are still multiple opinions on how to translate Mayan dates, or how to date events in ancient Egypt.

The attribute forcing a date not to be translated will solve, or at least help, problem 1. Problems 2 and 3 are not very severe, but could in any case be solved by a system allowing you to specify a date in multiple systems if you needed to. I don't know what to say about problems 4 and 5.

It would be nice to see a detailed proposal. Gdr 15:45, 2004 Aug 3 (UTC)

Use of numerals

I propose to replace the text of the second paragraph of section "Number names" with the following:

In ordinary text, whole numbers from zero through one hundred, and any such whole numbers followed by hundred, thousand, hundred thousand, million, and so on should be spelled out. A number other than these should be represented in numerals, except in sentences where it is the first term or where spelled-out numbers would crowd, as in a list. Numerals should also be used where appropriate in dates, times of day, currency, or scientific, mathematical, or technical prose.

This is far more comprehensive, precise, and accommodates necessary practice for certain fields. Spelling out numbers up to one hundred is the practice prescribed in the Chicago Manual of Style, which is for formal works which do not have a strict limitation on the amount of text, and elsewhere. The practice of only going up to ten is prescribed in the Associated Press Stylebook, where formality is sacrificed in order to more strictly limit the amount of the text. - Centrx 02:54, 12 Aug 2004 (UTC)

I strongly object to this for two reasons: 1) Using numerals for numbers greater than ten is widespread throughout Wikipedia; 2) spelling out the numbers makes the text quite clumsy looking, IMO. I see no reason to prefer the CMS to the AP. I do not understand what you mean by "accommodates necessary practice for certain fields". Can you explain? If you can identify specific contexts in which there is a clear advantage to spelling out the numbers, I could agree to doing so only in those specific contexts. But for general use, I would not want to see the numbers spelled out. older≠wiser 03:05, 12 Aug 2004 (UTC)

Using numerals for numbers greater than ten is widespread throughout Wikipedia

This does not mean that the standard should not be otherwise. Present use is no impediment and differing use between articles—which will be resolved—should not have any bearing on what is the proper standard.

spelling out the numbers makes the text quite clumsy looking

I don't see how this is the case. All of the numbers would be of a form where the smallest lettering would be "ten" and the largest would be "seventy-eight". I don't see how this looks clumsy,[sic] they are just words, and for reading it seems it might be less clumsy, for the words are spelled out in front of you. This is the English language in a formal work; it is appropriate to use words. We should not, for instance, replace all instances of "for example" with "e.g.". Note also that the proposed text indicates that numerals should be used rather than spelled-out numbers where those numbers would cluster thickly, that is where it would be clumsy looking.

I see no reason to prefer the CMS to the AP.

The reason I mentioned this is that, of the two major manuals of style which indicate numeral practice, the one which deals with text that is designed to be short prescribes 1-10 and the one which deals with text that is more formal and is less constrained on size prescribes 1-100. The Wikipedia does not need limitations on text size and it is appropriate for it to be formal.

Wikipedia is not a book. We prefer news style in many article[sic] over book style and I would argue that reading materials online is much closer to news style than to reading a book. What do manuals of style for online resources have to say in this matter? older ≠ wiser 21:15, 16 Aug 2004 (UTC)

I do not understand what you mean by "accommodates necessary practice for certain fields"

This statement was not regarding the limit on numerals; it was in reference to another provision of the paragraph. The present text of the page does not provide for the use of numerals between 0-10, or the proposed 0-100. The proposed text, however, provides that numerals should be used where appropriate in terms of date, time of day, or currency, or where appropriate in various technical fields, for example in the properties of chemical elements. So, please also note that there are several new and modified provisions in the proposed text which are not directly associated with the issue between 1-10 and 1-100. - Centrx 14:05, 12 Aug 2004 (UTC)

I second the above proposal. Whenever numbers can be written out with one or two words, numerals should not be used (e.g. one hundred, ninety, ten million). See The Chicago Manual of Style, "Numbers in Writing", Gude to Punctuation. -- Emsworth 20:02, 16 Aug 2004 (UTC)

I think we should be more flexible. We can perhaps recommend writing out small numbers but I think we should not require it. There are too many contexts where numerals are better.

The English had 102 ships under George Monck, the Dutch 98 ships under Maarten Tromp.

By the end of October, Tromp had 117 ships, including 12 fire ships.

The German panzer armies encountered the US 99th and 106th divisions.

The Germans suffered between 60,000 and 104,000 casualties.

Are these really better as the following?

The English had 102 ships under George Monck, the Dutch ninety-eight ships under Maarten Tromp.

By the end of October, Tromp had 117 ships, including twelve fire ships.

The German panzer armies encountered the US ninety-ninth and 106th divisions.

The Germans suffered between sixty thousand and 104,000 casualties.

I think not. Gdr 20:12, 2004 Aug 16 (UTC)

So then, add something like the following to the criteria: "For numbers that are applicable to the same category and used in the same context, where this rule prescribes that some of the numbers be numerals, for the sake of consistency use numerals for them all." - Centrx 02:29, 17 Aug 2004 (UTC)

Certainly, there is room for flexibility. (As for army divisions, incidentally, "1st Division," "10th Infantry," etc., appears to be common practice.) But, in general, numerals should not be used where one or two words could be employed to represent the number. I do not favour any attempt to impose any absolute requirement, either that only numerals should be used or that only words should be used (except in certain limited circumstances, where one or the other is clearly and definitely appropriate). -- Emsworth 20:24, 16 Aug 2004 (UTC)

I would prefer that this flexibility be incorporated into the proposed change. As presently worded, it seems pretty inflexible to me—I would not want to encourage any bot-assisted campaigns to change all occurences[sic] of numerals to words. There are plenty of cases where numerals are preferable. As I said above, I still have not seen any compelling reasong to prefer the CMS over the AP in this matter. older ≠ wiser 21:15, 16 Aug 2004 (UTC)

I think your last objection has already been addressed. The Manual of Style often uses the CMS as an authority. The CMS applies to formal writing, including encyclopedic writing, whilst the APS applies to newspapers, where space is at a premium. Needless to say, our style of writing fits the former. In any event, I think the following wording might be desirable:

Perhaps I have missed something, but I have not seen where the MoS gives preference to the CMS over other style guides. It is one among many style guides that are frequently referenced. There IS however, a stated preference for News style. Wikipedia is not a book. I STILL have not seen any COMPELLING evidence to prefer the CMS over AP or other news style type guide. The Economist, The Guardian, The Times (of London) all say spell out number 1-10 and use numerals above that. I'd say we are in very good company. I have no problem with allowing both styles (much like we allow both British and US usage), but I don't want to see the book style to become the recommended style for Wikipedia. older ≠ wiser 22:27, 16 Aug 2004 (UTC)

Where does it say that news style, as you construe it to include this issue of numerals, is the preference of the Wikipedia? Wikipedia:News style stresses having the "impact" of the news style, but does not dictate the encyclopedia articles follow the news style. The news style that is the stated preference is, as the news style article describes, "the order in which stories present information, their tone and the readers or interests to which they cater". In other words, the stated preference is for clear, accessible prose where sections are easily parsible by virtue of the pyramid style. It is not because of a preference for the format, that is nearly typographical format, of a newspaper. At the same time, why should the AP style for newspapers be our guide in contrast to the CMS which is more appropriate for works of serious length and is used, for instance, in publishing various encyclopediæ and other reference works.

What is your point in saying that "Wikipedia is not a book"? What exactly about a book is something that we should not have in the Wikipedia? Why do you (it seems, you personally) prefer that Wikipedia not have the features of a book? Usually, the reason for mentioning that "Wikipedia is not paper" is in order to stress that there are not limitations on length... In short, I think saying "not being a book" or "being like a newspaper" are not productive as analogies and, indeed, are in many respects contrary to the well-established form of the Wikipedia. - Centrx 02:29, 17 Aug 2004 (UTC)

I did not say that the stated preference for news style explicitly included the issue of numerals. However, I would say that it supports the case for prefering a news-type of style guide such as AP (or that of the Economist, the Guardian or the Times) over the CMS. The typographical format of reading Wikipedia online shares much more similarities with a newspaper than it does with a book. I gues[s] I simply do not agree that CMS is a "more appropriate" style guide for Wikipedia. older ≠ wiser 03:04, 17 Aug 2004 (UTC)

In what way is the format of the Wikipedia more like a newspaper than a book? Indeed, why is the format of the Wikipedia not more like a paper encyclopedia, a book of the same purpose as the Wikipedia? "Simply disagreeing" does not constitute an considerable argument. - Centrx 01:58, 18 Aug 2004 (UTC)

I don't think we're going to get anywhere with further discussion. You've laid out your argument, which I find unconvincing; and I've presented my rationale to the best of my limited ability and interest in the matter, which you have similarly found unconvincing. As I said, I'll abide by the consensus opinion. older ≠ wiser 02:34, 18 Aug 2004 (UTC)

You have not laid out a rationale. Regardless of numerals, I would like to know why you think that the Wikipedia is more like a newspaper than a book. Why is it more like a format with a limited space rather than a format with a vastly greater space? Why is it more like a format for delineating specific events and researching information directly relevant to current events rather than a format for thoroughly fleshing out an entire topic? Why is it more like a newspaper (on news paper) rather than an encyclopedia (in a book)? - Centrx 17:19, 21 Aug 2004 (UTC)

In general, whole numbers from zero to one hundred, and whole numbers followed by hundred, thousand, million, and so on, should be spelt out. Other numbers should be represented by Arabic numerals.
The rule above is not absolute. In several circumstances, it may be preferable to use only numerals, even for whole numbers between zero and one hundred, etc. Numerals may be used to prevent crowding (e.g. in tables). When multiple numbers are listed, it is acceptable to use only one form for all of them (e.g. "between 95 and 135," instead of "between ninety-five and 135.)
Arab numerals should be always be used for dates, times, years and currencies, and in scientific or technical contexts ("1000 km/s," not "one thousand km/s").
Numbers at the beginning of sentences should always be written out, or else the sentence should be restructured so that the number occurs in the middle.

Any suggested amendments? -- Emsworth 21:46, 16 Aug 2004 (UTC)

The caveat "in general" and a statement that the rule is not absolute in practice almost brings us back to the current standard. Change scientific to technical in the current standard and I don't see much wrong at all. Short is better.

From the Encyclopædia Britannica: Carom Billiards:

The 1993 World Billiard Association (BWA) three-cushion championship was won by an American for the first time in 40 years at the BWA's World Cup in Ghent, Belgium. Sang Chun Lee, a native South Korean who moved to New York City in 1987, won the crown despite lagging in cumulative tour scoring at the onset of the sixth and final tour stop. The 39-year-old Lee trailed both 21-time world titlist…

This is just an article that has more numbers than most I found. But use of digits rather than words appears throughout their current articles. For example, from Augsberg Confession:

Latin Confessio Augustana , the 28 articles that constitute the basic confession of the Lutheran churches, ...

Their article Twelve Tribes of Israel which can be seen fully if you log in, uses spelled numbers "six" and "two" but then refers to "The 10 tribes that settled in northern Palestine ..." Neither "twelve" or "12" actually occurs in the article.

Or check Oxford English Dictionary: A Day in the Life of the OED. It contains "another thirty or so" and "twentieth-century quotations" but also "roughly 70 quotations" and "batches 19 and 20".

The Style sheet for the Journal of the Royal Musical Association prefers digits for numbers 10 and highter[sic], but not when numbering centuries. The MIT Style Sheet also wants digits for numbers 10 and higher. These are just the first style sheets I found that covered the matter.

So leave it to user discretion as is currently done. There are too many academic and scholarly resources that follow the same style that APS recommends to make a case that it is currently more correct or academic to spell out numbers in full when they are less than a hundred.

In any case, the Wikipedia style sheet should not normally be repeating material found in most other style sheets except for cases where almost all style sheets agree and editors here still very often get it wrong. I don't see any point to adding material on such matters.

Jallan 23:33, 16 Aug 2004 (UTC)

So leave it to user discretion as is currently done.

Then this should be established in the Manual of Style. Many times, others have changed the numbers of my edits while pointing to this manual. Still, this is a guide and I see no reason why an appropriate practice should not be indicated in it while stressing that automated bots must not make changes to this effect by reason of the complexity of the matter.

the Wikipedia style sheet should not normally be repeating material found in most other style sheets

Why should we not be establishing guidelines of style that are appropriate to this unique platform? - Centrx 02:29, 17 Aug 2004 (UTC)

---Proposed modification of section "Number names" (which I think should be separated to "Use of numerals"):

In ordinary prose, whole numbers from zero through one hundred, and any such whole numbers followed by hundred, thousand, hundred thousand, million, and so on should be spelled out. A number other than these should be represented in Arabic numerals, except where indicated below.
If a number is the first term in a sentence, it should be spelled out. It is often better to restructure the sentence so that the number is not the first term.
Where spelled-out numbers would crowd, as in a list, use numerals.
When both spelled-out numbers and numerals would be in the same category and used in the same context, such as in a comparison, use numerals for all of the numbers. Note that numbers of different categories are treated differently.
Numerals should be used where appropriate in dates, times of day, currency, or technical prose, especially in tables or other areas where space is restricted.

Suggestions? - Centrx 02:31, 17 Aug 2004 (UTC) (removed examples - Centrx 01:58, 18 Aug 2004 (UTC) )

I don't think I can support any version that indicates a "should" reccommendation[sic] along the lines of "In ordinary prose, whole numbers from zero through one hundred, and any such whole numbers followed by hundred, thousand, hundred thousand, million, and so on should be spelled out." Since you seem to think this is only my preference, I would be willing to have a poll on the matter and accept consensus whatever that may be. older ≠ wiser 11:34, 17 Aug 2004 (UTC)

I can't support such a recommendation either. And Centrx's suggestion is much too long for what should be a simple recommendation along the lines of "Writing out numbers is preferred in these situations, use numerals in these other situations, otherwise use your judgement." Gdr 12:00, 2004 Aug 17 (UTC)

In fact, it only looks longer because of the examples and is shorter than Lord Emsworth's proposal when those are removed. Including examples is a separate issue and they might, for instance, be added to the present text if that is considered beneficial. They are provided, I suppose, only for interest and may be easily removed for they are not necessary. - Centrx 01:58, 18 Aug 2004 (UTC)

Personally, I quite strongly prefer a hard-and-fast rule of

If the value is over ten or non-integral, use numerals;
In a list where any item is over ten, use numerals;
For ages, scientific data, technical specifications, dates, use numerals.

I think it's clearer—for non-native English speakers, dyslexics, dyspraxics and those of us (including myself) with an astigmatism. I even prefer numerals for century names, though I can live with those.

Quite frankly, I find 18th Century much easier to read than Eighteenth century; I'm hardly illiterate (thankyouverymuch ;o) , I just find it much easier. And I am a native English speaker. Given the English-language Wikipedia is [to my ken] the most extensive, I find it hard to believe that the readership doesn't include non-native readers. To be quite honest, I don't understand the need or desire to use words when numerals are (a) more than sufficient and (b) quite acceptable in several style guides. — OwenBlacker 12:48, Aug 17, 2004 (UTC)

If, then, several style guides differ on the matter, let us leave it up to the preference of the original creator (except in circumstances where either numerals or words would be clearly correct). As with British/American English differences, the primary contributor's preferences for numerals above ten could be respected. -- Emsworth 13:45, 17 Aug 2004 (UTC)

Whilst I'd still find that harder to read, it's certainly an acecptable compromise. It's not like there aren't myriad issues where we "solve" the problem that way already (British/American English; Dashes etc[.]). That's fine by me, I guess. — OwenBlacker 16:39, Aug 17, 2004 (UTC)

I quite agree. That is what the current statement on the matter indicates with "may be represented",[sic] that is, do as you will on the matter of numbers from 10 upward.

Centrix asked: "Why should we not be establishing guidelines of style that are appropriate to this unique platform?" Unique is irrelevant. There are other encyclopedias on the web and other sites providing much factual information on the web. Also the current Wikipedia style guide is already unique, just as every style guide is probably unique. That's neither good nor bad. The current style guide allows individual freedom in some areas while being restrictive in others. I don't see that the question of when numbers should be rendered by Arabic digits and when they should be rendered by spelled-out words is one of those areas where guidelines should be more restrictive than they are now, especially not in a direction which goes against what many here prefer and which is also preferred by many style sheets and by the Encyclopædia Britannica. So far consenus is against Centrx's preferred use, tending towards leting the editor decide or towards using digits in far more numbers than Centrx prefers.

As to the question of robots, no matter what is decided in matters of style, bots should never be used to make changes in style unless the bot can be counted on to never change any quoted material. In this case since even numbers under 10 (ten) may be expected to be written with digits in technical material, a bot is even less likely to ever be a viable option for changing representation of numbers to fit a standard.

I am changing "non-scientific" to "non-technical" in the guide as I think that probably better represents what was intended and what is intended by those who have discussed the matter here. If anyone disagrees, change it back or discuss it here or both.

Jallan 19:17, 17 Aug 2004 (UTC)

Unique is relevant in that it is not necessarily the case "the Wikipedia style sheet should not normally be repeating material found in most other style sheets" when it may be at times necessary to specify a style on the Wikipedia that is not in accordance with those style sheets which do not pertain to "other encyclopedias on the web an other sites providing much factual information on the web" - Centrx 01:58, 18 Aug 2004 (UTC)

The very reason I brought this up is because numerous times others have reverted certain number edits while pointing to the style guide as the incontrovertible reason for doing so. No matter what else, this style guide must still be changed to reflect that it be the editor's choice. - Centrx 01:58, 18 Aug 2004 (UTC)

Perhaps you might provide examples of such reversions. It is sometimes hard to understand why someone sees a problem without concrete examples. I see "may be represented" as incontrovertible in indicating that number ten (10) and up can be represented either by digits or by words. Whose choice, if not the editor? Jallan 18:10, 18 Aug 2004 (UTC)

Discussion following consensus of the above

this first comment is in response to the last comment of the above, superordinate section

They are lost in various edit histories, but they were for numbers of the values we are discussing here and the justifying comments were very simply along the lines of see Manual of Style (dates and numbers) or according to..., etc. I think may be represented by numerals could be ambiguous because it does not specify that they may be spelled out. Also, it would allow, for instance, 43649 to be spelled out, which is something that would not really be appropriate. Also, should not five hundred be spelled out? So, how about

In ordinary prose, whole numbers from zero through ten, and any such whole numbers followed by hundred, thousand, hundred thousand, and million~~, and so on~~ should be spelled out. Whole numbers from zero through one hundred, and any such whole numbers followed by hundred, thousand, etc. may be spelled out or represented by numerals. Other numbers should be represented by numerals, except where they appear as the first word in a sentence, in which case they must be written out in full.
Numerals should be used where appropriate in dates, times of day, currency, or technical prose, especially in tables or other areas where space is confined.

? - Centrx 17:35, 21 Aug 2004 (UTC)

I could agree to most of this. I'm not sure about "whole numbers followed by hundred, thousand, hundred thousand, million, and so on should be spelled out." Why? I don't see that "ten million" is any better than 10,000,000 and in fact, given the differences between commonwealth and US meanings of million and billion it would be better to use numerals. older ≠ wiser 22:54, 21 Aug 2004 (UTC)

This is the written word, not a mathematical text, not a confined space. I don't see why one should be spelled out but not one hundred. As for US vs. Commonwealth, yes I do think it might be appropriate for numbers greater than billion (million is common between the two). Although, as the OED states, billion in the US sense is increasingly common in Britain. - Centrx 23:36, 21 Aug 2004 (UTC)

Um, no, million is still a source of confusion. As for space, so what. What about consistency. Why make exceptions for some numbers simply because they have zeroes. That makes no sense to me. older ≠ wiser 00:02, 22 Aug 2004 (UTC)

Simply, there is no confusion in the English language with million. Whereas the OED lists two unique definitions for billion, the US and British variety each, it does no such thing for million. Million as 1,000,000 has been the definition since the fourteenth century through today and the standard for British writers. Whereas billion is 1,000,000,000 in the US and 1,000,000,000,000 in Britain, what is million in Britian[sic] if not 1,000,000 and what is 1,000,000 if not a million? A British milliard is 1,000,000,000 like the U.S. billion, but the OED states that this term has largely been superseded by billion and nevertheless it is no confusion of million. As I write these long numbers I now see more the propriety of spelling out large numbers. As for the reason to have a different rule for "some numbers" with "zeroes", the exception is rather that non-round numbers are not spelled out and the others are standardly spelled out. The reason for this is that it is unwieldy to spell out four million twenty thousand two hundred and fifty-five whereas spelling out round numbers that can be expressed concisely, as four million is sensible. Note also that it is especially important to spell out approximate numbers. Rather than referring to the geology of 1,000,000 years ago, the reference should be to one million years ago (things like 1.53 million can also be used) whereas if the number would be indicated in numerals by this rule, it would be a relatively specific number. - Centrx 18:17, 22 Aug 2004 (UTC)

Yes, you are correct about million. I was misremembering the discussion about the milliard. However, I still would not want an absolute rule about this. I quite agree with you that it is better to spell out the term when dealing with approximations. However, if the what is meant is exactly 100, 1,000 or 1,000,000 then I see no advantage to requiring that the number be spelled out. older ≠ wiser 00:34, 23 Aug 2004 (UTC)

So[,]

In ordinary prose, whole numbers from zero through ten should be spelled out. Whole numbers from zero through one hundred, and any such whole numbers followed by hundred, thousand, hundred thousand, and million may be represented by numerals or spelled out, especially when approximate. Other numbers should be represented by numerals, except where they appear as the first word in a sentence, in which case they must be written out in full.
Numerals should be used where appropriate in dates, times of day, currency, or technical prose, especially in tables or other areas where space is confined.

But, what can we say about numbers greater than billions? Should 6,000,000,000,000 really represented by numerals when it is so long? - Centrx 22:48, 25 Aug 2004 (UTC)

Currency style/disambiguation

I hope I've found the right spot for this. I'm new to Wikipedia, but in my explorations so far I've noticed that some (most) authors (probably Americans like myself) have a tendency to use the dollar sign ($) when discussing monetary amounts in articles. The dollar is a popular currency name worlwide[sic] and therefore somewhat ambiguous. I can't seem to fine[sic] a style guide entry covering formatting of currency values, but would suggest that using the ISO 4217 currency codes rather than symbols and avoiding multiple inline conversions (since exchange rates are fluid) would be the best NPOV style.

ex. 1,000,000 USD rather than $1,000,000.

Thoughts? Comments? Autiger 22:39, 16 Aug 2004 (UTC)

That usage would depend on us believing that Americans know what USD means. I am not sure that is true. Rmhermen 23:13, Aug 16, 2004 (UTC)

I'm ignorant of this matter. But if you can cite prominant style sheets recommending this and provide examples showing that it is in use (outside of technical material directly dealing with currency exchange) then you can make a good case for recommending their use. I agree that when there is any amiguity[sic], currency amounts should not be expressed in US dollars without some indication that the article is not talking about Canadian dollars or Australian dollars or some other dollars. But "$1,000,000 (US)" and "$1,000,000 (Austalian)" is probably more understandable to more readers and just as unambiguous as the USD codes in such doubtful cases, e.g. if someone from Canada were recorded as buying property in the US. Jallan 23:56, 16 Aug 2004 (UTC)

Does it matter if Americans know what USD means, if it's hyperlinked? — OwenBlacker 12:50, Aug 17, 2004 (UTC)

Yes it matters. We try to make our articles readable, not masses of technical jargon. Rmhermen 14:52, Aug 17, 2004 (UTC)

I can't seem to find any other style guide support for the USD usage. Most make a POV currency assumption based on the geo-political area within which they are located. [I].e. US publications assume a US dollar, etc. All others are specified, as often before the monetary amount as after: ex. AU $1,000 vs. $1,000 AU. The UK Times Online goes so far as to assume the most widely known of any given monetary unit and only denote others: the dollar is US unless A$ (Australian) or C$ (Canadian)[;] the pound is UK[;] the franc is France. I suppose a POV assumption within the context of the article is usually going to be ok ([A] $34M building project in Chicago, IL, US is assumed to be US dollars.) In any case, should there be a style guide entry to codify or will people be left to their own devices? Autiger 15:51, 17 Aug 2004 (UTC)

Tbh, I'm quite a fan of The Times'[s] standard, where major reserve currencies can be left as just a symbol and disambig'd where necessary (so $100, €100, ¥100, £100, C$100, A$100, NZ$100 etc[.]). It could be helpful to suggest [but not require] that the currency symbol be hyperlinked, for clarity's sake, if it's not blindingly obvious from the article. — OwenBlacker 17:17, Aug 17, 2004 (UTC)

External reference: Times Online style guide see dollar entry. Autiger 17:35, 17 Aug 2004 (UTC)

Since this is text, why not just use "1,433,433 US dollars" or "one million Australian dollars" ? - Centrx 02:02, 18 Aug 2004 (UTC)

In Canada one hears "five hundred dollars US" more often than "five hundred US dollars" though both are used. The latter is more concrete and tends to call up an image of physical US dollar-bills, but not so much so that it feels odd in most circumstances. In writing (outside of currency exchange discussion which would be likely to use special codes) one would normally output "$500 (US)" or "$500 US". See Google Search: $500 US which comes up with samples of both. See also "a pair of boots cost $500 (confederate)" at Delisle Antiques: Civil War Items.

I would recommend this usage as most readible for non-technical use when the desire is simply to make clear which currency is intended when not obvious from context. Sometimes a phrase like "all prices in Australian dollars" early in an article is in itself sufficient. I think the style sheet should call attention to the problem, should mention the solutions discussed here, all of which have their place, and urge that an appropriate method of disambiguating currency should always be used when there is any reasonable possibility of misunderstanding on the part of the reader. That is what is most imporant. Which method does not matter too much, as long as it is clear and is appropriate to the article. It would seem odd to use something so modern and technical as the code USD when referring to prices in the US a hundred years ago. Jallan 18:45, 18 Aug 2004 (UTC)

In Argentina, the common format of denoting US Dollars is u$s, because Argentine Pesos also use the "$" sign.