Wikipedia talk:Manual of Style/Dates and numbers/Archive 125

This is an archive of past discussions. 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 120

←

Archive 123

→

On date autoformatting

I've just found an excellent example of why it's EVIL:

"Caroline Herschel died at Hanover on 1848-01-09 in full possession of her mental faculties."

Here's the link to me fixing the awful text: [1]

This has been in the article for over a year and a half. But date autoformatting meant noone noticed that the date was given, in prose, 1848-01-09. However, every non-Wikipedian reader of the article saw it that way. Shoemaker's Holiday ^{Over 200 FCs served} 21:59, 25 August 2009 (UTC)

That's one reason why I turned my prefs off. JIMp_talk·cont 22:39, 25 August 2009 (UTC)

(*muffled sounds heard through a handkerchief over my mouth*) Mmmmfff, mmnnunffth. Unngth. Nnnnnff!! Greg L (talk) 02:23, 26 August 2009 (UTC)

First symptom of the swine flu. :-) HWV258 22:18, 26 August 2009 (UTC)

I know the death knell has tolled, but since you brought it up: Because no one set a default date format for IPs doesn't make autoformatting evil, it makes this autoformatting a flawed implementation that few were interested in fixing. —Ost (talk) 23:10, 26 August 2009 (UTC)

- How many people read Caroline Herschel, how many read to the end, and how long will it take them to notice the unidiomatic comma now placed after January? The original blunder has the same excuse. Septentrionalis PMAnderson 18:07, 27 August 2009 (UTC)

It's gone now. JIMp_talk·cont 18:36, 27 August 2009 (UTC)

National variety of English for MOSNUM

This guideline contains a mixture of UK and US spelling, especially with respect to SI units. Apart from any sections that specifically address national varieties of English, shouldn't we make it consistent? (I don't care which variety is used.) --Jc3s5h (talk) 11:45, 21 August 2009 (UTC)

In the actual text of the guideline (black) sure, that should be made consistent. The examples (green) can and should be inconsistent, as they present text as found in various articles, and different articles will use different Englishes. --___A. di M. 12:29, 21 August 2009 (UTC)

There has long been a tendency for MOSNUM’s guidelines to give a large amount of deference to the style used by the first major contributor. I think that is OK to a point and serves a valuable purpose; it ensures contributing to Wikipedia is an enjoyable hobby, which is necessary for an encyclopedia that has an all-volunteer body of editors. I think that if an article has been created or greatly expanded by a particular, primary shepherding author, great weight should be placed on retaining the dialect of spelling used in that article. Certainly, the dialect should be consistent throughout (no “colour” in one paragraph and “realize” in another). However…
I’ve long held that great weight should also be given to *what* the article is about. If it is a generic article, like water, then we can expect the normal distribution with our readership (which is about 25% American, as I recall). Accordingly, “first major contributor” works fine and should avoid conflict amongst editors. However, if the article is about Spokane River Centennial Trail or Boston Red Sox, there will obviously be a much higher proportion of American readers and the dialect used in the article should be American English. If we had an article on the Bondville Miniature Village (something I visited while in England when I was in my 20s), it would, IMO, be most appropriate to be in British English. This mix of guidelines best serves, in my opinion, what is the primary objective of any encyclopedia: writing in a way that seems most natural, fluid, and least confusing for the target audience Greg L (talk) 19:10, 21 August 2009 (UTC)

??? A.'s comment was about the variant of English used within the text of the guideline itself, i.e., the idea was that the guideline itself should follow WP:ENGVAR. Also, "colour" + "realize" is a bad example because it's perfectly good British English as promoted by Oxford University Press. Now "colour" + "analyze", that would be atrocious. What you are describing is already a part of ENGVAR, and I am not aware of any problems with its implementation. Hans Adler 19:45, 21 August 2009 (UTC)

Maybe there isn’t any problem. I must profess I am not up-to-speed on how this issue fleshed itself out over the last several months. There used to be abundant conflict over the nuances. Maybe it's all settled. Greg L (talk) 20:04, 21 August 2009 (UTC)
Huh? WP:ENGVAR has been stable ever since I can remember; at least, I can't find any non-trivial difference between its current revision and the latest 2007 version. (And this discussion is supposed to be about the text of the guidelines themselves, not articles.) --___A. di M. 20:45, 21 August 2009 (UTC)

The ArbCom decision on which this is based (like the text of MOS) gives greater weight to an established period of stability (which it is not worth having Anglo-American wars over). Often this will go back to the first major contributor, but we only look at him when there has been no period of stability. This may appear less arbitrary. Septentrionalis PMAnderson 20:25, 21 August 2009 (UTC)

Most if not all of the style guides happen to use US spelling. Of course, the green examples can and should be in a variety of spellings. Tony (talk) 03:44, 22 August 2009 (UTC)

(slight diversion) For what it's worth, Talk:War of 1812, which naturally has British, Canadian and U.S. readers and editors, with a topic that relates to all three countries, recently reaffirmed a broadly-accepted consensus, following WP:ENGVAR, that British spelling should be followed in the article because its first major contributors used British spelling. The one exception was locations like New York Harbor, whose spelling follows the country they're in, unless it can be clearly shown that the locals used and use a different one. —— Shakescene (talk) 05:46, 23 August 2009 (UTC)

Probably an example of never being stable, like the infobox which keeps saying different things about who won, depending on which nationalist has been through last. Septentrionalis PMAnderson 23:33, 27 August 2009 (UTC)

....only to be reverted as quickly as possible by common trans-Atlantic consensus. —— Shakescene (talk) 05:17, 28 August 2009 (UTC)

That seems reasonable, as the Wikimedia Foundation and its servers are based in the United States. (That argument holds no weight for article-space, but for project-space it seems a reasonable standard.) Powers ^T 13:47, 28 August 2009 (UTC)

Apply the same rule as we do to articles. Simple. Why not? Most (all?) of these MOS pages have been predominantly (completely?) written in US English (examples not counted) for most (all?) of their lives. Let 'em stay that way (or whatever other consistant way each of them've been stable at). Inertia. JIMp_talk·cont 15:11, 28 August 2009 (UTC)

Suggestion to avoid use of {{val}} in the MoS

The {{val}} template is used to make it easy to format a value according to the MoS. Using the template in the MoS itself causes a "circular referrence" where the MoS depends on the output of {{val}} and {{val}} is supposed to depend on what the MoS says. I suggest the MoS uses hard-coded examples; otherwise anyone could change {{val}} and the MoS would change with it. — SkyLined (talk) 23:43, 22 August 2009 (UTC)

I'm not sure how likely this is; if {{Val}} is changed by consensus, MOS probably should change; if it's done as a means of pushing one user's opinion about format, the user will certainly be reverted and probably banned.

I was about to suggest substing {{val}} where it occurs, but we don't want all those ifeqs in our text. Septentrionalis PMAnderson 00:30, 23 August 2009 (UTC)

If you want to subst {{val}}, you'll have to subst those {{#ifeq ...}}'s as well (and any other templates) until you end up with the HTML output of {{val}}. Also, different browsers and even different version of the same browsers may render the same HTML differently (there have been issues caused by this in the past). To prevent all these issues, it would be best for somebody to render an example correctly and create an image of the output. This image can then be used in the MoS because all browsers should render the image exactly the same, regardless of any template. — SkyLined (talk) 01:13, 23 August 2009 (UTC)

What PMA said: If MOS or MOSNUM changes its guidelines, then {val} should change to reflect that. There’s no need to worry about the ramifications of what happens if you put the horse before the cart. Unless there is a clear, widespread, and convincing consensus to change {val}, editors should have every expectation that it will be stable and their articles aren’t going to change. Greg L (talk) 04:01, 23 August 2009 (UTC)
- And all our guidelines can do is to control what HTML we use. We cannot go beyond that; Wikipedia is a collection of HTML strings. If browsers differ in the interpretation of some strings, we have the choice of not using the strings, or of accepting that some users will get something odd. This is why we long avoided the use of thin spaces; too many browsers failed to render them.Septentrionalis PMAnderson 12:13, 23 August 2009 (UTC)

Special:ExpandTemplates can convert {{val|299792458|u=m/s}} to 299,792,458 m/s automatically. I don't think using images would be useful, for the reason PMA gave above. --___A. di M. 14:42, 23 August 2009 (UTC)

Moreover, I don’t particularly see the need for editors to have to suffer the hassle of going back to every article they’ve used {{val}} to now create code like this: {{nowrap|1=6.02214179(30) × 10−23 [[kilogram|kg]]}} (taking care to choose a true minus sign from the tool box for the negative exponent rather than use the visibly shorter keyboard hyphen) to obtain 6.02214179(30)×10⁻²³ kg.
Why? Because there is clearly no consensus in the community to change {{Val}}. Coding {{val|6.02214179|(30)|e=-23|ul=kg}} is much simpler. Like I wrote above, editors have every expectation of stability in the articles in which they’ve employed the {val} template. That little “padlock” icon in the upper right-hand corner of many of our templates—including {val}—is a *pinky promise* guarantee to the community of stability in a template so they can start using it with confidence; “anyone” can not change {{val}}. The select few who have the ability to modify templates are expected to behave with the utmost care and responsibility.
Behavior that instills FUD about the stability of a now-well-used template is most unfortunate and such an offender should expect to have their privileges stripped. When there is as much widespread unanimity to change {val} as there was to design and make it in the first place, we’ll let the gate keepers know. Greg L (talk) 19:54, 23 August 2009 (UTC)

- - I don't particularly see the need for editors to have to suffer the hassle of going back to every article they’ve used {{Val}} on. Nobody's suggested that; SkyLined suggests hardcoding the examples on this page, only, to set the standard; {{val}} is a tool for achieving that standard. This seems harmless, but it's not worth my taking the trouble to do it; if anybody else does, provided they don't change what this page actually says, that's fine by me. Septentrionalis PMAnderson 20:13, 23 August 2009 (UTC)

Yeah, I understood all that. I’m objecting to the premise underlying his making the suggestion for the need to horse around on WP:MOSNUM with the {{val}} template: …otherwise anyone could change {{val}} and the MoS would change with it. We shouldn’t have to worry about “anyone” changing {val} since there is no mandate from the community to do so. Greg L (talk) 20:27, 23 August 2009 (UTC)

I am sorry, I didn't make my point clear enough so the discussion has mostly been about some minor points that I was trying to make, not the main point. My main point is that if we use {{val}} in the MoS, you can never really know if {{val}} works correctly nor if the MoS you're looking at is the MoS everybody else is looking at. Some examples:

We know {{val}} does not work correctly on IE6 (discussion here). Somebody who would read the MoS using IE6 could therefore make incorrect assumptions about how numbers should be formatted. If we use screen-grab images as the base line, this cannot happen.
In the future, somebody may make a valid change to {{val}} and confirm the change works in HIS/HER browser only. This can potentially introduce a browser specific issue in another browser. To the editor, the output of {{val}} will look correct and follow the MoS but both may appear incorrect to other users. These other users can never really know the output of {{val}} is incorrect, because the MoS uses {{val}} and therefore has the same issue. If we use screen-grab images as the base line, this cannot happen.
Consider the discussion about thin spaces below; I may like the way it looks in my browser and everybody else may like the way the exact same HTML looks in their browsers, so we all agree to use that specific HTML. However, as soon as we switch browsers, we may notice that we agreed on completely different things. If we use screen-grab images as the base line, this cannot happen.

If we never run into any of the issues I am describing, then using images will have been a waste of time for the person who created them. But if we do run into the issues I am describing, we will:

have errors in wikipedia until they are noticed, which I think we should prevent at all cost,
have to undo the work of the original editor, which wastes his/her time,
have to have very long and boring discussions about how to fix it, which wastes everybody's time,

Because I believe the issues I described will happen at least once in the future, we will be better off using images. The only valid reason (that I can think of) against this idea is that it takes some initial work and makes maintaining the MoS slightly harder because changes to the formatting of numbers may require updates to the example images. However, I expect changes to the images will mean changes to {{val}} will be needed as well, so there is a chance that we'll run into the problems I described as well. Since these problems waste more than than changing images, I do not think it is a very good reason to object. — SkyLined (talk) 15:21, 25 August 2009 (UTC)

But the articles use HTML, not images. So saying "It is recommended that numbers be marked up this way" makes sense, but saying "It is recommended that numbers look like this" is meaningless at worst, and useless at best, because editors won't know how to make numbers look like that. BTW, it is completely impossible make numbers look the same exact way for all readers: details such as the choice of fonts will vary (e.g. the diagonal stroke in "7" is straight in my font and curved in Greg's), so we must make an abstraction somewhere. Separation of presentation and content is a key principle in modern Web design. (The issue is the same as narrow and broad transcriptions of pronunciations: saying that mat is pronounced [mæʔ͡t̚] (a narrow transcription) is a precise description of the movements of the tongue in an utterance of that word, and it will only be accurate in some dialects and in some contexts. For example, young speakers in England use a lower vowel than that transcription suggests; many speakers will pronounce the t in the mat is differently from that in the mat goes; people with a cold will pronounce the m differently than when they are fine, etc. Saying that it is {{pron-en|mæt}} (a broad transcription) means that the first sound is the same as in my and the last two are as in bat, which is practically almost always true. Narrow transcriptions would be useless in dictionaries, but broad ones are very widely used.) So I think that using pictures is useless, because we cannot recommend precise appearances, only implementations.

I never suggested we recommend precise appearance or require the HTML implementation to look exactly like the example image. I am suggesting that we describe what the style should be (as we currently do) and provide an images as an example (instead of the HTML examples that use {{val}} and can differ from browser to browser). {{val}} can serve as a recommended implementation. — SkyLined (talk) 20:41, 26 August 2009 (UTC)

Who changed {val}?

Per agreement reached here on WT:MOS Archive 97, {{delimitnum}} and {{val}} were supposed to have thinspaces ( ) on both sides of the times (×) symbol. Many people objected to the full-width spaces and wanted no spaces alongside the × symbol. The compromise, which made everyone happy, was to use thinspaces. My recollection is that {val} originally conformed with this agreement and generated scientific notation with the proper thinspaces. Perhaps I am mistaken, but this clearly isn’t the case now. Did someone change it? If so, please change it back.

Below, the top-most expression has the exponent, hard-coded with thinspaces on both sides of the × symbol. The one below it was created with {{val}}:

6.02214179(30) × 10⁻²³ kg
6.02214179(30)×10⁻²³ kg

Greg L (talk) 20:20, 23 August 2009 (UTC)

Some browsers have problems with thin spaces; on such browsers, if an expression like "6.02214179(30) × 10⁻²³ kg" were found at the end of the line, the browser would rather extend the line to fit the " × 10⁻²³" part in it, than force the entire expression at the start of the next line. The current implementation achieves the same visual appearance without this problem. --___A. di M. 20:31, 23 August 2009 (UTC)

OK. If some browsers don’t like the thinspaces, then that’s a good reason not to use them. Let’s address, though, an issue of fact: that {{val}}’s current technique appears the same as thinspaces. On both Safari 4.0.3 (the very latest) and Firefox 3.5.2 (as far as I know, the very latest), the {val} template looks exceedingly similar to regular non-breaking (full width) spaces (as shown below:

6.02214179(30) × 10⁻²³ kg (this one is hard-coded with non-breaking, full-width space)

6.02214179(30)×10⁻²³ kg (this one uses {{val}})

Using Safari, there is a little bit of difference ( {val} being thinner), but on Firefox (on a Mac), the gaps on both sides of the × symbol are absolutely identical in width—even when the text is enlarged one and two sizes.

Using CSS to get Firefox to match the apparent visual width of narrow spaces, I have to use of 0.2 em:

6.02214179(30)×10⁻²³ kg (hard-coded with 0.2-em-width gaps)

6.02214179(30) × 10⁻²³ kg (hard-coded with  )

On my Mac using both Safari and Firefox, the above two appear either absolutely identical or very, very close. On my system, this is the compromise appearance editors agreed to on WT:MOS. What do you see on your system? What width does {val} currently used? Greg L (talk) 21:00, 23 August 2009 (UTC)

With Firefox on Ubuntu: the first two examples have identical spaces before the times sign, but in the second one the space after the times sign is two-pixel wider than in the first one; the fourth one has a space before the times sign which is one-pixel wider than in the third one, but the spaces after it are the same in the last two examples. BTW, I think that the width of space characters depends on the font, not on the browser (except if a font doesn't have a glyph for a particular character, "sane" browsers will substitute it with a glyph from another font of their own choice). --___A. di M. 21:20, 23 August 2009 (UTC)

Passed through Special:ExpandTemplates: it uses 0.3em at the left of the × and 0.15em at its right. I think they had better be symmetrical; also, I think that they'd better be no narrower than the gaps within the numbers (0.25em), as seeing 1.234567×10¹² (difference exaggerated for clarity) gives the visual impression that only the 567 is being multiplied (even if that makes no sense). --___A. di M. 21:26, 23 August 2009 (UTC)

Let’s see… Yes, different fonts too. Some browsers resolve to 0.05 em while others only to 0.1 em. And, as you’ve seen, gaps preceding sanserif “1”s look different than those next to the other digits. While working on {val}, these differences were exploited to achieve an appearance that came across as a reasonable compromise for everyone. How does the following appear to you (and anyone else) on a variety of systems:

6.0221412570×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412571×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412572×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412573×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412574×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412575×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412576×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412577×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412578×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412579×10⁻²³ kg (hard-coded with 0.25-em left of × and 0.2-em right)

6.0221412579×10⁻²³ kg (made with {val} template for comparison with the value immediately above it)

Reasonably balanced and narrow on both sides of the ×? I have some errands to run. But I will spend more time looking at this on my iPhone and with other browsers. Greg L (talk) 22:24, 23 August 2009 (UTC)

In my font all the digits have the same width, including 1 (which has a serif despite being a sans-serif font, I guess for this very reason). I know there are fonts with narrower 1. (A site had some text after an automatically updating count-down in the same line, and when I viewed it at the university the text after the count-down shifted back and forward whenever an 1 appeared or disappeared in the count-down; what a wart.) But I think they are a minority of fonts, and also I suspect that having {{val}} adjust the width of the gaps based on the following digit would be very complicated. --___A. di M. 23:50, 23 August 2009 (UTC)

I’m sorry, I didn’t explain myself clearly. The above progression isn’t sensitive to what digit the significand ends with. It is merely a case where the two gaps sandwiching the × symbol have different values. Isn’t this what {val} is currently doing? If so, the only difference is that these gaps are narrower. They are 0.25 em on the left of the × and 0.20 em to the right—completely irrespective of what occurs in the significand. I showed all ten possible digits that could precede the ×-symbol to ensure that—overall—it seemed balanced. The code is fixed at ×10. Different browsers can have a precision of 0.05 em and still others go only to a tenth. This is being exploited here to keep the appearance as identical as possible across platforms.
My objective here is only to make the gaps closer to the original compromise achieved on WT:MOS, where some editors wanted a full‑width non‑breaking space and another camp thought it looked too much like a formula and wanted no spaces. The thinspace was well-received and all agreed to that compromise.
Here’s another try:

6.0221412570×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412571×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412572×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412573×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412574×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412575×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412576×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412577×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412578×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412579×10⁻²³ kg (hard-coded with 0.2 em left of × and 0.15 em right)

6.0221412579 × 10⁻²³ kg (thinspace characters left and right of the ×. Compare to the value above it.)

6.0221412579×10⁻²³ kg (made with {val} template for comparison with the two values above this line)

6.0221412579 × 10⁻²³ kg (non‑breaking spaces left and right of the ×. Compare to the three values above this line.)

Again, this is fixed code that isn’t dependent upon what’s in the significand. Do the gaps sandwiching the × symbol appear reasonably narrow (similar to a thinspace) and balanced on both sides of the × symbol? It’s getting close, for me. But I’m going to look at this too with some more browsers. Greg L (talk) 01:57, 24 August 2009 (UTC)

P.S. So far, it appears to me that {val} produces results that are well balanced on both sides of the × symbol, but it produces gaps that are quite similar to the width of a full-size non-breaking space (rather than the thinspace it originally was). Greg L (talk) 05:39, 24 August 2009 (UTC)

Tests:

6.0221412579 × 10⁻²³ kg (thin spaces)
6.0221412579×10⁻²³ kg (0.25em each side)
6.0221412579×10⁻²³ kg ({{val}})

___A. di M. 10:13, 24 August 2009 (UTC)

Tests (upon which the below screen shots were based):

6.0221412579 × 10⁻²³ kg (thin spaces)
6.0221412579×10⁻²³ kg (0.25em each side)
6.0221412579×10⁻²³ kg (hard-coded with 0.25 em left of × and 0.15 em right)
6.0221412579×10⁻²³ kg ({{val}}
6.0221412579 × 10⁻²³ kg (non‑breaking spaces each side)

I should have been taking screen shots so I could tell what is going on here. The {val} template, as of this writing right now, looks to be doing a ~~fine~~ satisfactory, balanced job on both sides of the ×. In both Safari and Firefox, it is more balanced than a thinspace and it is narrower than a full-width, non-breaking space. ~~I like it. Or… I like it now.~~ Spacing of 0.25 x 0.15 em seems to look better. Greg L (talk) 18:12, 24 August 2009 (UTC)

P.S. Based on my tests, it appears to me that 0.25 x 0.15 em best achieves the expectations of those who participated in the WT:MOS compromise. What are you seeing? Greg L (talk) 17:20, 24 August 2009 (UTC)

Here is what I am seeing on a Mac running OS X 10.5.8 (the latest):
The below screeshot is what I see with Safari 4.0.3 (the latest) and was captured 17:21:31, 24 August 2009 (UTC):

The below screenshot is what I see with Firefox 3.5.2 (I think is the latest) and was captured 17:22:07, 24 August 2009 (UTC):

As you can see, 0.25 x 0.15 em on my platform combinations appears to have a balanced appearance that is narrow so it ought to meet the expectations of those who participated in the WT:MOS compromise between no regular spaces and no spaces at all. How does 0.25 x 0.15 em look on yours? Please post screen shots if necessary. Greg L (talk) 17:58, 24 August 2009 (UTC)

The below screenshot is what I see with Safari on iPhone OS 3. It was captured 18:20, 24 August 2009 (UTC):

As you can see, 0.25 x 0.15 em on the iPhone also looks really good. Greg L (talk) 18:28, 24 August 2009 (UTC)

(edit conflict) It looks best to me, too. This was taken with Firefox 3.0.13 under Ubuntu 9.04:

I s'pose this should be taken to Talk:Val rather than here, though. --___A. di M. 18:31, 24 August 2009 (UTC)

Gap proposal

Tony, Srleffler, and SamBC were the three editors who had input on this issue, as memorialized here on WT:MOS Archive 97. What I am trying to do is carry out my self-appointed fiduciary responsibility, of sorts, in shepherding {{val}}. Those three had conflicting ideas on what looked good and both instantly bought into the compromise solution I showed them, which—at that time—utilized thinspaces. If those three think it looks fine, then I’m certainly good. I’ll contact the three of them here shortly and see who else might weigh in on this before suggesting anyone move on tweaking the width of those two gaps in {val}. Greg L (talk) 21:58, 24 August 2009 (UTC)

Comparisons

6.0221412579×10⁻²³ kg (no spaces)
6.0221412579 × 10⁻²³ kg (full-width, non-breaking spaces)
6.0221412579 × 10⁻²³ kg (Original compromise: thin spaces)
6.0221412579×10⁻²³ kg (Proposed: The tweak of the CSS version of thinspaces)

This ~~last~~ proposed one looks very good to me. In the MOSNUM examples, one might not link all or any of the "kg"s. Tony (talk) 22:56, 24 August 2009 (UTC)

Yeah, we used the “ul=kg” (unit link equals kilogram) for these examples. We could have just as easily made it “u=kg”, which wouldn’t have linked. Just showin’ off. Greg L (talk) 01:25, 25 August 2009 (UTC)

Thanks for all your work on this issue, Greg. I like the proposed CSS thinspaced version the best. I actually like it better than the no-spaces version, perhaps because the example also has thin spaces between every third digit. With that spacing, having thin spaces around the times makes sense.--Srleffler (talk) 02:55, 25 August 2009 (UTC)

(*Overheard at an auction*) Going once… Greg L (talk) 02:37, 26 August 2009 (UTC)

The proposed CSS version looks good to me on IE7 and Chrome 2.0.172.43. My main concern about the CSS version is that when copied and pasted, you get 6.0221412579×10−23, which could be mistaken for a multiplication followed by a subtraction—{{val}} does this already. Having some sort of space character (instead of a CSS gap) might serve as a (cryptic) clue that the − (being unary minus in an exponent) actually has a higher operator precedence than ×. As a better resolution, we could add a caret to the copied output, using something like ^ → ^ (a hidden caret). This would be especially valuable when copying a formula containing {{val}}.
Generally speaking, does anyone have any thoughts on whether there's any stylistic issue with having different spacing between the terms in a {{val}} expression, and an equation (which can be formatted with regular spaces, per WP:MOSMATH)? TheFeds 03:59, 26 August 2009 (UTC)
My personal preferences would be using the same spaces in exponential notation as in any other multiplication; as for the objection,
Consider 1.37×10¹³ × 4.29×10⁹ vs 1.37 × 10¹³ × 4.29 × 10⁹, and for completeness 1.37×10¹³×4.29×10⁹. Now, the unspaced (third) version is fairly obviously horrible, while the partially-spaced (first) version clearly indicates that you're talking about two distinct numbers being multiplied, with those numbers written in standard form, while the fully spaced (second) version looks like 4 separate numbers
I'd say that, since multiplication is associative, (1.37 × 10¹³) × (4.29 × 10⁹) equals 1.37 × (10¹³ × 4.29) × 10⁹, so that's a non-issue. But the horse appears to have long been dead. (And I like the suggestion of an invisible caret.) --___A. di M. 11:04, 26 August 2009 (UTC)

Regarding IE7 and Chrome 2.0.172.43, that’s good to know. Thank you very much.
As regards copy/paste, the CSS gaps in the significands are extremely convenient because they don’t appear in the pasted values, so Excel can use them directly. The gaps on each side of the × symbol are a moot point in Excel since you have to edit to change “×10” to “E” anyway. If an editor is doing work in Word or some other program and wants pretty looking results, they’ll just have to do a little work for themselves when they’re doing their homework.
As for chaining: What A. di M. said. Editors can also use middot; e.g. The “Theory of Everything” formula was eventually found to be nothing more than the product of 3.17×10⁷ · 6.2568945×10⁻²⁷ or use a hybrid; e.g. (3.17×10⁷) · (6.2568945×10⁻²⁷). In many circumstances, editors would start using math markup for this sort of thing. It’s rare that any single tool can “slice & dice” your food and lift the engine and trani out of your car. Greg L (talk) 14:43, 26 August 2009 (UTC)

(*Overheard at an auction*) (And with specific regard to exactly how wide the CSS gaps should be on both sides of the × symbol): Going twice… Greg L (talk) 19:09, 27 August 2009 (UTC)

Seeing no objections and universal agreement from the interested parties: Sold. Greg L (talk) 02:19, 29 August 2009 (UTC)

Changes applied at User:Greg L's request. — SkyLined (talk) 08:42, 29 August 2009 (UTC)

Defining 'internationally accepted units' and getting rid of the phrase 'region-specific'

I propose that the following wording be considered for the policy:

In place of this:

Current wording

Which units to use

In articles which are not region-specific, prefer internationally accepted units. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for particular classes of measurements. However, on region-specific topics, use the units used in the place the article is about, for example US customary units for US-related articles.
- When different parts of the English-speaking world use different units for the same type of measurements, add a conversion so that all English-speaking readers will be able to understand one of the units. For example, the Mississippi River has a length of 2,320 miles (3,734 km); the Murray River has a length of 2,375 kilometres (1,476 mi). (See § Unit conversions below.)

I propose considering this wording:

Michael Glass's proposal

Which units to use

Apart from US and some UK-related articles where US or Imperial measures are used, prefer internationally accepted units. These are units of the International System of Units (SI), non-SI units accepted for use with SI, units based on fundamental constants (e.g. Planck's constant), and other units that are widely used internationally for specific purposes.
- In general, add a conversion so that all English-speaking readers will be able to understand one of the units. For example, the Mississippi River has a length of 2,320 miles (3,734 km); the Murray River has a length of 2,375 kilometres (1,476 mi). (See § Unit conversions below.)

My intention is not to change the policy but to express it more clearly and concisely. Any comments? Michael Glass (talk) 12:34, 18 August 2009 (UTC)

Sounds reasonable to me.Tomwsulcer (talk) 10:55, 30 August 2009 (UTC)Tomwsulcer

Pmanderson's proposal

I would split the difference:

Which units to use

Apart from US and some UK-related articles where US or Imperial measures are used, prefer internationally accepted units. These are units of the International System of Units (SI), non-SI units accepted for use with SI, units based on fundamental constants (e.g. Planck's constant), and other units that are widely used internationally for specific purposes.
- When English-speakers use more than one unit for a given type of measurement, it is generally advisable to add a conversion so that all English-speaking readers will be able to understand one of the units. For example, the Mississippi River has a length of 2,320 miles (3,734 km); the Murray River has a length of 2,375 kilometres (1,476 mi). (See § Unit conversions below.)

This provides a rationale for the ruling, and allows for exceptions; there aren't many, but I foresee MGlass's text being used to demand conversions between calendar and tropical years, and other totally silly demands. Septentrionalis PMAnderson 12:51, 18 August 2009 (UTC)

I don't want to go round the houses, but I don't like "Apart from US and some UK-related"... many US articles use metric units, and using exceptio probat regulam this could suggest that SI is actively discouraged in US and some (unspecified) UK articles. This just does not cut the mustard; you can't get rid of the "regional" bit altogether, ugly though it be. For a start, write "Apart from most US-related" (since some use SI etc); and indeed since it simply says "prefer" why not cut that qualifying clause and put elsewhere? Below is not perfect but an attempt to show what I mean:

Which units to use

Prefer internationally accepted units. These are units of the International System of Units (SI), non-SI units accepted for use with SI, units based on fundamental constants (e.g. Planck's constant), and other units that are widely used internationally for specific purposes.
- In general, add a conversion so that all English-speaking readers will be able to understand one of the units. For example, the Mississippi River has a length of 2,320 miles (3,734 km); the Murray River has a length of 2,375 kilometres (1,476 mi). (See § Unit conversions below.)
Articles about US or UK subjects should use US or Imperial measures when appropriate, with conversions to SI unless the context makes that ridiculous.
- Take care to consider Canada and Ireland, which although largely metricated still use US or Imperial measures in some parts of daily life.

SimonTrew (talk) 12:56, 18 August 2009 (UTC)

Certainly an improvement. I would make the first sentence Wikipedia generally prefers internationally accepted units - although the phrase internationally accepted is both tendentious and incorrect; units accepted by the US and Canada, or Britain and Canada, are internationally accepted. If it is left as it is, some good soul will go through and switch George Washington to kilometers, quoting the first sentence in isolation. Septentrionalis PMAnderson 13:17, 18 August 2009 (UTC)

Yes, I did consider hoisting the "Articles about US... to the top of the example (even before "prefer internationally...", possibly). I agree about "internationally" and said so in the earlier discussion, but perhaps not as clearly. I think we can cut "internationally accepted units" completely since we immediately give its definition. For the other uses of "internationally" we can say simply "worldwide" (or is that equally contentious? Surely nobody will expect penguins in Antarctica to be getting out their slide rules?)

How about this? (Again, not intended as a final suggestion more something to bite on.)

Which units to use

Articles about people or places strongly associated with one place or time should use the units appropriate to that place or time. For the US this will generally mean US units, for the UK, sometimes Imperial units.
- Take care to consider Canada and Ireland, which although largely metricated still use US or Imperial measures in some parts of daily life, and used non-SI units for much of their past.
With that considered, Wikipedia prefers units of the International System of Units (SI), non-SI units accepted for use with SI, units based on fundamental constants (e.g. Planck's constant), and other units that are widely used worldwide for specific purposes.
- In general, add a conversion so that all English-speaking readers will be able to understand one of the units. For example, the Mississippi River has a length of 2,320 miles (3,734 km); the Murray River has a length of 2,375 kilometres (1,476 mi). (See § Unit conversions below.)

SimonTrew (talk) 14:06, 18 August 2009 (UTC)

How about

For articles not associated with a place or time, especially scientific articles, Wikipedia normally uses units of the International System of Units (SI), non-SI units accepted for use with SI, units based on fundamental constants (e.g. Planck's constant), and other units that are widely used worldwide for some specific purpose.

which is self-contained. Septentrionalis PMAnderson 14:28, 18 August 2009 (UTC)

(edit conflict) I hadn't thought about the possibility of misunderstanding the phrase "internationally accepted" that way [i.e. "units accepted by the US and Canada, or Britain and Canada"]. I still find it's unlikely that a reasonable person could accidentally misunderstand it, but we'd better prevent unreasonable wiki-lawyers from purposefully misunderstand it. I still don't like the "shopping list"-like explanation, which suggests that all SI units, all units based on fundamental constants, etc. are accepted (while the megasecond isn't accepted, and the Planck mass is only accepted in advanced theoretical physics contexts). Also, I'd rather go with "parts of the English-speaking world" than with "English speakers": no conversion could make all English speakers understand a measure such as "4.7 microfarads", for there are many who don't know what electric capacitance is; but anyone who knows what it is would measure it in submultiples of the farad, regardless of where they're from; so a conversion for that measure is unnecessary and useless. Let me give a try:

----

Which units to use

Except in the cases mentioned below, prefer the units in most widespread use worldwide. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as inches for display sizes and years for long periods of time.
- When discussing topics strongly associated with one place or time, use the units appropriate to that place or time. In articles about the present, for the US this will usually mean United States customary units, and for the UK this generally means Imperial units for some classes of measurements and metric units for others (see, for example, the Times Online style guide under "Metric").
When some parts of the English-speaking world would use a different unit than the one used in the article, generally add parenthetical conversions so that readers from those regions can understand the measurement: for example, the Mississippi River has a length of 2,320 miles (3,734 km); the Murray River has a length of 2,375 kilometres (1,476 mi). (See § Unit conversions below.)

----

~~I'm not sure about "general articles"; I think clearer alternative could be found.~~[Edited to incorporate point made above by Trew, which I hadn't read yet.] As for cases such as kilojoules v kilocalories in Italy (and I suppose Canadians and Irishmen can find other examples of that), it says "locally used", not "locally recommended". Per WP:BEANS, let's wait until some editor cites this guideline and some obscure law requiring [kilojoules|some SI unit] in [the EU|Ireland or Canada] to replace [kilocalories|some Imperial unit] with [kilojoules|some SI unit] in an [EU|Ireland- or Canada]-related article before we explicitly make that point. --___A. di M. 14:47, 18 August 2009 (UTC)

I like the general principle Simon put forth: Articles about people or places strongly associated with one place or time should use the units appropriate to that place or time. Arago used toises, and we should describe in toises - translating into meters and yards. Septentrionalis PMAnderson 15:31, 18 August 2009 (UTC)

Emended to incorporate that point. BTW, I think that "people or places" is too restrictive, as beer glass is neither a person nor a place (not in the obvious sense of "place", at least); also "articles" should be toned down, because an article discussing several topics can use different units for each one of them, as in the Irish road speed limits example (mph for historical limits, km/h for modern ones). --___A. di M.

I agree with A. di M. that, yes, the phrasing around "internationally accepted" is to avoid unreasonable wikilawyering (and I would love to get rid of the "internationally" and prefer, say "accepted worldwide"); but wikilawyering is exactly what that wording was intended to avoid, i.e. deliberate misinterpretation. I also agree that "people or time" (I didn't say people but places, btw, but it holds) is unduly restrictive, but of course we don't want to enumerate every possible kind of association. I also see the point about "articles", what I was trying to do here was restrict the recommendations to article space, not to e.g. MOSNUM talk space where obviously it is necessary to mix units, nor (which I think you had more in mind) to articles where it is right and proper to mix units.

How about "Articles or parts of them that are about things strongly associated with one place or time..." i.e. substitute "things" for the first "place or time" but keep the second "place or time" since that, I think, *is* what essentially defines which units are used, e.g. the size of beer glasses depends on the size the beer is served in which depends on the place. There may be a case for putting "people" in there too, since i can imagine for example the article on Anders Celsius is so strongly associated with the unit of measure that it is the person who causes the association, not particularly where or when he lived. But I don't want to enumerate it much further than that, if we can help it.

"Things" perhaps sounds vague, but I think is perfectly adequate. I suppose there will be wikilawyering about whether concepts are things and so on (yes they are. A thing is a noun. And most WP articles are about nouns. Even an article, say, Wikilawyering (this is in WP space but the redirect is from article space) is a noun, it may look like a verb but is an agentive noun (some would call it a gerund). I suppose there are some articles that are not in any sense about things, though.

The kilocalorie thing should probably be dropped from this thread and maybe start another for this special (unique?) case. I hate to want to write a special case into MOSNUM simply for it, and surely a reasonable editor would take WP:COMMON (as I did at Bacon in the example I gave above, and although that article is reasonably frequently edited, it has not been reverted). My only concern here really is that a picky GA or FA reviewer will point out an inconsistency here, where the inconsistency really is there in daily life and must be fairly reflected in WP.

Best wishes SimonTrew (talk) 12:30, 19 August 2009 (UTC)

Maybe "topics" is the word you are looking for instead of "things". --___A. di M. 16:54, 19 August 2009 (UTC)

SimonTrew's proposal

Yes, "topics" is much better. I just start a new subsection since we've not had a complete proposal written out for a while, and this is very similar to one above (has it been edited? I don't recall it being quite so similar) so here goes (the bold and strike are to indicate changes that may otherwise be too suble to be noticed; and are not intended to indicate proposed markup):

----

Which units to use

Except in the cases mentioned below, prefer the units in most widespread use worldwide. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as inches for display sizes and years for long periods of time.
- When discussing topics strongly associated with places, time or people, use the units appropriate to them. In articles about the present, for the US this will usually ~~mean~~be United States customary units, and for the UK ~~this generally means~~ Imperial units for some ~~classes of measurements~~topics and metric units for others (see, for example, the Times Online style guide under "Metric").
When some parts of the English-speaking world ~~would~~ use a different unit ~~than~~from the one used in the article, ~~generally~~ place conversions afterwards in parentheses so that all readers ~~from those regions~~ can understand the measurement: for example, the Mississippi River ~~has a length of~~is 2,320 miles (3,734 km) long; the Murray River ~~has a length of~~is 2,375 kilometres (1,476 mi) long. (See § Unit conversions below.)

----

I am not that het up about "different than", but since others may insist on "different from", we might as well do that and avoid needless argument. The other rewordings are basically just simplifications or to remove being over-specific (e.g. "readers from those regions" -> "all readers", surely the intent is that anyone reading the article can undertand the measures in it.) SimonTrew (talk) 12:16, 20 August 2009 (UTC)

Minor points: I'd still rather go with "readers from all over the world" or something like that, than with "all readers"; see the example about microfarads somewhere else in the thread. (BTW, is "time" rather than "times" in the inner bullet a typo?) --___A. di M. 12:42, 20 August 2009 (UTC)

Excellent proposal, explains the spirit very well and makes clear how all the details flow from it. Hans Adler 12:40, 20 August 2009 (UTC)

Thanks, Hans.

To A. di M. I don't think "time" was a typo, but after spotting it (which took some scanning for me) I do think "times" is better. I guess since it took me a while to spot, it is somehow acceptable in my dialect and doesn't seem particularly strange to me – time being continuous it can be used in the singular – but since people and places are plural (and the verb is in the plural) it seems more than sensible to put it into the plural, yes.

As for "all over the world", I still think "all readers" is good enough, but certainly don't want to endanger what I think is a good proposal worked out between several of us here for what is (as you say) very minor, and if you think the emphasis is needed then I can accept that. I will suggest "readers worldwide" (i.e. repeat "worldwide" rather than rephrase unnecessarily – and longer – as "all over the world"), but I am not going to lose sleep over it.

Thanks for the constructive criticism I don't know, shall we just change it now or soon, or does this need more discussion? SimonTrew (talk) 22:19, 26 August 2009 (UTC)

Headbomb's proposal

Leave it as is. It ain't broken, and at this point all we're doing is listing WP:BEANS scenarios. The original wording of "Use international units [...] usually this means SI, SI-related & units accepted with SI" is as both as clear and as vague as it needs to be. The subtleties are covered in the bullets. Headbomb {^ταλκ_{κοντριβς} – WP Physics} 11:31, 21 August 2009 (UTC)

I disagree. It is broken, because it fails to make clear what "regionally specific" etc. means, and just opens the door to wikilawyering about whether something does or does not obey MOSNUM. I think our intent here is to be clear, concise, and allow maximum room for manoeuvre by sensible editors, while deliberately cutting off wikilawyering, and I hope with my proposal (and of course the input of several others to refine it) we have done that rather well; and have done it in discussion with no effect on the policy until we get consensus, which I hope we are close to. SimonTrew (talk) 22:04, 26 August 2009 (UTC)

General comments on region vs. internationally accepted

All these wordings are confusing in that they say "use", but in most cases, what is really meant is "list first", because conversions are usually provided. It would be nice to think that editors would read the manual from end to end and remember everything, but that just isn't going to happen, so wording that does not require the editor to read a different part of the manual to understand that "use" usually means "list first" would be better. --Jc3s5h (talk) 14:34, 18 August 2009 (UTC)

That's why I added a sub-bullet about unit conversions to the first bullet in the "Which units to use" subsection , where there was none. --___A. di M. 15:06, 18 August 2009 (UTC)

Adjusting the wording in the light of the previous discussion

I agree that the wording is confusing in saying 'use' but meaning 'list first'. It is also inconsistent with later dot points. This is how the passage reads at the moment

Which units to use

Except in the cases mentioned below, prefer the units in most widespread use worldwide. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as inches for display sizes and years for long periods of time.
- When discussing topics strongly associated with places, times or people, use the units appropriate to them. In articles about the present, for the US this will usually be United States customary units, and for the UK Imperial units for some topics and metric units for others (see, for example, the Times Online style guide under "Metric").

When some parts of the English-speaking world use a different unit from the one used in the article, place conversions afterwards in parentheses so that readers from all over the world can understand the measurement: for example, the Mississippi River is 2,320 miles (3,734 km) long; the Murray River is 2,375 kilometres (1,476 mi) long. (See § Unit conversions below.)

Which units have priority

Except in the cases mentioned below, put the units in most widespread use worldwide first. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as years for long periods of time.
- With topics strongly associated with places, times or people, put the units most appropriate to them first. In US articles about the present, this will usually be United States customary units; for the UK Imperial units for some topics and metric units for others, and a mixture of units for others (see, for example, the Times Online style guide under "Metric"). The most appropriate units for articles may be deduced from the units most used in the sources for the article.

When some parts of the English-speaking world use a different unit from the one used in the article, place conversions afterwards in parentheses so that readers from all over the world can understand the measurement: for example, the Mississippi River is 2,320 miles (3,734 km) long; the Murray River is 2,375 kilometres (1,476 mi) long. (See § Unit conversions below.)

Comment

This version replaces 'prefer' with 'put.....first'. I have also removed the reference to screen sizes as I thought this was a rather trivial example. Years are important, the use of feet for the altitude of aircraft is important, but the size of a television or mobile phone screen is not so important. With UK articles a useful guide to what unit has priority may be to follow the majority of sources of information for the article. Michael Glass (talk) 13:53, 22 August 2009 (UTC)

"Put first" only makes sense when conversions are used, which won't be the case whenever something is measured with the same unit throughout the English-speaking world (hours, volts, picofarads, ohms, typographic points, dots per inch, yadda yadda yadda). "Use" always makes sense; the bullet about conversion makes clear than in "the Mississippi River is 2,320 miles (3,734 km) long", by "the one used in the article" we mean the mile. The addition of the point about source sounds fine. --___A. di M. 15:01, 22 August 2009 (UTC)

Based on your suggested changes:

I don't think it is necessary to change the title - we do already have a unit conversions section directly below this and we do already mention unit conversions in the text. For the same reason, I prefer prefer to put... first, but am not overly fussed about it.

I think leaving it at years is not really enough. It is, IMO, useful to include a measurement that is an exception to the SI-only rule in most of the world but not all of it - to show that this applies to more than just the most obvious cases. Inches for screen display size is a good example. If you can think of a better one, then I don't object to it being used instead.

Your sources bit appears to require source-based units in all circumstances on UK-based article. I don't think this is a good idea. Sources are often written for a non-native audience, and may reflect a units that would not ordinarily be used in the context of the topic concerned. Taken to a relative extreme, requiring source-based units could mean we have to compare a 2,000-foot (610 m) hill with a 650-metre (2,130 ft) hill because those are the units the two (completely different) sources use - something which I do not believe would look particularly professional.

Better to be a bit vaguer, IMO, to allow editors to determine the best units in the context of the article. If consensus is that converting the sourced units is the best idea, then we should convert the sourced units. We do still have WP:V and the section on sourcing below, so it's not like sourcing is not part of the equation. Pfainuk talk 15:09, 22 August 2009 (UTC)

My responses:

I agree that 'Put first' only makes sense when conversions are used. In other words, the policy would only apply where it was relevant. If the language is more precise, this is an advantage, not a defect.

if we mention unit conversions in the text then it is only sensible to mention unit conversions in the title, since this is what the text is about.

I think the use of the foot to measure altitude in aviation is a better example than the screen size of television sets.

My mention of sources The most appropriate units for articles may be deduced from the units most used in the sources for the article offers guidance to editors; it is not a straitjacket. Note that it says may be deduced not is to be deduced.

My mention of sources is consistent with present policy: 'If editors cannot agree on the sequence of units, put the source value first and the converted value second. If the choice of units is arbitrary, use SI units as the main unit, with converted units in parentheses.' [2]

Michael Glass (talk) 22:57, 22 August 2009 (UTC)

I see that you say "may be" as opposed to "is to be", but this remains the implication to my mind.

You argue that this is consistent with current policy. I don't totally agree, but I'll take this for the sake of argument. If it is consistent with current policy, why does it need reiterating specifically for British units? Better, surely, for the current guideline to remain and apply to all cases equally. This would avoid any implication that the rules for UK-related articles are any different from the rules for other articles. If the guideline is unchanged, why does this need to be added? Better, in my view, to avoid repeating ourselves too much.

My understanding was that the foot was universal (or nearly) when doing plane heights. Looking at Altitude, apparently not - so I will accept the foot in this case. Pfainuk talk 08:49, 23 August 2009 (UTC)

You are quite right that we don't want to single out the British for special treatment, in fact that is the opposite of what we are trying to do here I think. The problem at the moment is that it kinda does single out both the US and the UK as special cases, where the rewording really cleans out a lot of that, while stating that it is OK to use US Customary or Imperial when it is appropriate.

I kinda agree with your sentiment about being "vaguer", although I would say it is quietism (a word I nicely got out at Scrabble on Saturday, to the tune of 84 points), i.e. don't legislate about things that any sensible editor can happily work out for themselves; Glass is kinda right that it is guidance not law, but on the other hand wikilawyering will say MOSNUM says this MOSNUM says that. So best to be quiet when there is no guidance to offer. I think you could cut MOS in half, really, by just deleting overly specific rules, and you may have noticed on perhaps this rather mundane para that in every proposed rewording it gets shorter.

On the aviation point, you are right that it is measured in feet, or often expressed in "levels" (Level 250 would be 25,000 feet above sea level). A modern aircraft autopilot will hold the aircraft on that level with astonishing accuracy. I feel it may be better to stick with the aviation example as one that people know probably reasonably well, but switch to an example that has nautical miles (aviation charts measure it thus, and they are a non-SI unit accepted by SI and defined as 1,852m exactly). We could switch to ships, which tend to be at sea level (you may have noticed) but that is possibly somewhat favouring the British (of which I am one) because we are surrounded by the stuff and are kinda used to the sea; in other places, even in Europe, there are people who have never seen the sea – my Hungarian girlfriend had never seen the sea until she came to England – so the aircraft example is probably better to stick with I think.

And on a minor point, we should not call them "planes" in the actual text since that has too many other meanings, and "airplanes" vs "aeroplanes" is bound to stir up needless bickering. Aircraft is I think the preferred WP term, and certainly is the term used in the industry (which I worked in for about 9 years) and of course then covers helictoptes, microlights &c. i.e. suitably vague or quietist. I am sure you just used that in the discussion and would not use it in the wording itself, I mention it only for completeness.

Best wishes SimonTrew (talk) 22:42, 26 August 2009 (UTC)

Another refinement of the suggested wording

Taking into account the criticisms above, here is another refinement of the suggested wording and the title:

A more comprehensive title is provided.
The wording has been tweaked to make it clearer in one or two places.
The words 'in the present' have been removed from the direction about US articles because I couldn't see that they said anything useful.
The dot points have been rearranged into two lists. This enabled me to remove one redundant sentence and to tidy the rather ragged layout in the present policy.

Which units to use and how to present them

Except in the cases mentioned below, put the units first that are in the most widespread use worldwide. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as years for long periods of time or the use of feet in describing the altitude of aircraft.

With topics strongly associated with places, times or people, put the units most appropriate to them first. In US articles, this will usually be United States customary units; for the UK Imperial units for some topics and metric units for others, and a mixture of units for others (see, for example, the Times Online style guide under "Metric").
If editors cannot agree on the sequence of units, put the source value first and the converted value second. If the choice of units is arbitrary, use SI units as the main unit, with converted units in parentheses.
Generally, use units consistently (e.g., write a 10-kilogram (22 lb) bag of potatoes and a 5 kg (11 lb) bag of carrots, not a 10-kilogram (22 lb) bag of potatoes and a 11-pound (5 kg) bag of carrots).
Nominal and defined values should be given in the original units first, even if this makes the article inconsistent: for example, When the Republic of Ireland adopted the metric system, the road speed limit in built-up areas was changed from 30 miles per hour (48 km/h) to 50 kilometres per hour (31 mph). (The focus is on the change of units, not on the 3.6% increase.)
Avoid ambiguous unit names (e.g., write imperial gallon or US gallon rather than gallon). Only in the rarest of instances should ambiguous units be used, such as in direct quotations, to preserve the accuracy of the quotation.
When some parts of the English-speaking world use a different unit from the one used in the article, place conversions afterwards in parentheses so that readers from all over the world can understand the measurement: for example, the Mississippi River is 2,320 miles (3,734 km) long; the Murray River is 2,375 kilometres (1,476 mi) long. (See § Unit conversions below.)

In scientific articles, use the units employed in the current scientific literature on that topic. This will usually be SI, but not always; for example, natural units are often used in relativistic and quantum physics, and Hubble's constant should be quoted in its most common unit of (km/s)/Mpc rather than its SI unit of s⁻¹.
Some disciplines use units not approved by the BIPM, or write them differently from BIPM-prescribed format. When a clear majority of the sources relevant to those disciplines use such units, articles should follow this (e.g., using cc in automotive articles and not cm³). Such use of non-standard units are always linked on first use.
Use familiar units rather than obscure units—do not write over the heads of the readership (e.g., a general interest topic such as black holes would best be served by having mass expressed in solar masses, but it might be appropriate to use Planck units in an article on the mathematics of black hole evaporation).

Michael Glass (talk) 13:05, 23 August 2009 (UTC)

I'll accept this. I think it would be better if the sourcing point was second last in the sub-list, so as to list the more general rules first, and then to proceed to what to do if unsure. But this is a fairly minor point - not worth too much hassle. Pfainuk talk 19:57, 23 August 2009 (UTC)

Thanks for the input. I'll go with the word order above and see what the result is. Michael Glass (talk) 22:51, 23 August 2009 (UTC)

I've been away from a few days and while I am glad my proposed changes have been largely adopted, I am not so happy with the rest of it. I would have preferred that the whole lot was discussed rather than make the change and "see what the result is". Our responsibility here is not to go changing the MOS, which affects in theory every page on the English WP, without very good reason. While it may seem to others that the discussion then gets rather boring, those others can, in my opinion, go and do something else, since the MOS must be relatively stable, authorititive, not unduly prescriptive, clear, and concise. If it changes on the whim of one editor, it is about as much use as a snake in an arse-kicking competition. In my opinion, it is simply not acceptable to change MOSNUM and "see what the result is". The result from at least one editor is annoyance that after careful discussion and several iterations from several editors, from a discussion that you yourself started, you just come in and change it to your preferred wording anyway (taking along the suggestions that were made in the mean time).

I have to use this stuff. I edit mostly technical articles about physics and chemistry and all sorts of things; putting Coca-Cola formula from bizarre metric measures into amazingly accurate drachms (i.e. because, we think, from overly-specific conversions to metric in the source), caramel color or pencil, all of which need some conversions and definitions. The aim I have in all of those is, as thoreaux said, simplify, simplify.

I am annoyed with this "except in the cases mentioned below"; it is simply ugly (not your fault it has been in every proposal including mine). I would like to find a better alternative, but I think maybe just to cut it entirely; surely the exceptions stand for themselves? Not sure here as we can't have conflicting advice, but it is very ugly. Since the para mentions there are exceptions, links to them and gives an example, I am suggesting that it is quite obvious that the following indented bullets are also exceptions? But I am not wholly confident with that view; what do others think?

So let me suggest my alternative to your ~~fait accompli~~proposal. I have not struck caps when they need to be added or removed, please take that as read to save my unnecessarily cluttering it:

~~Except in the cases mentioned below,~~ put the units first that are in the most widespread use worldwide. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as ~~years~~ for ~~long~~ periods of time longer than an hour or ~~the use of feet in describing the altitude of aircraft~~nautical miles to describe the distances that aircraft travel.

With topics strongly associated with places, times or people, put the units most appropriate to them first. In US articles, this will usually be United States customary units; for the UK Imperial units for some topics and metric units for others, and a mixture of units for others (see, for example, the Times Online style guide under "Metric").
~~If editors cannot agree on the sequence of units,~~ put the source value first and the converted value second. If ~~the choice of units is arbitrary~~there is no clear choice which value to put first, use SI units as the main unit, with converted units in parentheses.
~~Generally,~~ use units consistently ~~(e.g.,~~: write a 10-kilogram (22 lb) bag of potatoes and a 5 kg (11 lb) bag of carrots, not a 10-kilogram (22 lb) bag of potatoes and a 11-pound (5 kg) bag of carrots).
Give Nominal and defined values ~~should be given~~ in the ~~original~~source's units first, even if this makes the article inconsistent: for example, When the Republic of Ireland adopted the metric system, the road speed limit in built-up areas was changed from 30 miles per hour (48 km/h) to 50 kilometres per hour (31 mph). (The focus is on the change of units, not on the 3.6% increase.).
Avoid ambiguous unit names (e.g., write imperial gallon or US gallon ~~rather than~~instead of gallon). ~~Only in the rarest of instances should ambiguous units be used, such as in direct quotations, to preserve the accuracy of the quotation.~~But do not change units in quotations; place conversions in square brackets directly afterwards within the quotation if you feel it is useful to do so.
When some parts of the English-speaking world use a different unit from the one used in the article, place conversions afterwards in parentheses so that readers ~~from all over the world~~everywhere can understand ~~the measurement~~it: for example, the Mississippi River is 2,320 miles (3,734 km) long; the Murray River is 2,375 kilometres (1,476 mi) long. (See § Unit conversions below.)

In scientific articles, use the units ~~employed in~~used by the current scientific literature on that topic. This will usually be SI, but not always; for example, natural units are often used in relativistic and quantum physics, and Hubble's constant should be quoted in its most common unit of (km/s)/Mpc rather than its SI unit of s⁻¹.
Some disciplines use units not approved by the BIPM, or write them differently from the way that BIPM says they should be.BIPM-prescribed format. When a clear majority of the sources relevant to those disciplines use such units, articles should follow this (e.g., using cc in automotive articles and not cm³). Such use of non-standard units are always linked on first use.
~~Use~~Prefer familiar units ~~rather than~~over obscure onesunits- do not write over the heads of the readership (e.g., a general interest topic such as black holes would best be served by having mass expressed in solar masses, but it might be appropriate to use Planck units in an article on the mathematics of black hole evaporation).

As far as the scientific bits at the end go, your version seems to add more vagueness than was there in the first place. What is a clear majority? Why the irrelevant examples for Hubble's constant and so forth and references to a particular pair of topics (relativistic and quantum physics)? It seems that you are just muddying the waters here rather than clearing them. That being said, "use familiar units" is worthwhile, but again we are back to physics with black hole evaporation, and Planck units (whatever they are, I have never heard of them in 20 years of doing sub-atomic physical science, perhaps I am just thick), that's just too specific and confusing. Cut that crap out, the exempli gratia in the opening para explains the principle and we don't have to lord it over the readership because we all know about black hole evaporation. You are hoist with your own petard there because you say use familiar units and then express in completely unfamiliar units. If your point was deliberately to confuse then you succeeded, but it is not the goal of MOSNUM to confuse, I think, but to clarify.

Best wishes. SimonTrew (talk) 23:28, 26 August 2009 (UTC)

Simon, most of your proposed changes (above) are OK with me. Some of the wording (e.g. on scientific terms) you object to is unchanged from the previous version. They were just not noticed before. I agree with the idea of simplifying rather than complicating things. I would be happy to take up your revisions almost in full. I would insert is the words In general at the beginning and make a few other minor changes as shown below:

In general, put the units first that are in the most widespread use in the world. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as for periods of time longer than an hour or nautical miles to describe the distances that aircraft travel.

With topics strongly associated with places, times or people, put the units most appropriate to them first. In US articles, this will usually be United States customary units; for the UK Imperial units for some topics and metric units for others, and a mixture of units for others (see, for example, the Times Online style guide under "Metric").
Put the source value first and the converted value second. If there is no clear choice which value to put first, use SI units as the main unit, with converted units in parentheses.
Use units consistently: write a 10-kilogram (22 lb) bag of potatoes and a 5 kg (11 lb) bag of carrots, not a 10-kilogram (22 lb) bag of potatoes and a 11-pound (5 kg) bag of carrots.
Give nominal and defined values in the source's units first, even if this makes the article inconsistent.
Avoid ambiguous unit names (e.g., write imperial gallon or US gallon instead of gallon) but do not change units in quotations; place conversions in square brackets directly afterwards within the quotation if it is useful to do so.
When some parts of the English-speaking world use different units from the one used in the article, place conversions afterwards in parentheses so that readers everywhere can understand it: for example, the Mississippi River is 2,320 miles (3,734 km) long; the Murray River is 2,375 kilometres (1,476 mi) long. (See § Unit conversions below.)

In scientific articles, use the units used by the current scientific literature on that topic.
Some disciplines use units not approved by the BIPM, or write them differently from the way that BIPM says they should be.
Prefer familiar units over obscure ones.

Michael Glass (talk) 05:01, 27 August 2009 (UTC)

I don't like this because it changes words couched in appropriate ambiguity into a self-contradictory set of rules that appear to have no exceptions beyond WP:IAR.

There will be situations where it is impossible to follow the rules given here as absolute:

Use the most appropriate units in a given geographical or historical context
Use the source value first
Use units consistently
Give nominal and defined units first, even where this is inconsistent

In a previous message I named an example - where we compare a hypothetical 2,000 feet (610 m) hill with a 650 metres (2,130 ft) hill. If those are the measures given in the sources, there's no way we can meet the guidelines we are given.

My proposal would be:

Generally speaking, you should determine what unit is the most appropriate primary unit in the context of the measurement you are using, and then proceed to convert this into all units that are in common use in the English-speaking world in that context, so that all can understand it.
To determine which units should be primary, the following guidelines should be considered:

In general, the primary unit is that which is in the most widespread use in the world. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as for periods of time longer than an hour or nautical miles to describe the distances that aircraft travel.

Where a topic is strongly associated with a given place, time or person, use the units most appropriate to them first. In US articles, this will usually be United States customary units; for the UK Imperial units for some topics and metric units for others, and a mixture of units for others (see, for example, the Times Online style guide under "Metric").

Give nominal and defined values in the units in which they are defined first, even if this makes the article inconsistent. Otherwise, use units consistently: write a 10-kilogram (22 lb) bag of potatoes and a 5 kg (11 lb) bag of carrots, not a 10-kilogram (22 lb) bag of potatoes and a 11-pound (5 kg) bag of carrots.

In scientific articles, use the units used by the current scientific literature on that topic.

Where a given discipline uses units not approved by the BIPM, or writes them differently from the way that BIPM says they should be written, use the units as commonly used in that discipline.

Prefer familiar units over obscure ones.

If there is no clear choice as to which unit is primary, put the source value first and the converted value second. If the choice is arbitrary, use the SI unit.
You should avoid using ambiguous unit names (e.g., write imperial gallon or US gallon instead of gallon), but do not change units in quotations. Place conversions in square brackets directly afterwards within the quotation if it is useful to do so.

Sourcing is my last bullet, because so far as I can see most of this is listing cases that overrule the sources-first rule. Even if a given source puts energies in foot-poundals, we still probably want it put in Joules - unless the second bullet (to use units appropriate to a time period or person) applies. Pfainuk talk 17:42, 27 August 2009 (UTC)

Pfainuk, you make some telling comments about the proposed wording. You are quite right in noting the inconsistency in trying to balance competing requirements of both consistency of presentation and accuracy reflecting the sources. These differences became far more obvious in the revised version because the language was so much clearer, as clear wording makes any defect in logic far more obvious.

However, there are problems with your proposed solution.

"Put first" is less POV than "primary unit". "Primary unit" is also ambiguous, because it could also mean the unit used in the source of information. Therefore I believe "put first" is better wording.
Because different English-speaking countries use different units of measure, talk of "units that are in common use in the English-speaking world" is problematic.
Putting the the point about source value last is not helpful. One good way of establishing usage is to find out what the sources use. Otherwise the choice of units would be essentially arbitrary.

The proposal below makes the language less prescriptive. I have also made other changes, including omitting the point about nominal and defined values. Without the example the point is quite obscure; with the example it belabours the point. I have also removed the bolding and added the introduction from the policy as it is at the moment (in italics).

Units of measurement

Shortcuts
WP:UNIT
WP:UNITS
WP:MEASUREMENT

The use of units of measurement is based on the following principles:

Avoid ambiguity: Aim to write so you cannot be misunderstood.
Familiarity: The less readers have to look up definitions, the easier it is to be understood.
International scope: Wikipedia is not country-specific; apart from some regional or historical topics, use the units in most widespread use worldwide for the type of measurement in question.''

If there is trouble balancing these bullets, consult other editors through the talk page and try to reach consensus.

Which units to use and how to present them

Apart from the exceptions listed below, put the units first that are in the most widespread use in the world. Usually, they are the units of the International System of Units (SI) and non-SI units accepted for use with SI; but there are various exceptions for some measurements, such as for periods of time longer than an hour or nautical miles to describe the distances that aircraft travel.

With topics strongly associated with places, times or people, put the units first that are most appropriate to them. In US articles, this will usually be United States customary units; for the UK, Imperial units for some topics, metric units for others, and a mixture of units for others (see, for example, the Times Online style guide under "Metric").
In general, put the source values first and the converted values second. If the choice is not clear, put the SI units first, with the other units in parentheses.
Avoid inconsistent usage: write a 600 metres (2000 ft) hill with a 650 metres (2,130 ft) hill., not a 2,000 feet (610 m) hill with a 650 metres (2,130 ft) hill. If the source value is not put first, the source value should be recorded in a note to the text.
Avoid ambiguous unit names (e.g., write imperial gallon or US gallon instead of gallon) but do not change units in quotations; place conversions in square brackets directly afterwards within the quotation if it is useful to do so.
When some parts of the English-speaking world use different units from the one used in an article, place conversions afterwards in parentheses so that readers everywhere can understand it: so, the Mississippi River is 2,320 miles (3,734 km) long in an article about the US; and the Murray River is 2,375 kilometres (1,476 mi) long in an article about Australia. However, in a general article about the length of rivers, put SI units first. (See § Unit conversions below.)

In scientific articles, use the units used by the current scientific literature on that topic.
Some disciplines use units not approved by the BIPM, or write them differently from the way that BIPM says they should be.
Prefer familiar units over obscure ones.

Comment from Pfainuk

I have edited the above so that these are level 3 and level 4 headers, so that it's more obviously the same conversation. This is not intended to affect the proposal, but rather for convenience on talk. The first time I looked at this, I didn't notice the proposal because it was under a level 2 heading.

I notice that this proposal fails to take account of A di M's point that we're talking about which units to put first before we mention that we're supposed to be putting more than one unit. This was the intention of the opening paragraph in my version: start with the basic principle that we choose a unit to put first and then convert that unit into all the relevant alternatives, and then go into detail as to decide which unit to put first.

My other thought structure-wise was that all of the bullet points except the one on ambiguous unit names deal with which units should go first in the text. As such, it makes a bit more sense just to treat it as one list rather than two. The three bullets at the bottom would go below the point about consistency.

On the point on consistency, I would suggest that it is not necessary for a source value to be placed in a footnote. It may be appropriate to ensure that it is clear when the sourced value is not placed first - but use footnotes throughout and we'll potentially end up with articles littered with useless footnotes - thereby devaluing the useful ones.

On defined units - for the most part this should be common sense, and it is mostly covered by the people-times-and-places rule. I think it's best off going in, but I would hope that editors don't need to be told.

I would also point out that your proposal advises us to put the most widespread units in use in the world first, except that in general they should use the source unit, regardless of system. This is self-contradictory in a way that the current text is not. It may be a good idea to ditch the sources point entirely, and instead integrate it into the first paragraph or to place it, with the note on ambiguous units, in a separate paragraph at the end:

If it is unclear as to which units are most appropriate, put the units preferred by the source for the unit first. If the choice is arbitrary, put SI units first.

Placing this in a separate paragraph perhaps makes it clearer that this is an extension to, and not an exception to, the basic international-units-first rule (whereas the other points are exceptions to it). Sources-first is most appropriate when the other rules are not applicable since it inherently has the potential to introduce intra- and inter-article inconsistency when dealing with similar measurements in similar contexts. Pfainuk talk 10:23, 29 August 2009 (UTC)

Glass's point about source units is valid, in that if an editor doesn't know which unit is the appropriate one for the height of a hill in Great Britain, they can figure out by looking at what sources do. OTOH his wording is somewhat unclear about that. I would use something like: With topics ... (... under "Metric"). When in doubt, use the unit which most sources use for that type of measurement in that geographical context: it will be likely to be the most appropriate one. --___A. di M. 14:40, 29 August 2009 (UTC)

Besides being entirely unnecessary instruction creep, it will almost always be possible to find sources that use whichever units you prefer, and there are some editors who may use this kind of wording in MOS to justify changing the priority (order) of units within articles to suit their agenda. _wjemather ^bigissue 20:40, 29 August 2009 (UTC)

We need to take account of the fact that different English-speaking people use different units; that is why we need to supply both traditional and metric measures in many contexts. The question is which unit should come first. Mostly it is clear-cut, because most of the world uses metric units and there is a clear rule about US-based articles. With UK based articles it may not be clear and this is where it is useful to go by the sources.

Now for specific comments:

If source-shopping is the game, someone will soon come along and put that right by quoting a better source. A more common problem is that sources are not provided.
While in theory the guidelines may sound contradictory, in practice the best sources will agree with the other guidelines and there will be no contradiction.

If people are satisfied with the present wording it might be better to leave it be. I will try once more to see if I can improve on the wording, but if that doesn't work I'll let the matter rest. Michael Glass (talk) 22:51, 29 August 2009 (UTC)

After giving the matter some further consideration I feel it is best to proceed step by step with any changes of wording. The first change I would make is to change this:

Generally, use units consistently (e.g., write a 10-kilogram (22 lb) bag of potatoes and a 5 kg (11 lb) bag of carrots, not a 10-kilogram (22 lb) bag of potatoes and a 11-pound (5 kg) bag of carrots).

to this:

Avoid inconsistent usage: write a 600 metres (2000 ft) hill with a 650 metres (2,130 ft) hill., not a 2,000 feet (610 m) hill with a 650 metres (2,130 ft) hill.}}. If the source value is not put first, the source value should be recorded in a note to the text.

The reasons are as follows:

The example is more realistic. We are not telling people how to write shopping lists for vegetables.
It shows editors what to do if the sources of information are inconsistent.
It is important to note what the original says.

What do others think? Michael Glass (talk) 02:42, 31 August 2009 (UTC)

Again, I don't agree that we should put in a footnote every time the unit we use isn't the same as the unit the cited source uses. Such a rule is likely to mean that many articles, where the units used are those that are most appropriate for the context and not necessarily the units preferred by the sources, are likely to be filled with unnecessary footnotes. Putting in lots of useless footnotes detracts from the useful footnotes in the same way that having lots of useless links detracts from the useful links.

I would note on your first message, you say that in practice the best sources will agree with the other guidelines and there will be no contradiction. If we're assuming that the sources will follow the other guidelines, then there's no need for us to specifically mention sources - this is instruction creep. Either the point on sourcing is unnecessary, or it contradicts the other points. And both are a problem. Pfainuk talk 14:23, 31 August 2009 (UTC)

As for the first point, the measurement should cite its source anyway per WP:V, and it'd be no great deal if the citation instead of reading John Doe (1987). "Bla bla bla", in Blaology. it reads John Doe (1987). "Bla bla bla", in Blaology. (Measurements reported in inches in the original.); OTOH I don't think it is always necessary; but it definitely doesn't hurt. --___A. di M. 15:19, 31 August 2009 (UTC)

My "solution" for identifying source units is to include a comment like . (Maybe identifying the source in the comment by its ref name would be even better.) At least that way, a future editor will know what was intended, without having to refer back to the original document cited. If you've got Imperial and SI sources contributing to an article, I'd recommend sticking to one set of units (parenthesizing alternatives as necessary), unless the choice of units is critical to the understanding of the text. On the other hand, if it's necessary to alert the reader of the source unit, a footnote would be a good way to do it, but I don't think that it's important enough to mention for most situations. TheFeds 20:41, 31 August 2009 (UTC)

I notice that all the comments have been on footnoting changes between the text and the source. I will let that matter rest for the moment and note something else. Nothing has been said about replacing the present shopping list with a more appropriate example.

From this:

Generally, use units consistently (e.g., write a 10-kilogram (22 lb) bag of potatoes and a 5 kg (11 lb) bag of carrots, not a 10-kilogram (22 lb) bag of potatoes and a 11-pound (5 kg) bag of carrots).

to this:

Avoid inconsistent usage: (e.g., write a 600 metres (2000 ft) hill with a 650 metres (2,130 ft) hill., not a 2,000 feet (610 m) hill with a 650 metres (2,130 ft) hill.).

Any problems? If there are none, I will make this change to the text. Michael Glass (talk) 12:17, 1 September 2009 (UTC)

I have no objection to this change, other than that I think |adj=on should be added to the convert template, for purposes of grammar (a 600-metre hill). Pfainuk talk 20:14, 1 September 2009 (UTC)

Done, though not with the conversion templates. Michael Glass (talk) 11:32, 3 September 2009 (UTC)

Another proposal

At present the policy reads:

When different parts of the English-speaking world use different units for the same measurement, the "primary" unit should be followed by a conversion in parentheses. This will enable readers from all over the world to understand the measurement: for example, the Mississippi River is 2,320 miles (3,734 km) long; the Murray River is 2,375 kilometres (1,476 mi) long. (See § Unit conversions below.)

This could be more clearly expressed to explain that the different measures would be used in different articles. Here is my proposal:

When some parts of the English-speaking world use different units from the one used in an article, place conversions afterwards in parentheses so that readers everywhere can understand it: so, the Mississippi River is 2,320 miles (3,734 km) long in an article about the US; and the Murray River is 2,375 kilometres (1,476 mi) long in an article about Australia. However, in a general article about the length of rivers, put SI units first. (See § Unit conversions below.)

Any comments, suggestions or objections? Michael Glass (talk) 11:32, 3 September 2009 (UTC)

In an earlier discussion, it seemed unclear how many ordinary readers were in the U.S., how many in purely-metric countries (Anglophone and non-Anglophone) and how many in hybrid countries like Britain or Canada where both metric and non-metric units are sometimes used (regardless of official or technical usages, but rather how ordinary people think of quantities). Only if it's clear that a significant majority of Wikipedia readers think of the units in question in purely metric terms would I think that "put metric units first" in articles with no clear association with any one country would be justified. —— Shakescene (talk) 11:49, 3 September 2009 (UTC)

Entirely unnecessary as clarification follows immediately in the next paragraphs. _wjemather ^bigissue 16:54, 3 September 2009 (UTC)

Agree. This point as it stands does not significantly deal with which unit comes first - it merely demonstrates conversion. Better to leave the explanation about which unit should come first to the bullet points directly after. We don't have to make every point everywhere. Pfainuk talk 17:07, 3 September 2009 (UTC)

It is impossible to work out exactly how many readers come from which countries, let alone what units they might prefer. This wording might be better:

When some parts of the English-speaking world use different units from the one used in an article, place conversions afterwards in parentheses so that readers everywhere can understand it: so, the Mississippi River is 2,320 miles (3,734 km) long in an article about the US, and the Murray River is 2,375 kilometres (1,476 mi) long in an article about Australia. (See § Unit conversions below.)

It makes it clear that we are talking about different articles.
It is shorter than the present wording.
It avoids the use of "primary" unit, which could be mistaken for the source unit.

Any problems or objections to this change? Michael Glass (talk) 21:04, 3 September 2009 (UTC)

Now that I'm paying closer attention, I'm really confused. "Some parts of the English-speaking world" will always use different units from the one used in an article. One of the generally-accepted points (though not a formal consensus or conclusion) in the earlier discussion (which might have been about units or about WP:ENGVAR) is that one can't make assumptions about who wants to read about Canberra, Stephen Leacock, Manhattan or the Tower of London. There might be more interest among outsiders who know nearly nothing but want to learn more, than among compatriots of the article's subject.

So, for example, I gave metric conversions for feet and yards in the articles about Yankee Stadium (1923) and Yankee Stadium almost everywhere that didn't have them. (I think there was one table where it just wasn't practical.) —— Shakescene (talk) 21:35, 3 September 2009 (UTC)

Not always. Units of time are generally the same all over the world, for example - so seconds do not generally need to be converted into any other unit.

My comments on this proposal are:

When some parts of the English-speaking world use different units from the one used in an article, place conversions afterwards in parentheses so that readers everywhere can understand it:

We haven't got as far as putting a unit in the article yet because we haven't decided which one we need. Further, there may be several units used in a single article - which one are you referring to? The current wording is a whole lot clearer.

You argue that "primary unit" can be misconstrued as meaning the source unit. I don't doubt it can be: in the past I've seen an editor who managed to construe "any material challenged or likely to be challenged must be attributed to a... source" from WP:V to mean that the person challenging the material could not remove it without providing a source to back up the removal. But I think such an argument would be difficult to sustain, particularly given that we proceed to give a long list of guidelines as to how to determine which units are "primary". The term "primary" is not exactly badly defined here.

The remainder of your proposal goes on - essentially - to repeat the guidance directly below it. Why do we need to do that? Where is the benefit? I don't see any benefit, and I think that it's better to keep these trimmed and simple. We're not trying to demonstrate which units to use here - we're only trying to demonstrate how to present units. So, how about we replace both with the River Nile is 6,650 kilometres (4,130 mi) long?

Finally, I note you argue that your version is shorter than the present wording. It's actually two words longer. Pfainuk talk 16:46, 4 September 2009 (UTC)

Well, it takes up less space, that's for sure. The problem with giving only one example, as you propose, is that it could suggest that only articles with metric measurements first should have conversions. That's why the two examples are helpful here.

As we can't agree on what changes would be acceptable it is better to leave it at this time. Michael Glass (talk) 01:03, 5 September 2009 (UTC)

Choice of units

The present guidelines state:

If editors cannot agree on the sequence of units, put the source value first and the converted value second. If the choice of units is arbitrary, use SI units as the main unit, with converted units in parentheses.

I think it would be preferable to express this as:

In general, put the source value first and the converted value second. If the choice of units is arbitrary, use SI units as the main unit, with converted units in parentheses.

It would seem to me to be a good general rule to follow the sources. However, the guidelines should provide guidance rather than a straitjacket. What do others think of this? Michael Glass (talk) 01:17, 5 September 2009 (UTC)

Changing "if editors cannot agree on the sequence of units, put the source value first" to "in general, put the source value first" contradicts the rest of the section. Putting source values first leads to choppy articles unless the article only uses one source. Putting source values first should be a last resort, except for direct quotes. --Jc3s5h (talk) 02:32, 5 September 2009 (UTC)

I agree with Jc3s5h. As I argued before regarding this proposal, the rule becomes effectively:

Put the units first that are in the most widespread use in the world, except that in general you should put the source value first and the converted value second.

The general rule is simultaneously to put the most widely used units first and to put the source units first. If they're the same, then the sources point is redundant. If they're different, then the two points directly contradict one another. Both are a problem.

Better, in my view, only to fall back on source values in cases where none of the other rules apply - where it is not clear which is the most widespread unit and there is no guidance from the other rules. Pfainuk talk 07:52, 5 September 2009 (UTC)

I note the complaints that the proposed rule contradicts the rule to be consistent. No it doesn't, unless the sources are inconsistent, and if they are inconsistent the guideline explains what to do if the choice of units is arbitrary. This problem is unlikely to arise in most of the world, because most of the world uses the metric system for most ordinary measures, In the USA, the US customary measures are used, so it is not so likely to arise there. That only leaves Britain where there is likely to be any real confusion, and this is where the rule would be most useful. If most authoritative sources are Imperial, that's the way to go. However, if they are closer to the usage of the Times Guide then that would be OK, too. If metric, then go metric. Or, if the authoritative sources were really at sixes and sevens, then metric would also be the way to go. The rule of following the sources is helpful for British articles because it's more objective than simply relying on the editors to make this decision without reference to the sources of information. Michael Glass (talk) 15:19, 5 September 2009 (UTC)

A few points here:

The proposed rule does contradict the rule that units should be consistent. If the sources are not consistent, then we cannot have source-based units and consistent units. That is a contradiction. That the sources are inconsistent does not imply that the choice of units is arbitrary, and there may be a unit appropriate in the context of the article that is not SI.

As it happens, though, that's not what I argued. I argued that your proposed wording simultaneously tells people that, in general, they should use the unit used the source and the unit in most widespread use internationally, and that this is contradictory.

You argue there is likely to be no difference between these two rules. On the contrary, it's reasonably easy to think of examples where this could be an problem.

If a source for the distance between the Earth and the Moon is US- or UK- based, it may well give that distance in miles. In that case, there is a contradiction between the sources rule and the widespread-usage rule. A source describing the weight of a hare might well express a value in pounds if it was written in the US or UK. In that case, there is a contradiction between the sources rule and the widespread-usage rule. The muzzle velocity of an AK-47 is likely to be expressed in feet per second in an American source. In that case, there is a contradiction between the sources rule and the widespread-usage rule. In any case where this new rule makes a difference, it contradicts the widespread-usage rule.

Even if there is no difference (as you argue and I dispute) then there is no good reason to be adding a new rule, and instruction creep is a good reason not to. Either this rule is contradictory or it is instruction creep - and in either case it's a bad idea.

Finally, I would note that nowhere does your proposal state that the guideline should only apply to UK-related articles. It does not - it applies to all articles. If it did apply only to UK-related articles I would oppose it. UK-related articles should be treated as other articles, bearing in mind that the guideline says that appropriate units are to be used as primary when dealing with any topic that is strongly associated with a certain place, time or person. Pfainuk talk 16:37, 5 September 2009 (UTC)

Neither the present wording nor the proposed wording state that the guidelines should only apply to UK-related articles. If the widespread usage rule applies as you have stated, then all articles should be metric first, as is the case with your examples. If this rule is to be tempered by an appeal to time, place or person, then how is this to be tested but by looking at the sources? The kind of thing that i would question is when an article is in Imperial units when most or all of the sources are metric. Michael Glass (talk) 23:40, 5 September 2009 (UTC)

You say: If the widespread usage rule applies as you have stated, then all articles should be metric first, as is the case with your examples.

Unfortunately for your position, that's not what your suggested rule says. On the contrary, your suggested rule says both that imperial units and that metric units should be used as primary units in these cases. The widespread-usage rule says one thing and the use-sources rule says the other, and both apply. Which leaves the editor none the wiser.

You say that you would question a case where an article is imperial-first using sources that are metric-first. I have no doubt that this is not uncommon among US-related articles, where the imperial-first rule is well-established. That it sometimes happens on British-related articles, equally, is not entirely surprising.

Not all sources follow Wikipedia's MOS, and it seems particularly fair to assume that not all sources - particularly those used in UK-related articles - will follow it when we make an exception based on person, time or place. Sources are not the be-all and end-all. The British are well aware that imperial units are not used by most outside the UK and will commonly use metric units only in texts intended for a non-British audience (such as tourist guides) - including in circumstances where imperial units are actually more appropriate in a British context. But on Wikipedia we're not specifically writing for a non-British audience, and the source may well not use the most appropriate unit in the circumstances. Better, IMO to use the units that consensus deems to be appropriate, falling back on the sources if necessary as per the current wording. Pfainuk talk 08:55, 6 September 2009 (UTC)

The problems I see with your position are two assumptions that underlie them.

That when the British use metric units it is for the sake of tourists and people outside the UK.
That the consensus of editors is a superior guide to usage than the sources of information.

It's more accurate to say that usage in the UK is contested, with some actively campaigning for the full adoption of metric units [3] and some others fighting tooth and nail for the retention of the older units [4]. Somewhere in the middle is the Times Style Guide [5] which acknowledges "The Times should keep abreast of the trend in the UK to move gradually towards all-metric use, but given the wide age range and geographical distribution of our readers, some continuing use of imperial measurements is necessary."

Dismissing the use of metric units as simply an attempt to cater for tourists is not in accordance with the reality of the situation.

I think you will find that there is a similarly gradual movement towards metric measures on Wikipedia and most editors will be untroubled by this development. However, a few will resist this even when the sources of information are predominantly or even exclusively metric. In this situation a suggestion to follow the sources will offer some commonsense guidance when the question arises.

Michael Glass (talk) 23:58, 6 September 2009 (UTC)

Your first point is not my point. It may not have been as clear as one would have liked, but my point was not that British people use metric units for the sake of foreign tourists and no-one else. If it had been I would be arguing for full imperial units in British contexts - and I made it very clear in the last discussion that this is not something I think would be a good idea. My point was that, in literature for foreign tourists, the British will tend to use metric units in circumstances where imperial units would be more appropriate in a British context. This might, for example, include the speed someone is driving. Most Britons don't have a clear idea of what speeds in kilometres per hour mean in practical terms. The only circumstances under which one might think to use them would be when one's words are intended for a foreign audience. Using it in other contexts would appear to be the RL equivalent of being POINTy. But I'm sure you can find sources giving speeds in a UK context in metric measures.

Living in the UK, as I do, I am well aware that metric units are in use in some circumstances. But imperial measures are also used in some circumstances, and given that sources are not perfect, generally speaking - given that they do not always reflect Wikipedia's style guidelines - they not always the best guide to the most appropriate usage. Better to rely on a consensus of editors: we trust them to be able to choose which wording is most appropriate, why can't we trust them to choose which units are appropriate - falling back on source units where necessary as per the current wording?

All that said, I notice with interest that you fail to address the core issue with this wording. This wording clearly is not intended apply only in British contexts (indeed it's not clear that it's intended to apply in British contexts at all: they're an exception to one general rule, why not the other?) - it is intended to apply "in general". You have not adequately addressed the fact that, if we were to accept this rule, we would, in general, call upon users both to use most widely used units in a given context as primary and to use sourced units as primary - and that this is contradictory when the source uses a unit other than that which is most widely used in a given context. Pfainuk talk 16:53, 7 September 2009 (UTC)

I'm not sure we should be worrying excessively about what units the British would use. There are, after all, only 60 million of them, they are part of the European Union now, and they are officially under the metric system. There are a few hold-outs, of course, but they'll eventually die out. The only real holdout to metric is the United States, and Americans are at least theoretically using the metric system as well. We cater to Americans because there are 300 million of them, but I'm not sure how long our forbearance will last, at least not mine. The rest of the English-speaking world understands metric, so the default units for articles should be metric, and exceptions only made for the U.S.RockyMtnGuy (talk) 00:12, 8 September 2009 (UTC)

I don't agree. There are contexts in which imperial units are used overwhelmingly in the UK, and in these contexts it would be inappropriate to use metric units first. Most British people understand metric in theory, but do not understand it in practice - they know that a vehicle travelling at 70km/h will travel 70,000m in every hour, but they won't have a clear idea of how far 70km is in practical terms or whether 70km/h is fast, slow, or the right speed to be driving through a town, or down a motorway. Roads are signposted in miles and yards, speed limits are posted in miles per hour, and car speedometers give speeds in miles per hour (there is generally a km/h dial but it's generally far too small to be of any practical use).

I would note that the extent to which the UK adopts metric units is a political issue here, and systematically prescribing metric units in UK contexts regardless of whether they are the most appropriate unit in the context or not would imply that Wikipedia holds a POV on the matter.

That said, the current wording doesn't actually treat the US or UK as exceptions - rather, any context that is strongly tied to a person, place or time is an exception to the most-widely-used-internationally rule. It may be appropriate to use imperial first for all measures on an article about the Blitz, or mesures usuelles on an article about some aspect of the ancien régime of France (though these might equally come under the "obscure units" rule). Pfainuk talk 06:21, 8 September 2009 (UTC)

I really can't see how you can say that the current wording doesn't actually treat the US or UK as exceptions. Look at the wording:

Except in the cases mentioned below... ...In US articles... ...for the UK ...

Of course the wording treats both the UK and the US as exceptions to the general policy.

The policy for the UK gives the following latitude:

Imperial units for some topics and

metric units for others, and

a mixture of units for others (see, for example, the Times Online style guide under "Metric").

How is is possible to deny that this is so? Michael Glass (talk) 12:53, 8 September 2009 (UTC)

The exception is about all "topics strongly associated with places, times or people"; the UK and the US are just examples of that. (They are the most relevant ones because the units most appropriate to, e.g., present-day continental Europe topics will almost always be the same units most appropriate for non-regional topics. But they are by no means the only cases to which the exception applies.) --___A. di M. 13:04, 8 September 2009 (UTC)

Exceptions are exceptions are exceptions. There are general exceptions of course, but there are also specific exceptions for both the US and the UK. Read the policy and you can see that there is also a huge difference between what is allowed for the US and what is allowed for the UK. Can the UK have Metrics? Yes they can! Can the UK have Imperial? Yes they can! Can the UK have a mixture? Yes they can! That is quite exceptional, and it applies to no other country. And no amount of obfuscation and chop logic can hide it. Michael Glass (talk) 13:46, 8 September 2009 (UTC)

The United States can have a mixture of metric and non-metric units too. Otherwise Americans would be unable to measure electricity. --Jc3s5h (talk) 16:06, 8 September 2009 (UTC)

Those about the US and the UK are not "policy", they are examples of what "the units most appropriate to them" are in those countries. In cases where the "units most appropriate" to a US topic are metric (e.g. some engineering topics in which Americans would use metric, too), there's nothing to forbid the use of metric units in an article about it. I've just changed the page to include "for example". --___A. di M. 16:45, 8 September 2009 (UTC)

Some issues with the Currency section

I have seen some recommendations in the "Currency" section which contradict other parts of the MoS and common sense:

Conversions "rounding to the nearest whole unit": for large amounts of money that will be generally over-precise (considering also how wildly exchange rates can fluctuate), and it doesn't work for amounts which are less than one dollar (or one euro, or one pound). I'd rather say: "rounding to one or two significant figures, ... eg., 10,000 Swedish kronor (approx. €1000, US$1400, or £800 as of August 2009)".
"When possible, always link the first occurrence of a symbol for lesser-known currencies": I'd say "spell out the first occurrence", as we do with other units of measurement. Having the reader needing to hover on a link in order to obtain information is contrary to WP:ACCESS; "some editors consider it unnecessary to link the symbols of well-known currencies, but doing so can often be helpful to readers, as many countries use dollars or pounds as their base currency": if it means what I think, it goes against WP:EGG; a reader seeing "$500" in an article about Australia won't even notice the figure isn't in Australian dollars unless they happen to hover on the piped link, and the information is completely lost when the article is printed. If it doesn't mean what I think, it should be made clearer. Also, linking to United States dollar in articles not specifically about economy would border on WP:OVERLINKing.

What d'y'all think? --___A. di M. 14:30, 30 August 2009 (UTC)

I agree 100%. The precision in conversions of money should be governed by the same mathematical rules that govern any other conversion, which is properly a balancing act that ensures valid data is not lost and that false precision is not created.
As for spelling out a first occurrence of a new monetary unit, I also agree. Experienced Wikipedians tend to fall victim to the phenomenon of assuming that the typical reader is nearly as familiar with “*basic common* knowledge” as we Wikipedians are. You can see this in our computer articles, where editors will often instantly launch into symbol‑itis in lead paragraphs (The Banana Jr. 9000 was the first consumer-grade computer to come stock with 1 MB of RAM). The saving grace does not lie in that we have provided a link to what is an unfamiliar term for someone who has no knowledge about computers and wants to read up on the subject; novices to a subject should not have to click back and forth from article to article to wade through and parse key sentences.
Clearly, certain minimum skill sets and knowledge must be assumed in our readership; they should be familiar with the “United States” and “ice” and these terms need not generally be linked. Distinctions between US$400 and AU$475 when they are juxtaposed in the same sentence are exceedingly subtle, easily overlooked, and assume far too much familiarity with currency symbology than should be expected of novices to the subject. Currencies of countries that are not primarily English speaking, the currencies of primarily English-speaking countries that are less well known, and the currencies of English-speaking countries whose symbols can easily be confused within the same article should generally be spelled out on first occurrence and their symbols parenthetically introduced.
As per your example, parenthetical conversions, such as 10,000 Swedish kronor (approx. €1000, US$1400, or £800 as of August 2009) are not spelled out in order to reduce wordiness. This broad principle is used for other units of measure on Wikipedia, such as The typical sack of sugar in the United States is sold in five-pound bags (2.3 kg) and the principle does not change just because the units of measure are currency.
A notable exception, in my opinion, would be in articles that are clearly about the United States or England. Take the example of an article on the World Trade Center. It seems to me that it should be sufficiently non-confusing for the intended readership if we write New Jersey Governor Robert B. Meyner objected to New York getting a $335 million project or, in the London Bridge article, …it cost approximately £1,000,000 to build. The only other notable exception that comes to mind to this basic principle—that the first occurrence of the primary monetary unit should be spelled out—would be a highly advanced economics article directed to an obviously expert readership. Greg L (talk) 16:36, 30 August 2009 (UTC)
As for the first paragraph, the fact is that, unlike the number of millimetres in an inch, the number of Czeck crowns in a US dollars is not a constant; so what is "valid data" today will be nonsense tomorrow. We should convert a length of 34.5678 millimetres to 1.36094 inches to preserve the precision of the original; but converting the the exact value of the Nobel Prize in Physics from 10,000,000 Swedish kronor to $1,435,234 (I'm making these numbers up) will be meaninglesss as the last digits will vary wildly, unless we add "(as of 30 August 2009 at noon at the London Exchange)" or something. So I think it is saner to just chop the value off to one or two sigfigs and to notice the conversion as approximate. As for the second paragraph, I agree, but I'd have choosen a better example: someone who doesn't know what RAM is wouldn't be helped by "random access memory", and OTOH there are many people who have an idea of what RAM is but can't recall what the acronym stands for. It'd be like suggesting that we use "lysergic acid diethylamide" rather than "LSD". As for the last sentence, I don't think it's useful to mention that explicitly; an economy article using so many currencies that even spelling out the first occurrence of each one would be cumbersome would be something so rare that WP:IAR is sufficient to handle that. --___A. di M. 15:45, 31 August 2009 (UTC)

Most of what you wrote are intricacies that need to be factored in, but shouldn’t throw tar at the basic principles. Rounding to one significant digit in a monetary conversion—particularly when it is accompanied with a “August 2007 exchange rate” clause— ~~is far too imprecise~~ might be too imprecise in some circumstances.
As for justifying launching straight into symbol‑itis by citing the example of LSD, that’s not an appropriate example (although I now have a better appreciation for the hazards of arguing with you). It is simply common technical writing practice observed by everyone from Encyclopedia Britannica to the Associated Press to Aviation Week & Space Technology to write out “National Aeronautics and Space Administration” before using NASA. The same would apply in any encyclopedic treatment on a computer-related article: one spells out “megabyte” on first use as well as “random access memory”. However, for terms that are universally known by the general public only by their abbreviations, such as DNA and LSD, one flips this principle around; one uses the abbreviation first and then parenthetically spells it out.
It’s important to not seize upon examples and use them to justify practices without really considering all the implications. You have an influential role on WT:MOSNUM and others put great credence in your opinions. We already have far too many editors launching straight into editing WP:MOSNUM without first properly discussing things on WT:MOSNUM, where they can learn about broader implications and nuances that muddy the waters.
This reminds me of the king who would have people come up to him at social events and tell him this or that and he didn’t have a wise, kingly-sounding response. He asked his wise men to suggest a response that would be universally true and wise-sounding that would be suitable in nearly any circumstances. They came back and told him to respond “All things must come to an end.” I’ve recently discovered another phrase that will work for most any situation where someone is complaining about some right-wing practice or belief or some left-wing practice or belief, or any sort of issue that is imbalanced and extreme in some way. It goes “Well… when you dig down into the details, it’s a bit more complicated than that.” Try it yourself; it works exceedingly well. And that principle applies here, when it comes to diverging from the general practice of spelling things like “MB” out on their first occurrence and parenthetically introducing the acronym. Indeed, “LSD” is a flipped exception to the rule (as is “DNA”). So the same goes here: Well… when you dig down into the details, it’s a bit more complicated than that. Greg L (talk) 18:46, 31 August 2009 (UTC)
I wrote "to one or two significant figures" [emphasis added] actually. Well, maybe that's too vague, but I don't think that you will ever need more than that. A reader reading about the value of the Nobel Prize in Physics (or whatever) won't particularly care whether it was closer to $1,403,000 or to $1,404,000 on a particular day at a particular location; telling him that it's roughly $1,4 million is enough for him to have an idea, and it won't need to be updated every week. (Exchange rates can vary in the third significant figure in much less than one month, and we don't want to have to note down the day a conversion was made.) This is my opinion, at least; what do other editors think? (As for "RAM", I guess that many more people know it by its acronym than by its full name, too; spelling out "megabyte" is fine, I was just pointing out that "RAM" is a bad example for the general principle—with which I agree—citing "LSD" as an even worse example. Sorry if I wasn't clear. Anyway, I didn't mean to start a discussion on how to refer to RAM, which wouldn't even belong here.) --___A. di M. 19:38, 31 August 2009 (UTC)

Our views seem to be converging—notwithstanding that our writings seem to get in the way. It is always good to look to Encyclopedia Britannica when thinking about what is good, project-wide style. I aspire to own the EB set one day but currently have to drive to the library. It’s noteworthy that they never launch into “MB” and “KB” in their computer articles; they always keep on writing out “megabyte” and “kilobyte” from stem to stern. I haven’t looked at how they handle LSD, but I am quite confident they use “LSD” from stem to stern and mention “lysergic acid diethylamide” only in passing. I learned, long ago when writing business plans, to reinforce trade-specific terminology by spelling it out several times and to introduce the abbreviation parenthetically only when the term is being repeated often enough that it becomes nearly tedious and the reader is mentally prepared—even anxious—for the damned abbreviation. I think this compromise technique is superior to Encyclopedia Britannica’s practice of always writing it out clear to the end of the article.
As for precision in monetary conversions, your proposal of two significant digits passes my initial *grin test* for what would be appropriate as a general guideline. To be sure, I suppose I would need to see some examples in context—perhaps even some conflicting examples where one practice seemed more appropriate in one case, and yet another practice seemed more suitable in another case. Sometimes the rule-set can be obscure until you, uhm… dig down into the details.
Memorializing common sense into technical writing guidelines can be a tedious effort. Generally, my preference in these sort of matters is to not prescribe rigidity in guidelines and to instead provide a global proscription of what not to do and give examples of good practices and bad practices. For instance:

When showing a a conversion to another monetary unit, avoid excess or false precision. For instance, do not write The Manhattan Project cost the U.S. $2 billion at the time (equivalent to £14.1 billion in 2007).

To me, this allows a bit more freedom for editors to use common sense and affords experienced editors a bit more latitude to fix errors. Greg L (talk) 21:00, 31 August 2009 (UTC)

BTW, here's my proposal, with changes from the current version marked:

Currencies

Shortcuts
WP:$
WP:£
WP:€
MOS:CURRENCY

See also: Wikipedia:WikiProject Numismatics/Style § Article titles

Which one to use

In country-specific articles, such as Economy of Australia, use the currency of the country.

In non-country-specific articles such as Wealth, use US dollars (US$123), the dominant reserve currency of the world. Some editors also like to provide euro and/or pound sterling equivalents, formatted as described in the next section.

Formatting

~~Fully identify~~Use the full name of a currency on its first appearance (~~AU$52~~52 Australian dollars); subsequent occurrences ~~are normally given without the country identification or currency article link~~can use the symbol of the currency (just $88), unless this would be unclear. The exception to this is in articles related entirely to US-, EU-, or UK-related topics, in which the first occurrence may also be shortened ~~and not linked~~ ($34, €26, and £22, respectively), unless this would be unclear, and in places where space is limited such as tables, infoboxes, and parenthetical notes. ~~Avoid over-identifying currencies that cannot be ambiguous; e.g., do not place EU or a similar prefix before the € sign.~~When there are different currencies using the same symbol, use the full abbreviation (e.g. US$ for the United States dollar and AU$ for the Australian dollar, rather than just $) unless the currency which is meant is clear from the context.

Do not place a currency symbol after the value (123$, 123£, 123€), unless the symbol is normally written as such. Do not write $US123 or $123 (US).

Currency abbreviations that come before the number are unspaced if they consist of or end in a symbol (£123, €123), and spaced if alphabetic (R 75).

If there is no common English abbreviation or symbol, use the ISO 4217 standard.

Ranges are preferably formatted with one rather than two currency signifiers ($250–300, not $250–$300).

Conversions of less familiar currencies may be provided in terms of more familiar currencies, such as the US dollar, euro or pound sterling. Conversions should be in parentheses after the original currency, rounding to ~~the nearest whole unit~~one or two significant digit and noting the conversion as approximate, with at least the year given as a rough point of conversion rate reference; e.g., ~~1,000 Swiss francs (US$763 in 2005)~~10,000 Swedish kronor (approx. €1000, US$1400, or £800 as of August 2009).

For obsolete currencies, provide if possible an equivalent, formatted as a conversion, in the modern replacement currency (e.g., decimal pounds for historical pre-decimal pounds-and-shillings figures), or at least a US-dollar equivalent as a default in cases where there is no modern equivalent.

When possible, always link the first occurrence of ~~a symbol for~~ lesser-known currencies ~~(₮146)~~(146 Mongolian togrogs); some editors consider it unnecessary to link the symbols of well-known currencies, but doing so can often be helpful to readers, as many countries use dollars or pounds as their base currency, and not all readers are familiar with the euro.

The names of currencies, currency subdivisions, coins and banknotes should not be capitalised except where normal capitalisation rules require this (for example, at the start of a sentence).

The pound sterling is represented by the £ symbol, with one horizontal bar. The double-barred ₤ symbol is ambiguous, as it has been used for Italian lire and other currencies as well as that of the British. For non-British currencies that use pounds or a pound symbol (e.g., the Irish pound, IR£) use the symbol conventionally preferred for that currency.

--___A. di M. 20:02, 31 August 2009 (UTC)

It seems to me that a modern equivalent (i.e. taking inflation into account) to an obsolete might be more useful than a straight conversion according to the rate at the time of transfer. Convert five pounds twelve shillings sixpence Australian, for example, not to $11.25 but to what five pounds twelve shillings sixpence is worth today. JIMp_talk·cont 01:00, 4 September 2009 (UTC)

I thought that was what "an equivalent, formatted as a conversion, in the modern replacement currency" was supposed to mean; but now I see it's not clear enough and should be clarified. (BTW, in Italy I see amounts in euro converted to liras with the original exchange rate of 1,936.27 L/€ all the time all the time, as if there hadn't been seven years of inflation since the lira was last used. I've never understood the point of that. I know there are people still thinking in liras, but IMO it's high time they learned to think in euros, as they'll have to use them to the end of their life.) --___A. di M. 11:02, 4 September 2009 (UTC)

[Just for your information, English speakers usually say (and write) lire as the plural of lira. Liras is actually rather unusual. Put it down, I guess, to travellers, Italian stock, Italian-speakers and local collectors of stamps, coins and books, as well as all the pastas we pluralize (linguine, lasagne, etc.), although pizze for pizzas is as unusual as "liras" for lire and paste is too easily confused with the paste that sticks together.] —— Shakescene (talk) 04:50, 5 September 2009 (UTC)

Dithering incomprehensibility?

MoS has just imported much of MOSNUM's Units of measurement section, since we had some pretty iffy stuff that hadn't been reviewed for a while. Wikipedia_talk:Manual_of_Style#Another_query. One import sticks out, though, concerning the old imperial versus metric units in UK-related articles:

... for the UK Imperial units for some topics and metric units for others, and a mixture of units for others (see, for example, the Times Online style guide under "Metric").

The Times online style guide says not to mix the systems within an article. Hmmmm. It strikes me as ludicrous that the primary and converted units should be switched here and there in the course of an article: heck, choose one as primary and stick to it, surely? To do otherwise is a really unprofessional look. MoS had this, until yesterday, which is much better, IMO:

UK-related topics may have either SI (generally preferred) or imperial units as the primary units.

Seems more professional and a lot simpler, don't you think? Can MOSNUM import MoS's previous point? I'm encouraging MoS to go back to it rather than to retain what we've just imported (on this matter alone—the rest is better). How about it? Tony (talk) 11:47, 8 September 2009 (UTC)

The Times says "try not to mix the two systems in a single article. In general, we should prefer the metric" [emphasis added] and then goes on with a longish list of exceptions. There's no point in avoiding giving the distance between Oxford and London in miles only because that'd be inconsistent with the fact that mean temperatures in Oxford are given in Celsius degrees: they are different types of measurements, and people in Oxford do normally use miles for distances and Celsius degrees for temperatures. How 'bout: With topics strongly associated with places, times or people, put the units most appropriate to them first. In US articles, they usually are United States customary units; for the UK, they usually are metric units for most measurements, but imperial units for some measurements such as road distances and draught beer (see, for example, Metrication in the United Kingdom and the the Times Online style guide under "Metric"). --___A. di M. 12:23, 8 September 2009 (UTC)

Good luck!Michael Glass (talk) 13:01, 8 September 2009 (UTC)

ADM, are you really suggesting that miles (km) be swapped for kilometres (mi) within an article, higgledy-piggledy? Bit excessive, isn't it? If there are hard-line anti-metric editors on an article, I'd rather it were all imperial primary, metric converted. Tony (talk) 13:11, 8 September 2009 (UTC)

I don't think anyone is suggesting that if one is measuring road distance, one would switch between "miles (km)" and "kilometres (mi)". But one might write Mr. Banks used 8 8 litres (1.8 imp gal) of petrol to drive 62 miles (100 km) since petrol is sold in litres. --Jc3s5h (talk) 16:02, 8 September 2009 (UTC)

While I grudgingly accept for the moment the line that UK articles may choose either system, I think an insistence that the real-life mess that country has made of metric conversion should not be paraded in a serious online publication such as WP. I say: normally choose one or the other as primary for the whole article, just as MoS said until yesterday. That in itself is a significant concession by many editors—UK and non-UK. Tony (talk) 16:23, 8 September 2009 (UTC)

I am not. I'm saying that there's no reason to forbid, say, the use of degree Celsius and miles in the same article, if it is about a place where those are the units in most widespread units for temperatures and road distances. I still believe that "comparable" measurements should use the same unit in each article. But I can't see how the "miles and degrees Celsius" combination is any worse than the "miles and degrees Fahrenheit" combination. (By "comparable" I don't just mean "of the same physical dimension", I mean "which makes sense to compare in the given context": I can see no particular reason to avoid giving a road distance as 1.4 miles and a mountain height as 7,400 feet, even in the same sentence. And using two units whose ratio is 1,609.344 instead of 5,280 isn't inherently worse.) --___A. di M. 16:40, 8 September 2009 (UTC)

Tony's positon means that any article mentioning a measured electrical quantity must list metric units first throughout, even if it is strongly associated with the United States (or, the article must be considered abnormal). An example of such an article is Hoover Dam. --Jc3s5h (talk) 16:54, 8 September 2009 (UTC) modified 17:04 UT.

What about articles measuring time in seconds? :-) --___A. di M. 16:58, 8 September 2009 (UTC)

I don't know about electrical quantity, but road distances and draught beer seem to be just fine with the chosen system as primary, consistent with the rest of the article. Remember, this is just a debate about which one goes in parentheses. Switching them around for beer, but back again for milk, and then whirling back for building heights (or whatever the mess is) is likely to make the text slightly more difficult to read. Non-UK readers would be forgiven for thinking these were editing slips. In any case, I see no consensus to change what MoS had before. Tony (talk) 02:48, 9 September 2009 (UTC)
How is John had 1 imperial pint (0.57 L) of beer in a pub in a 97 metres (318 ft) building any harder to read than if the metres and the feet were swapped? The pint isn't defined as a "simple" number of cubic metres, but it isn't defined as a "simple" number of cubic feet, either; even if it did, I can't see how it matters here. --___A. di M. 16:25, 9 September 2009 (UTC)

So what is the wording to be, if we must accept this unique inconsistency? And the other global point in that section about consistency in articles will have to be modified, yes? I do note that this is a change in MoS's advice from last week's. No one seemed to notice it until I brought it up. Tony (talk) 15:06, 10 September 2009 (UTC)

It's not "unique inconsistency". I can't see how measuring heights of buildings in metres and volumes of beer in pints is any more or any less inconsistent than using feet and pints, or metres and litres. To be really consistent you'd have to use metres and cubic metres, or feet and cubic feet, and of course no-one would do that. (Using different units for measurements of the same type, such as John weighs 150 pounds (68 kg) and James weighs 79 kilograms (174 lb), would be serious inconsistency which I'd avoid unless necessary, but no-one's advocating that.) As for wording, I think the bullet present now (as of 15:40, 10 September 2009 (UTC)) starting with "With topics strongly associated ..." is fine. --___A. di M. 15:40, 10 September 2009 (UTC)

nbsp for dates?

It was recently suggested to me at a FAC that dates should be given non-breaking spaces. Is that correct according to the MoS? Thanks. Parsecboy (talk) 00:23, 9 September 2009 (UTC)

I believe it's on the "consider doing" side rather than the "you must" side of the scale; and I think I'm right in saying it only ever applied to month and day / day and month, not between them and the year. Tony (talk) 02:50, 9 September 2009 (UTC)

I use a non-breaking space (or try to if I don’t forget) because…

The whiz-bang fanada-yadda-yadda later broke ground on 6
February and was finished in less than a year.

…looks better when it reads

The whiz-bang fanada-yadda-yadda later broke ground on
6 February and was finished in less than a year.

So a non-breaking space looks better. Requiring it in MOSNUM might not be all that practical since dates are written so often and because there are so many editors who wouldn’t want to bother (or be knowledgeable about such advanced techniques). However, permitting the practice so that more experienced editors can make it part of their clean-up seems OK. That’s my 2¢. Greg L (talk) 20:17, 9 September 2009 (UTC)

short dates using slashes

I'm in the middle of a dispute at The Shells over the policy regarding dates formatted as MM/DD/YYYY. I was under the impression that this format is deprecated due to ambiguity issues and should be converted to YYYY-MM-DD (or spelled out) in every case. I would appreciate it if someone could take a look at the current discussion and provide some guidance as to whether this is policy or not. Thank you. ~ Paul^T+/_C 07:29, 10 September 2009 (UTC)

I don't think it is discussed in MOSNUM, but I think it should to cover instances such as this. Dabomb87 (talk) 12:24, 10 September 2009 (UTC)
It is recommended that dates be written as 11 September 2009 or September 11, 2009 (to take today's date as an example) in order to avoid ambiguity. The use of DD/MM/YYYY (or MM/DD/YYYY) is ambiguous shouldn't be used. The use of 2009-09-11 is okay where concision is required (e.g. in tables), or where dates can be sorted. I'll have a look at the page you indicate and add $0.02-worth if appropriate. HWV258 22:40, 10 September 2009 (UTC)

MOSNUM needs to be merged into MoS

Dear colleagues

At MoS, we realised that the Units of measurement section was in a bad mess. Michael Glass and A. di M. have both assisted by pasting in the more up-to-date version from MOSNUM's Units of measurement.

This highlights the absurdity of having two different pages, when MoS main covers just about all of the scope of MOSNUM. There is insufficient difference in the two scopes of MoS main and MOSNUM to warrant the fragmentation of guidance, discussion and monitoring. More importantly, it is undesirable for the two talk pages to be fragmented.

The MoS pages as a whole are in an uncoordinated mess, and it would go some way towards serving the project better to merge MOSNUM into MoS main. All of the main sections (except, oddly, Currencies, which is a pretty important one for general editors), are there.

Why don't we make things easier for ourselves? Any highly specialised guidance in MOSNUM that is not here could easily be sequestered into either a separate subsection or—better IMO—an appendix, here. The merger should not add much text to MoS main (indeed, everything needs significant rationalisation on the micro-scale). Tony (talk) 10:56, 10 September 2009 (UTC)

The main page of the MoS is already bloated enough without detailed guidance about "Scientific notation, engineering notation, and uncertainty" and whatnot, which more than 90% of editors will be very unlikely to care about. If anything, I'd do the reverse: I'd remove obscure items in the main MoS which are also found in MOSNUM (e.g. "Uncalibrated (bce) radiocarbon dates", the bit about units used in scientific literature, etc. IOW, I wish the second and third sentence of WP:MOS, "This main page contains basic principles. Additional subpages of the Manual of Style, listed and linked in the menu on the right-hand side of this page, explore some topics in more detail.", were true. --___A. di M. 13:08, 10 September 2009 (UTC)

I understand what you are saying, but the current situation in which two small groups of editors are divided into two talk pages about scope that entirely overlaps is unconscionable. I believe the solution is to (1) audit the differences in wording and iron them out by negotiation; (2) sequester the more specialist information on numbers and dates in both pages into link-targeted appendices at MoS; and (3) delete MOSNUM.

This would probably reduce the size of MoS a little (aside from the appendices), which is a critical need. We need to make the style guides more accessible and less daunting to the majority of editors. The pages are gaining a bad reputation for bloat and inaccessibility.

~~In essence, I'm proposing that the binary structure (MoS vs MOSNUM) be changed so that it's main text vs appendix, in MoS alone.~~ Tony (talk) 15:02, 10 September 2009 (UTC)

The scope of WP:MOS and of its subpages is already clear enough: "This main page contains basic principles. Additional subpages of the Manual of Style, listed and linked in the menu on the right-hand side of this page, explore some topics in more detail." The real problem is that WP:MOS contains loads of stuff which isn't supposed to be there according to that. For example, WP:MOS could just say that conversions from metric to US/Imperial and vice versa are needed when using a unit which isn't used in some parts of the English-speaking world (one sentence), while pointing at WP:MOSNUM for more detailed info on how the present the conversions, etc. --___A. di M. 17:49, 10 September 2009 (UTC)

Quoting Tony: The MoS pages as a whole are in an uncoordinated mess. I couldn’t agree more. I also agree with A. di M. that detailed guidance on numbers is going to be hard to shoe-horn into one venue. The “mess” is going to be chronic, in my opinion, as long as MOS and MOSNUM are not locked down with gate-keepers and remain free for anyone to edit whenever they please. All it takes is some editor with a deep conviction that the BIPM will soon endorse esperanto and how words from this *universal language* should be permitted in our articles (citing WP:ENGVAR or some other inane argument) in order to thoroughly waste everyone’s time here and further mess up MOSNUM. (*sigh*). Greg L (talk) 15:05, 10 September 2009 (UTC)

Shoehorn? But it's already in MoS. How much extra is there here that's not? Tony (talk) 15:08, 10 September 2009 (UTC)
Another possibility is to retain only the specialist stuff here, still under the name MOSNUM. It would be much smaller. But I think an appendix at MoS would be a much better way to organise it. Tony (talk) 15:13, 10 September 2009 (UTC)

These topics are what I see as being too involved for MoS, and suitable for retaining in a specialised MOSNUM (linked to from MoS):

Uncalibrated (bce) radiocarbon dates
Some of the calendar section (?)
Time zones (if retained at all)
Delimiting (but a short summary to be inserted into MoS)
Natural numbers
Repeating decimals
Non-base-10 notations
Scientific notation, engineering notation, and uncertainty (but a short summary in MoS)
Avoiding ambiguities (it's already in shortened form in MoS)
Most of "Units and symbols often written incorrectly"
Quantities of bytes and bits
Adopting suggestions from standards bodies
The table from "Common mathematical symbols" (?unsure)
Geographical coordinates

It's quite large enough for a MoS subpage. What urgently needs to be removed is the common-person stuff that is already in MoS. Editors need to know a lot of stuff about numbers and dates, but the field crosses into engineering and science in a way that clutters the dummies' guide badly. Tony (talk) 15:31, 10 September 2009 (UTC)

Quoting Tony’s second post: Another possibility is to retain only the specialist stuff here. I think something along these lines makes sense. Upon reflection, I think I agree with Tony’s basic premiss that stripping out too much of the basic essentials of number and date-related stuff from MOS and putting into MOSNUM tends to keep the information away from all but the most diligent and thoughtful editors. So…
Perhaps the thing to do is to harmonize MOSNUM into MOS, but instead of having a “MOSNUM” for the specialized stuff, we merely have sub-pages of MOS that expand upon lengthy topics (such as scientific notation). I suspect little (or even none) of this need to have “Click here for main article”-stuff would be required if there was a concerted effort to streamline MOS and MOSNUM, which suffers from bloat‑itis in many areas.
I suspect one of the “yuck factors” for many of us is the unconscious realization that there is a somewhat different set of editors that frequent MOS and harmonizing will result in social chaos—similar to neanderthals and modern humans finding each other hunting the same game in France 80,000 years ago (“you funny looking”). Perhaps we can keep the dates and numbers section segregated below a border via transclusion so that MOSNUM can keep its discussions on a separate talk page. I think it would prove exceedingly unwieldily if we had an integrated talk page for all issues. Greg L (talk) 15:54, 10 September 2009 (UTC)

I've struck through some of my earlier comments. This does seem to be the way to go: two separate pages still, without the massive overlap. Tony (talk) 16:01, 10 September 2009 (UTC)

Or transclusion, like how we handled RfCs, which occurred in their own articlespace but the results appeared in a unified page. We could still have entirely separate sub-pages for the particularly complicated and expansive topics. Greg L (talk) 16:06, 10 September 2009 (UTC)
Transclusion would miss the point (that main MoS should be a summary). --___A. di M. 17:49, 10 September 2009 (UTC)

A possible solution would be to centralize discussion on all MoS topics, but retain subpages for content. (Redirect sub-talk-pages to one location.) That way, you'd be able to avoid editors at MOSNUM operating under different principles and objectives than the editors at MOS, leading to inconsistent guidelines. There might be issues with the talk page getting filled too quickly, however.
Nevertheless, I support simple guidance at MOS, with complicated guidance in main articles such as MOSNUM. TheFeds 16:42, 10 September 2009 (UTC)

I agree entirely. Let's put the readers' needs first: we have general editors, who need to know the usual stuff about numbers, dates, currencies, units of measurement; and we have editors who write technical, engineering, scientific articles, and those who need to know about certain arcane historical dates et al. There is an urgent need to rationalise, trim, make user-friendly. One obvious way of achieving this is to draw a boundary between needs of regular, everyday, general editors who don't need to know about radiocarbon dating and whether to use (km/s)/Mpc or s⁻¹; and the tiny proportion of editors who do need advice on this. Putting it all together just gives ammunition to those who say the guides are inaccessible: I agree with those people.

So if "summary" means such a new boundary, so be it. Tony (talk) 02:45, 11 September 2009 (UTC)

Common Era

I thought the Manual recommended CE notation, but I am obviously mistaken. I am probably opening a can of worms, and pardon any cluelessness, but why does the MoS take such an ambiguous stance about the use of the Common Era year notation? It is becoming overwhelmingly dominant in academia. Most popular periodicals mandate its use in their house styles. Even a large number of explicitly Christian publishers now make use of the CE notation. While I can appreciate the pragmatic impossibility of mandating the Common Era notation, should we not follow the real world example and prefer its use? Or am I completely missing something? Or am I just hopelessly naive that a consensus could be built for such? Vassyana (talk) 14:08, 10 September 2009 (UTC)

MOSNUM is “ambiguous” because there is such conviction on both sides of the issue. My preference is to write in a manner that reads most naturally for a given readership (article). Greg L (talk) 14:50, 10 September 2009 (UTC)
- Vassyana: the issue apparently caused a lot of grief a few years ago. I think the current line, which favours neither but insists on within-article consistency, is ideal. Tony (talk) 15:33, 10 September 2009 (UTC)
  - Thanks for the clue in. If it was something that lead to that much strife, it's probably best to let sleeping dogs lie. I'll dig through the archives myself, but does anyone know where the bulk of those discussions took place? Vassyana (talk) 16:44, 10 September 2009 (UTC)

I don't know where the main discussion took place, but I do have Common Era and Anno Domini on my watchlist. There is a constant stream of (usually IP) editors who stop by to alter those articles to favor one notation or the other.

Also, for purposes of editing Wikipedia, I don't think academia should be equated to the real world. --Jc3s5h (talk) 16:58, 10 September 2009 (UTC)

Exactly. Wikipedia is an encyclopedia; it isn’t a religious publication and we therefore don’t automatically follow the practices of religious publications—even if God in Her infinite wisdom were to anoint it with holy oil. Nor do we follow what academic articles do because they are “academic” and are the product of smart people with leather patches on their elbows. And if the BIPM anointed one method and declared the other smelled like baby diapers, that doesn’t matter either.
Often we can look to the AP Manual of Style and the practices of Encyclopedia Britannica and World Book to see what English-speaking peoples are routinely exposed to and what therefore seems most natural and encyclopedic. Too often, ninth-graders smitten with the “Forget what Encyclopedia Britannica does, I can change the WORLD by editing Wikipedia” are the source of unnecessary grief here.
My preference, again, is to always use whatever writing style and words are most natural and interrupts the train of thought the least. I know: shocking. Greg L (talk) 19:38, 10 September 2009 (UTC)

My preference is to follow the ISO 8601 standard and use positive numbers for years > 0, negative numbers for years < 0, and 0000 for year 0. However, that is too rational for much of the population here on Wikipedia, and they tend to have what my grandmother would have referred to as a conniption fit when the idea is suggested. It does have the advantage of completely ignoring the religious issues, which is always good.RockyMtnGuy (talk) 21:54, 10 September 2009 (UTC)

And your grandmother would be justified to have a conniption fit. The purpose in any good technical writing is to communicate with minimal confusion. If we adopted every keeno-sounding idea from every notable alphabet-soup standards organization (BIPM, ISO, IEC, NIST, etc.) just because it is logical and cool and all that; all we would be doing is trying to put Wikipedia in the forefront of the effort to Lead the way and show the world the path to a better and more logical future that is liberated from the many shortcomings and inconsistencies of English and gets the English-speaking world ready to join the United Federation of Planets.©™®. Under this banner, we could write stuff like how A 67 centiuno (67 cU) majority is required in the U.S. Senate to…. Fortunately, the IUPAP guy who proposed the “uno” had the courage to retract it after it flew like a wet noodle in the real world.
Unfortunately, the IEC didn’t retract their proposal to say A memory chip with a capacity of 256 kibibytes and we had a pack of editors pushing that here on Wikipedia for three years. (*sigh*)
The proper thing to do is not pretend that we are here to help change how the world communicates (our aborted effort with “kibibyte” showed Wikipedia doesn’t have that sort of influence), and simply *communicate* using good grammar, clear writing, and in a fashion that results in the least *!* brain interrupts and is most natural for the target readership.
Wikipedia needs to follow the way the English-speaking world communicates. Sometimes, that means we can look towards the Associated Press Manual of Style, sometimes towards how magazines that are devoted to a particular discipline communicate to their readership (like computer magazines), and sometimes it means looking towards other encyclopedias like Encyclopedia Britannica and World Book to get a clue. Even the best of we Wikipedia editors could benefit from looking to these other sources for tips.
Sometimes, arguing with persistent editors here on WT:MOSNUM seems exceedingly unwise of me. Greg L (talk) 22:55, 10 September 2009 (UTC)

RockyMtnGuy, would you like to share with us how you obtain consent from your data exchange partners whenever you use an ISO 8601 year that preceeds 1583. Of course you do that, because you have read the standard and saw that consent is required by the standard. I'm sure you also read the part about always using the Gregorian calendar, and always make sure you do so, even though the Julian calendar is increasingly likely to be found in ordinary writing in Europe and the Americas when writing about events before the 19th century. --Jc3s5h (talk) 23:09, 10 September 2009 (UTC)

My read of his post is that he isn’t seriously proposing this—just that he’d “like-ta.” Greg L (talk) 23:12, 10 September 2009 (UTC)

My reading of his post is that he isn't proposing it for Wikipedia but that he does use it, including for early and negative years, outside Wikipedia. So I'm curious how he handles gathering all the consent, and dealing with the confusion among the consumers of his early dates, about how come the dates arn't in the Julian calendar, just like they often are in history books. Alternatively, I'd like to get a sense of how may people know the correct name of the standard is indeed ISO 8601, but haven't actually read it.

I blame all the confusion on the ISO. They co-opted a format that was either already in fairly common use, or was so logical that it was bound to come into general use. Then they made a bunch of fussy rules when they should have known it was beyond their ability to inform the world of the rules. Finally they compounded the confusion by charging exorbitant fees for copies of the standard. --Jc3s5h (talk) 23:22, 10 September 2009 (UTC)

In the computer business ISO 8601 is the preferred date format because 1) it's unambiguous, 2) you can sort dates easily, and 3) It involves less funky date arithmetic. I wish somebody had told Microsoft that before we spent all that money on the year 2000 problem.

I didn't have to convince my data exchange partners to use ISO 8601. The standard setting bodies for the industry I was in (the Canadian oil industry) got together with the government regulators, made a decision, informed us that we were all using ISO 8601 for data exchange, and that was it. These type of industry-wide decisions were non-negotiable.
The year 1583 is just a red herring - realistically, you don't exchange data for dates before that time, and if you did, you really can't be sure which calendar is in use. If necessary, you can convert the dates for any year to any calendar.
Historically, different countries adopted the Gregorian calendar at different times. England and Scotland were on different calendars for about 150 years. The provinces of Canada adopted the Gregorian calender on different dates - in fact, Nova Scotia changed to the Gregorian calendar in 1605, then reverted to the Julian calendar in 1710, and then changed back to the Gregorian calendar in 1752.
The Julian calendar itself was not in use before -0044, so I'm not sure it's more valid that the Gregorian calendar for ancient dates.

If you are writing character dates for the general public, either the American or British character date format is okay, but if you are using computers the only numeric date format that is really foolproof and unambiguous is the ISO format. And, yes I do use ISO data format for all-numeric dates.RockyMtnGuy (talk) 06:21, 11 September 2009 (UTC)

I agree. According to the ISO themselves, ISO 8601 deals with “Data elements and interchange formats – Information interchange”. It is clearly intended to standardize how to format and exchanging digital time & date data and simultaneously account for time-zone differences by comparing offsets to UTC. It has nothing whatsoever to do with manuals of style for how to write out human-readable dates in publications.
I don’t know why we are discussing ISO 8601 on MOSNUM; MOSNUM is a style guide and not a bulletin board for Wikipedia developers trying to resolve data exchange protocols between servers. Greg L (talk) 19:05, 11 September 2009 (UTC)

P.S. Our own Wikipedia article on the subject seems to have a serious misrepresentation in the first sentence: It reads as follows:

“

ISO 8601 is an international standard for date and time representations…

”

I find that word “representations” to be exceedingly misleading and possibly the result of POV-pushing by a propeller-head. The clear thrust of the standard is about information-system exchange of temporal data. The ISO isn’t suggesting that there be some world standard for publications like encyclopedias that we write The attacks on the twin towers occurred on 2001-09-11T08:46:00. While following such a practice here might indeed grease our admission into the United Federation of Planets, it clearly isn’t good writing style. Greg L (talk) 19:05, 11 September 2009 (UTC)

Phrases such as "if you are using computers the only numeric date format that is really foolproof and unambiguous is the ISO format" are unclear, because we use computers for a great deal of general purpose writing. If you are editing Wikipedia, you are using a computer. If you are reading Wikipedia, you are very likely using a computer. I think the thrust of the ISO 8601 standard is the representation of dates and times related to commercial events that occur within a computer, or which are recorded in a computer contemporaneously. The crowd of people who deal in such things (and I strongly suspect, wrote the standard) don't give a damn about the Julian calendar, cannot be made to care about the Julian calendar, and proper use of ISO 8601 for historical dates will never be enforced. In my mind, I picture a speaker getting up at an ISO 8601 drafting meeting to talk about which calendar should be used, and all the heavy-hitters leaving the room to take a bathroom break.

There are also problems with scientific and astronomical uses of time, but scientists (including astronomers) are better equipped to deal with standards bodies than the general public. --Jc3s5h (talk) 19:34, 11 September 2009 (UTC)

I fixed the ISO 8601 article. Whoever wrote that lead paragraph either was POV-pushing or didn’t understand the basic nature of the standard. Greg L (talk) 20:20, 11 September 2009 (UTC)
Not necessarily: the standard itself uses the word "representation", and maybe whoever wrote that sentence in the article just naively used the same word, overlooking the fact that it could be misunderstood. --___A. di M. 20:46, 11 September 2009 (UTC)

Not necessarily “not necessarily”. Someone had this colossal pile in it:

The application of the standard is intended to be very broad. It applies to all written communications that contain dates, times, and time intervals regardless of the communication medium (printed, electronic, or hand written) or the location of the sender and receiver (either within an organization, between organizations, or across international boundaries). The application of the standard was never meant to be limited to dates and times processed, stored, and displayed by computers. It applies to all industries and all forms of human activity where accurate and unambiguous representations of dates, times, and time intervals are needed when communicating internationally, nationally, locally, internally, or even privately.

Rather than put a {fact} tag at the end of the section, I deleted it for being what it was: a fairy tale. Read the above—again. Focus on this part: “It applies to all industries and all forms of human activity where accurate and unambiguous representations of dates, times, and time intervals are needed when communicating internationally, nationally, locally, internally, or even privately.” I think I might memorialize that on my home page as the world-record example of misinformation on Wikipedia. It bears zero relationship to reality, as disclosed by the ISO themselves at ISO/Date and Time. Greg L (talk) 22:16, 11 September 2009 (UTC)

Well, that was nonsense.

BTW, a while ago, I read a page by ISO explicitly stating that ISO 8601 wasn't meant so supplant dates written in natural languages. By what you say I guess that's the page you linked, but I can't find that statement in it. (I can't search for it with my browser because I don't remember anything of its wording.) Where is it? --___A. di M. 11:30, 12 September 2009 (UTC)

I tried Google searches using a clue you provided: “natural language” and didn’t immediately come up with anything by the ISO. It appears any further effort on my (or your) part on that article would be short-lived and frustrating; the first page of edit history shows that some 90% of edits are being done by I.P. editors.
It is this sort of experience that makes me long for the practice observed by the German-version of Wikipedia: it’s my understanding that they have all articles guarded by gate keepers. That prevents articles from being mucked up by some kid who wears Spock ears at Star Trek conventions and leaves notes for his mom that read like this:

Mom,
I went to Steve’s house because he got a mint-condition PDP-11 computer when he was at his Linux meeting last night. I’ll be home by 2009-09-12T19:30:00.

In light of the German-version of Wikipedia and how everything there is guarded with a gate keeper, that sorta makes my suggestion that WP:MOSNUM be guarded by a gate keeper seem quite tepid. Greg L (talk) 17:59, 12 September 2009 (UTC)

Edit warring over date format

Can someone please guide me or help me here? I am using one date format in citations, only to have an editor edit war with me over it. Is my below analysis correct? If so, how can I address the matter.

The date format style of citation that I am using, specifically MM/DD/YYYY is not prohibited by Wikipedia from what I can see, but is not directly discussed.

The guidance here states, however:

"Articles on topics with strong ties to a particular English-speaking country should generally use the more common date format for that nation."

Here, the article is about a band from the USA. As it is written in [6], the MM/DD/YYYY style of citation is, in the United States:

"common or prescribed—particularly in military, academic, scientific, computing, industrial, or governmental contexts. See Date and time notation by country#United States."

This guidance also states, per the Wikipedia Arbitration Committee

"it is inappropriate for an editor to change an article from one [acceptable style] to the other without substantial reason." Furthermore, "Edit warring over optional styles ... is unacceptable. If an article has been stable in a given style, it should not be converted without a style-independent reason. Where in doubt, defer to the style used by the first major contributor."

The guidance repeats these admonitions by in effect largely repeating itself when it states, a second time:

"Retaining the existing format -- If an article has evolved using predominantly one format, the whole article should conform to it, unless there are reasons for changing it based on strong national ties to the topic. In the early stages of writing an article, the date format chosen by the first major contributor to the article should be used, unless there is reason to change it based on strong national ties to the topic. Where an article that is not a stub shows no clear sign of which format is used, the first person to insert a date is equivalent to "the first major contributor"."

In this instance the article was stable in the MM/DD/YYYY format. I had created the article and was the major contributor. Another editor, with whom I had just disagreed on a separate issue, has spent the last few days wikistalking me to all my edits, and seeking to delete or revise dozens of them, including all of the date format edits, despite my pointing him to the language of the guidance.

He argues "The date format you are trying to use is ambiguous and does not present an international view of the information." With that, he reverts all of my edits.

Am I correct here that since this is an article with a strong tie to the US, it is appropriate for me to use the common date format for that nation, and that it is wrong for him to wikistalk me to edit war by constantly reverting the date format? If so, what recourse do I have? This of course is especially disturbing given the background I just described. Many thanks.--VMAsNYC (talk) 02:10, 11 September 2009 (UTC)

You are absolutely correct to use US not international, but without the spelling-out of the month, many English-speakers will have trouble knowing which is the month and which is the day: March 5 or 3 May?. Tony (talk) 02:18, 11 September 2009 (UTC)

Just to add a point about citation templates, and using numbers. Here is an example of VM's edits. If you change 2009-08-05, which people know means 5th day, 8th month (this standard has a name, but I forget what it is) to 5/8/09, no one knows what the latter means. I agree with you that you should be allowed (and are allowed) to write August 5, 2009, but 5/8/09 is ambiguous. This is one of the problems (one of the many) with citation templates. SlimVirgin ^{talk|contribs} 02:48, 11 September 2009 (UTC)

I had understood that the reason the guidance by its terms allows for the common style of the English speaking country (and this is clearly a common style of the US), is that the subject of the article is clearly US, as stated in the first sentence, which is an indicator to the reader that the common US style is being used.--VMAsNYC (talk) 03:10, 11 September 2009 (UTC)

But readers in other countries don't necessarily know that, and we can't assume that they will. I think that the US vs. international date format primarily refers to the order of the month and the day (1 January versus January 1) rather than the actual format of the date. Dabomb87 (talk) 03:12, 11 September 2009 (UTC)

SlimVirgin, there is a standard named ISO 8601, but like all standards, it does not apply your use of the YYYY-MM-DD format unless you claim it applies. It would be unwise to claim it applies until you read it, because it is full of pitfalls. --Jc3s5h (talk) 03:14, 11 September 2009 (UTC)

Which is why I'm glad MOSNUM recommends using only human-readable formats (spelled out months) rather than the other stuff with hyphens or slashes. Dabomb87 (talk) 03:16, 11 September 2009 (UTC)

"I had understood...". It may be unpalatable at the moment, but it should now be clear to you that you have misunderstood. Slashed dates are not used on WP (for the reasons stated in the close vicinity). Please try to rise above the dispute you have been involved in, shake hands, and move on with other useful work at WP. Cheers. HWV258 03:39, 11 September 2009 (UTC)

Let me say I'm happy to abide by the rule, and indeed by any consensus that is not at odds with the rule, even if it is different from my understanding. But there seems to me to be a stark difference between what this policy actually says and what you are explaining to me. They just don't match. Perhaps the answer is that the policy needs to be changed, but as it is now it does not match what you are saying as I read it.

First, this guidance says "Articles on topics with strong ties to a particular English-speaking country should generally use the more common date format for that nation". (emphasis added). And wikipedia tells me that the MM/DD/YYYY style of citation is, in the United States, "common or prescribed." That's a direct match -- the word "common" appears in both sentences. To me, that means I can read it as saying "An article with strong ties to the US should generally use a common date format for the US (e.g., MM/DD/YYYY).

What you are saying is very different. It is along the lines of "Articles on topics with strong ties to a particular English-speaking country should generally use "a more common date format for that nation as long as it is specifically mentioned in this Guidance." That is much more limiting, but that is not what this guidance says now.

Nor does the guidance say that the examples given in the following chart as "Correct" are the only acceptable date formats.

Incorrect	Correct
9th May the 9th of May	9 May
May 9th	May 9
May, 2001	May 2001
9 May, 2001 09 May 2001	9 May 2001
May 9 2001 May 09, 2001	May 9, 2001

Maybe that was intended, but as I read it it says that certain formats should be changed to other formats ... but does not suggest that it has mentioned all acceptable formats in those boxes.

In fact, my wikistalker was not even changing the format to a format that appeared in the above chart. He changed it to the "YYYY-MM-DD style" -- and all the guidance tells us with regard to that style is that it is uncommon in English prose, and should not be used within sentences, but may be useful in long lists and tables for conciseness. But there is no mention of whether it is acceptable in reference citations (even though reference citations are discussed a little further down in the guidance) -- so on what basis are people suggesting that that format is in fact blessed by this guidance?

So what in the guidance itself allows the other editor to edit war with me, reverting all my edits (with regard to this US band) from a format that is clearly a common US format to a YYYY-MM-DD format?

If there is doubt about this issue, doesn't the admonition in the intro of the guidance come into play, namely "Where in doubt, defer to the style used by the first major contributor"?

I'm clearer now on what the 2 or 3 people who have responded here think, but have difficulty reading the guidance as saying the same thing. Sorry if I'm missing something -- feel free to point out the error in my analysis.--VMAsNYC (talk) 07:30, 11 September 2009 (UTC)

if i'm following things right, it seems like you're mistaking a Wikipedia article describing date formats used in the "outside world" for a recommendation of forms to use on Wikipedia. but the WP:MoS is where the preferred Wikipedia style is described, and it does include an explicit statement that months need to be spelled out: WP:MONTH. this misunderstanding shows that that's not quite explicit enough, and/or that that needs to be reiterated in other date-related sections. meanwhile, in an article about a US band, the US date format should indeed be used – and on Wikipedia that means month-spelled-out-in-full DD, YYYY. Sssoul (talk) 07:44, 11 September 2009 (UTC)

I'm just quoting the precise language of this guidance, to wit: "Articles on topics with strong ties to a particular English-speaking country should generally use the more common date format for that nation." It doesn't say "the more common date format for that nation, as long as it is one of those mentioned in this guidance." Nor does it state (unless I'm missing it -- very possible) that those date format listed (in the box? in the guidance?) are the only acceptable date formats (and not just examples of good and bad formats).

In any event, am I understanding you correctly that it was innappropriate for another editor to edit war with me (on an article relating to a US band) by inserting YYYY-DD-MM formatted dates (even if it would have been ok had he put in, say, month dd, yyyy dates)?--VMAsNYC (talk) 07:54, 11 September 2009 (UTC)

smile: why does Lily Tomlin's "mommy never told me that i shouldn't shave the kitty" come to mind? (that's meant as mild humour, not any kind of disrespect – i'm a literal-minded sssoul too!) yes, the guidance for Wikipedia articles means "the more common date format for that nation, in the month-spelled-out style used on Wikipedia". edit warring (you were involved as well, right?) is almost never a good idea, but this is the MOSNUM talk page; and see the discussion below regarding YYYY-MM-DD format in sentences. Sssoul (talk) 08:34, 11 September 2009 (UTC)

smile-back-atcha: why does Chicago's "Not Guilty" come to mind -- nah, once he reverted me four times I asked him to stop edit warring, but I stopped before doing the same (though he brought me to the brink). And yep, I'm a fan that if there is a rule in people's minds that matches the sentiments here, it might be reflected more precisely in the guidances (e.g., if there is an exhaustive list of what is permissible, just state it).--VMAsNYC (talk) 09:04, 11 September 2009 (UTC)

Since the article that is being edit-warred over is Written Roads, an American rock band, the readership is likely to have a heavy American representation. Articles are best when they are written with an intended readership in mind. I used Euro-style dates in Kilogram, Fuzzball (string theory), and Thermodynamic temperature, (and several other general-science articles) because—even though I am American—the articles are not closely associated with a particular country. However, if I was writing an article on Spokane River Centennial Trail or Boston Red Socks, I would certainly use the American date format. And, as others have pointed out above, you should spell out the name of the month. Unless you have a table with a narrow column for dates, you should always endeavor to write out the month, regardless of whether it is month/day or day/month. Greg L (talk) 04:14, 11 September 2009 (UTC) ^{(Yes, I can discuss this topic.)}

Any suggestions for how I address the fact that my wikistalker is now deleting any reference of the band in other articles, even if sourced?--VMAsNYC (talk) 04:47, 11 September 2009 (UTC)

Fortunately, I am so loath of wikilawyering and other games played by editors who apparently have no life, I am not expert in these matters. However, wikistalking is prohibited. I may well be corrected on this, but I’d try WP:ANI. You don’t have to be extraordinarily formal and follow all the whiz-bang fancy techniques other complainants utilize when the do their wiki-Perry Mason; just use plain-speak, keep your cool, document the editing behavior of the stalker with diffs, and ask for relief. Greg L (talk) 05:04, 11 September 2009 (UTC)

Ugh. I'm with you -- I'm not really a fan of wikilawyering (other than in the sense of reading through these policies like a lawyer in an effort to glean their intent). And this fellow is far more experienced in these things than I am. He has spent the last few days focused almost solely on changing all of my edits that he could, even in bizarre fashion at times with no substantive impact but it makes you wonder why he makes these edits (all to my edits, after I disagreed w.him on another issue). So, as to the bizarre non-substantive edits, for example he changed my edit Trio to Trio, then changed my entry of all-female to all-female, then changed "publisher= Seventeen Magazine" to "work= Seventeen Magazine". But today he increased his activity to more substantive and disruptive edits -- deleting mentions of the band The Shells in other articles ... see, for example, [7] and [8]. I'm not looking to punish the fellow with a legal proceeding on Wikipedia -- I just want him to stop, and go away. Isn't there a way to get someone to talk to him, short of going through complicated procedures?--VMAsNYC (talk) 06:15, 11 September 2009 (UTC)

Quoting you: I just want him to stop, and go away. Isn't there a way to get someone to talk to him, short of going through complicated procedures? I assume that tracking his I.P. address and hiring thugs to beat the living crap out of him is not something you are prepared to resort to, so… no. Your options are 1) live with it, 2) abandon Wikipedia as a hobby, or 3) go the proper route (ANI) and put a stop to prohibited behavior that prevents you from contributing to Wikipedia. Option 3 doesn’t take much time so long as you keep it simple; far less, in fact, than you have so-far devoted to this thread. Greg L (talk) 18:34, 11 September 2009 (UTC)

BTW, I could be all wet on this and Neon white is correct about “bad grammar” (although he/she means “bad punctuation”), but I believe that when one wants to talk about all of the female bands that exist in the United States, it is is written all female bands I’ve found in the U.S. are of interest to me. If you want to talk about bands that are entirely female, it is properly groups that are all-female bands are of interest to me or all‑female bands are cool. Were I have been handling that, I would have made the title All‑female band and provided a redirect for the non-hyphenated search term. (Hmmm… I just noted that “non‑hyphenated” is hyphenated.) Not only is the article title currently punctuated incorrectly via the omission of the hyphen, but the three instances of “all female” in the body text need to be hyphenated because they are followed by the noun “band”. I’m not sure whether you or your stalker was responsible for which version, but the article is incorrect now. This is an all‑encompassing rule of hyphenation. ;-) Greg L (talk) 21:37, 12 September 2009 (UTC)

Actually, VMAsNYC, whatever about all his other edits, he was right to change "publisher= Seventeen Magazine" to "work= Seventeen Magazine". That's not a "bizarre non-substantive edit". For the name of a newspaper or magazine, "work" is the correct parameter (or you can use "newspaper"), because that puts the name in italics, as is required. "publisher" doesn't put it in italics, and is supposed to be used only where you need to specify what company publishes the newspaper or magazine (not usually necessary). See Wikipedia_Talk:Citation_templates#Work_vs_Publisher....? and the section underneath it also. (And by the way, "cite news" would have been the right template to use, rather than "cite web"). -- Alarics (talk) 15:49, 13 September 2009 (UTC)

Proposed tweaks to Template:Convert's default precision

See the discussion at Template talk:Convert#Some suggestions for changes to the default precision. --___A. di M. 21:01, 12 September 2009 (UTC)

I give up

Looking at edit histories, it seems that all you want to do is argue about MOSNUM rather than MAKE ARTICLES BETTER. I tried. God knows I tried. Now can you please carry on in your little circle of wheter an inch is more than am mile while I go and make Wikipedia better?

In just one month I have realised why people abhor MOS. It is people masturbating, basically. We come to MOS for advice and it changes day by day. Which is about as much use as a snake in an arse-kicking competition. Let the whole thing go away, let it die, nothing good will come of all this bickering. Go away from MOS and do some real work improving the encyclopaedia, not laying down your laws.

I give an example: at {{convert}}. User:Jimp who kinda runs those templates (with the help of many others) responds to comments, says it is right or wrong, if it is missing he adds it. That is CONSTRUCTIVE. Sometimes we have short discussions about how it should be put, and come to a consensus. That is CONSTRUCTIVE. More constructive work is done in Convert in practically improving the encyclopaedia than was ever done at MOSNUM.

S. SimonTrew (talk) 14:50, 11 September 2009 (UTC)

Quoting you: We come to MOS for advice and it changes day by day. Indeed; this is a serious shortcoming on MOSNUM. Until it is locked down and overseen by a couple of gate keepers, which forces editors to constructively discuss and work towards consensus, MOSNUM will forever suffer from instability. That’s not pessimism; it’s simple acknowledgement of reality. It is unrealistic in the extreme to expect (hope) otherwise. Greg L (talk) 18:26, 11 September 2009 (UTC)

Maybe the reason MOSNUM is unstable is that it tries to be too comprehensive. There's no significant disagreement about how dates should be formatted in the vast majority of contexts, and it takes only a handful of sentences to codify that consensus. But when you get into exceptions and special contexts and the variety of people's experiences and edge cases and...well, it's just unnecessary to codify all that, and if you do, it leads to bloat that makes MOSNUM drudgery to read and use, so of course the result is that it won't be. The last thing MOSNUM needs is "gatekeepers" (a.k.a. owners) ensuring its continual refinement ad infinitum.

Remember, policies and guidelines are supposed to be clear and terse, and "editors are expected to use common sense in interpreting and applying these rules". It's unnecessary to attempt to codify all the exceptions and edge cases; that's what the edit-revert-discuss cycle is for. In fact, I'd say that getting into such detail is seriously counterproductive, because P/G bloat (1) discourages people from reading and attempting to follow them, or (2) distracts them from what they intended to edit. I'm sorry to say, there have been plenty of times I went to MOS to try to do right by an edit, and by the time I located the right page and read through it thoroughly to be sure I wasn't overlooking some exception, I was too distracted to feel enthused about going back and making the edit after all. I get a lot more done when I ignore MOS, and so far none of my edits has been challenged because of it. But even if such a challenge did happen, as long as the intention was to alleviate some real flaw in my edit (as opposed to forcing me into compliance with some picayune rule enumerating terpsichorean angels) I would welcome it. Unconventional (talk) 01:50, 14 September 2009 (UTC)

I find your sentiments to be not at all uncommon and I respect them. However, manuals of style, such as the Associated Press’ Manual of Style are quite comprehensive. And the AP MoS is intended to be used by professional writers at newspapers, the vast majority of whom have degrees in journalism. Given that Wikipedia is “the encyclopedia to which *anyone* many contribute,” I find arguments predicated on the notion that “common sense”-alone will result in a quality product here on Wikipedia to be overly optimistic and uncompelling.
Moreover, given that exceedingly small groups of editors—only a handful really—can hijack MOSNUM and produce guidelines that sanctify garbage like making Wikipedia the only general-interest publication on this pale blue dot to use the IEC prefixes (a computer with 256 kibibytes (KiB) of RAM), and then spread such naive and shortsighted notions across literally hundreds of articles (making Wikipedia look utterly foolish), it is, IMO, important to prevent problems before they start. This is especially true with contentious issues in which some editors see things in black & white and have strong convictions. Greg L (talk) 02:17, 14 September 2009 (UTC)

As for having a style guide as comprehensive as that intended for professionals, your arguments elude me. This is not the AP; we don't have an efficient authority structure to dictate adherence to the style guide, we don't have professional writers (let alone with journalism degrees) committed to it, and we don't reject content on the basis of imperfect form. Why you would think that what works for AP should work for WP escapes me.

My argument is not that "common sense alone" will result in a quality product. Rather, the agent is common sense supplemented by AGF and consensus-building, i.e. the edit-revert-discuss cycle. Moreover, your argument that common sense is not enough fails to answer my argument that over-elaboration makes the problem even worse as many editors ignore the MOS altogether. A concise, stable, seldom-disputed MOS would be much more likely to be consulted and respected, if you ask me. The endless revisions and arguments over minutiae only serve to brand it as unfinished and excessively pedantic, to non-professional writers.

It also seems ironic to me that you would propose on the one hand that "one or two gatekeepers" should act as guideline stewards while on the other hand you bemoan that "small groups of editors" could have undue influence. Guidelines should reflect community consensus, and I don't see how "one or two" improves on "small groups", unless you suppose that the one or two are better than the rest. That sort of elitist attitude doesn't really fit well in a wiki. Am I misunderstanding the role you have in mind for the gatekeepers?

I do respect your desire to make WP a quality product, and I share it, but I believe WP's open-door policies inevitably imply that this can only occur by improving, not avoiding, less than professional contributions. "Anyone can edit" is the antithesis of quality control, so there will always be new contributions in need of polishing. Thus, WP will never achieve the high lustre of the New York Times or a paper encyclopedia, at least where new content is concerned. I don't think you really grok that fact. Even so, if Wikipedia truly looked "utterly foolish", would it be one of the most-consulted web sites in the world? Surely not. People chuckle good-naturedly at its foibles, but that only underscores how much they expect from it. The exception proves the rule. Unconventional (talk) 04:36, 14 September 2009 (UTC)

I disagree with your notion that Maybe the reason MOSNUM is unstable is that it tries to be too comprehensive. The implied solution (make it less comprehensive for an all-volunteer force of amateur editors) would just make Wikipedia look like crap (bringing upon Wikipedia, even more of that attribute you light-heartedly dismiss as People chuckle good-naturedly at [Wikipedia’s] foibles). There is apparently too little common ground for me to further engage you on this issue. We will have to agree to disagree. Goodbye and happy editing. Greg L (talk) 04:54, 14 September 2009 (UTC)

As you wish, though I doubt we are so far apart as you seem to think (I very often agree with you on other topics). Thank you for your first, less dismissive reply. Unconventional (talk) 15:22, 14 September 2009 (UTC)

Summary of MOSNUM started

Following Tony1's proposal, I've started writing a very brief summary (my goal is approx. 25% of the full text) of WP:MOSNUM, intended to eventually replace sections from 10 to 13 of WP:MOS. Feel free to collaborate at User:A. di M./MOSNUM. --___A. di M. 21:05, 14 September 2009 (UTC)

You da man. Greg L (talk) 21:39, 14 September 2009 (UTC)

Request for edit of protected section; YYYY-MM-DD format

I propose that

YYYY-MM-DD style dates (1976-05-31) are uncommon in English prose, and should not be used within sentences. However, they may be useful in long lists and tables for conciseness. (If the only purpose why they are used in a particular table is ease of comparison, consider using {{sort|2008-11-01|1 November 2008}}.) Because some perceive dates in that style to be in conformance with the current ISO 8601 standard, that format should never be used for a date that is not in the (proleptic) Gregorian calendar, nor for any year outside the range 1583 through 9999.

be changed to (new text underlined)

Of the all-numeric date formats, only the YYYY-MM-DD style is reasonably unambiguous; other all-numeric styles should be avoided except in direct quotes. The YYYY-MM-DD style dates (1976-05-31) are uncommon in English prose, and should not be used within sentences. However, they may be useful in long lists and tables for conciseness. (If the only purpose why they are used in a particular table is ease of comparison, consider using {{sort|2008-11-01|1 November 2008}}.) Because some perceive dates in that style to be in conformance with the current ISO 8601 standard, that format should never be used for a date that is not in the (proleptic) Gregorian calendar, nor for any year outside the range 1583 through 9999.

because some editors don't get that listing the acceptable date style implies that other styles are discouraged.

Any objections? --Jc3s5h (talk) 03:27, 11 September 2009 (UTC)

I think it is an improvement. However, it might be worth noting that many users use ISO-format dates in references, too. Dabomb87 (talk) 03:33, 11 September 2009 (UTC)

I believe much of the use of YYYY-MM-DD format in citations is in citation templates, and the templates in turn were damaged by the misguided belief that date autoformatting would take care of all date format problems. I believe that most of that damage has been repaired, and most citation templates will accept whatever format the editor types in. Nevertheless, those who want that format in citatons can try to argue that the citations are a long list, so that format is OK. The change would not affect the success or failure of such an argument. --Jc3s5h (talk) 03:50, 11 September 2009 (UTC)

apparently it also needs to be stated explicity that in sentences, the month is to be spelled out, not given in numerical form. that's stated at WP:MONTH, but the misunderstanding above indicates that it's not clear enough that that also applies to full dates and date-month items. so maybe something like this:

Of the all-numeric date formats, only the YYYY-MM-DD style is reasonably unambiguous; other all-numeric styles should be avoided except in direct quotes. Within sentences, the month needs to be spelled out, but the YYYY-MM-DD format may be useful in long lists and tables for conciseness. (However, if the only reason for using this format in a table is ease of comparison, consider using {{sort|2008-11-01|1 November 2008}} instead.) Because some perceive dates in the YYYY-MM-DD format to be in conformance with the current ISO 8601 standard, that format should only be used for dates in the (proleptic) Gregorian calendar and in the year range 1583 through 9999.

(i don't understand what that parenthetic "consider using ... instead" is advising - DMY format? the use of that "sort" template? i've smoothed the syntax a bit, but someone else needs to make the point clearer, whatever it is.) Sssoul (talk) 06:49, 11 September 2009 (UTC)

I'm not happy with "reasonably", and can the text be more concise? Does this mean every use of ISO 8601 will have to be changed? There's an awful lot of it.

This involves MoS main page as well; it is a textbook example of why the pages need to be rationalised. Tony (talk) 07:18, 11 September 2009 (UTC)

I'd use:

Do not use date formats such as 09/11/2009, as they are ambiguous (it could refer to either 9 November or September 11).
When dates are used within sentence, spell out the month (11 September 2009 or September 11, 2009); in places where space is limited, such as in narrow table columns, infoboxes, and footnotes, use the YYYY-MM-DD format (2009-09-11), or abbreviate the month to three letters (11 Sep 2009 or Sep 11, 2009).
- Tables which do not use the YYYY-MM-DD format should use Template:Sort or Template:Dts so that the rows can be sorted correctly.
Because some perceive dates in the YYYY-MM-DD format to be in conformance with the current ISO 8601 standard, that format should never be used for a date that is not in the (proleptic) Gregorian calendar, nor for any year outside the range 1583 through 9999.

--___A. di M. 09:31, 11 September 2009 (UTC)

Extending the YYYY-MM-DD format to footnote references would be a change from the current guidance. I don't support it, for the same reason that it is not supported in the current guidance. If we have to spell the date out, that's fine, but (certainly for the US, where it is not common, as pointed out in the guidance) we should stick with the approach of the current guidance. Furthermore, footnotes certainly are not in table format, and in most wikipedia articles they are not long lists either (and I doubt that is what the draftspeople of the guidance meant by long lists). The archive history of the guidance is full of acknowledgements that this format is not approved for footnote references. Plus -- the issue was just discussed last month at [9], with many editors expressing hesitation about expanding its use because they perceived it as less user-friendly than spelling out the months (including hesitations about expanding its use to footnote references). --VMAsNYC (talk) 09:45, 11 September 2009 (UTC)
But the current guideline does not forbid YYYY-MM-DD in publication and access dates of references (as they are not "within sentences"); and in practice there are thousands of articles which use that (e.g. this one and this one which recently appeared as TFA). I don't see the point of going around and change them all (even if my personal preference would be for "11 Sep 2009", so that it'd not require inconsistency when you don't know the day of the month; now they use "2009-09-11" and "September 2009" which looks weird).--___A. di M. 09:59, 11 September 2009 (UTC)

"or abbreviate the month to three letters"—I think Greg will have something to say about that. Jun? Can we lose "in places"?
If by losing "in places" you just mean to remove those two words leaving the rest intact (i.e. "where space is limited"), I have no problem with that. For the other point, what about replacing "use ... or abbreviate" to "consider using ... or abbreviating"? That way, readers will understand that they needn't abbreviate "June" to "Jun" just for the sake of it. (OTOH it would still be useful in a table column in which all other months are abbreviated.) --___A. di M. 11:55, 11 September 2009 (UTC)

I think we should use fully spelled out months and human-readable formats when possible. If you look at Wikipedia:Featured lists, many of the articles listed there use full date formats without abbreviation, and aren't the space hogs some think they are. Dabomb87 (talk) 12:49, 11 September 2009 (UTC)
I'd say "when reasonable" rather than "when possible". When is it impossible? I guess it never is. There's nothing forbidding us to use full spelled-out dates in this table, but that'd be far more trouble than it's worth: for example, that'd mean than on all but the wider displays each row of the table would take two lines, for example. --___A. di M. 13:04, 11 September 2009 (UTC)

Reply to A. di M.--No disrespect meant here, but what's funny is that when I just made an argument elsewhere based on the fact that there were a plethora of articles that were in accord with my view, another editor terseley responded "I don't care that WP:OTHERCRAPEXISTS". —Preceding unsigned comment added by VMAsNYC (talk • contribs) 11:22, 11 September 2009 (UTC)

When "other crap" exists in a substantial fraction of all Wikipedia articles (probably, in about half of all articles including at least one full date in the "References" section), that shouldn't be dismissed that quickly, I think. --___A. di M. 11:55, 11 September 2009 (UTC)

I did a few searches to find out how often the (I now understand out of favor, though it certainly was not clear to me from the guidance) MM/DD/YY format appears in Wiki articles. Impressive as well.--VMAsNYC (talk) 12:26, 11 September 2009 (UTC)

(outdent) back to A di M's proposed wording: i find the double negatives in "that format should never be used for a date that is not in the (proleptic) Gregorian calendar, nor for any year outside the range 1583 through 9999" unnecessarily convoluted/confusing. can we try "that format should only be used for dates in the (proleptic) Gregorian calendar and in the year range 1583 through 9999" instead? Sssoul (talk) 11:43, 11 September 2009 (UTC)

Sounds all right to me. --___A. di M. 11:55, 11 September 2009 (UTC)

Very minor drafting point -- does that mean inclusive?--VMAsNYC (talk) 12:29, 11 September 2009 (UTC)

Revised proposal

{editprotected}} Taking the previous points into account and shamelessly stealing good ideas, I propose:

* YYYY-MM-DD style dates (1976-05-31) are uncommon in English prose, and should not be used within sentences. However, they may be useful in long lists and tables for conciseness. (If the only purpose why they are used in a particular table is ease of comparison, consider using {{sort|2008-11-01|1 November 2008}}.) Because some perceive dates in that style to be in conformance with the current ISO 8601 standard, that format should never be used for a date that is not in the (proleptic) Gregorian calendar, nor for any year outside the range 1583 through 9999.

be changed to:

Do not use ambiguous date formats such as 09/12/2009 (9 December or September 12), even when the day number is greater than 12.

When dates are used in running prose, spell out the month (12 September 2009, September 12, 2009); where space is limited, such as in narrow table columns, infoboxes, and footnotes, use YYYY-MM-DD (2009-09-12) or abbreviate the month (12 Sep 2009 or Sep 12, 2009). Tables that do not use YYYY-MM-DD should use Template:Sort or Template:Dts so the rows can be sorted correctly.

The YYYY-MM-DD format should be used only for dates in the (proleptic) Gregorian calendar and in the year range 1583 through 9999, because YYYY-MM-DD is regarded by some readers as conforming with ISO 8601.

In reply to VMAsNYC, I have always understood the word "through" to mean inclusive; do you have reason to believe otherwise? --Jc3s5h (talk) 13:46, 11 September 2009 (UTC) Revised 14:31 UT in response to Sssoul.

Hmmm. I only made the comment because I've seen phraseology along the lines of "1-10, inclusive." But I can't recall if I've seen exclusive ... and for the life of me I can't seem to google the rule. Perhaps in those instances that I recall the author was paid by the word ... Merriam Webster indicates one meaning of through is to and including, which matches what you are thinking.--VMAsNYC (talk) 15:49, 11 September 2009 (UTC)

looks close! ~~i think that first bullet point needs to be phrased differently though, because some literal-minded types will decide it means they can happily use 23/11/2009, since it's not ambiguous. how about:~~

~~Do not use all-numeric MM-DD or DD-MM date formats; in sentences, spell out the month in full: September 11, 2009 or 11 September 2009~~
~~Where space is limited, such as in narrow table columns, ... [and so on]~~ fixed now, thanks ...

Sssoul (talk) 14:10, 11 September 2009 (UTC)

I have changed the proposal in response to Sssoul. I didn't use her exact wording because I have observed that in talk page discussions, anything with an "M" in it may be interpreted as a placeholder for a numeric month, abbreviated month, or spelled-out month. YYYY-MM-DD should be OK since it is accompanied by an example. --Jc3s5h (talk) 14:34, 11 September 2009 (UTC)

Minor point: "proleptic Gregorian calendar" refers to the extension of Gregorian calendar to cover dates before 15 October 1582, so in this case "(proleptic)" can be safely omitted. (BTW, I'd replace "use ... or abbreviate" with "consider using ... or abbreviating", to show that we aren't forbidding to write "4 June 1987" in an infobox containing no other full date just for the sake of it.) --___A. di M. 14:57, 11 September 2009 (UTC)

another minor point, but using a different date in the example would probably be good - 12 September, for example Sssoul (talk) 15:40, 11 September 2009 (UTC)

Agree on not using 9/11 because of the idiomatic order for that particular date. As for proleptic Gregorian calendar, I would think it could be called that if using the Gregorian calendar for a date after 1583 but before a country strongly connected to the article adopted the Gregorian calendar. --Jc3s5h (talk) 17:24, 11 September 2009 (UTC)

So where is the exact text that is proposed, please? Tony (talk) 16:04, 12 September 2009 (UTC)

I'm not sure we really need the PER template here, as the page is only semi-protected and all of the people here seem to be autoconfirmed, unless I'm missing something. -- Soap ^Talk/_{Contributions} 16:24, 12 September 2009 (UTC)

Ah, yes. It's a transcluded section within a larger page. The section is full-protected; the page isn't. Annoying but necessary, I suppose. -- Soap ^Talk/_{Contributions} 16:39, 12 September 2009 (UTC)

Comment. The old and the new proposal both suggest templates like {{Sort}} and {{Dts}}, but these templates generate output that has serious WP:ACCESSIBILITY problems. For example, "{{dts|1776|July|4}}" generates "July 4, 1776" which a screen reader will read aloud as something like "zero one seven seven six hyphen zero seven hyphen zero four July four, seventeen seventy-six". (A text browser such as Lynx will render it as "01776-07-04 July 4, 1776".) The MOS shouldn't recommend practices that have accessibility problems. Please remove the bullet "Tables which do not use the YYYY-MM-DD format should use Template:Sort or Template:Dts so that the rows can be sorted correctly." from the draft text. Eubulides (talk) 16:23, 12 September 2009 (UTC)

If the templates don't work, I suggest you nominate for deletion. After they are deleted, come back here and get this guideline revised. --Jc3s5h (talk) 16:49, 12 September 2009 (UTC)
Just fixed Dts.[10] I would fix Sort, too, but it's protected. --___A. di M. 17:44, 12 September 2009 (UTC)

Alas, that fix doesn't work, at least not for Lynx. It renders the output of "{{dts|1776|July|4}}" as "01776-07-04 July 4, 1776". I'll follow up at Template talk:Dts. The conversation here has quickly turned into the usual format flamefest and I doubt whether much progress can be made here right now. Eubulides (talk) 07:01, 14 September 2009 (UTC)

(outdent) not taking any stand on the templates, just responding to Tony's question: so far the proposed wording is:

Do not use date formats such as 09/12/2009, as they are ambiguous (it could refer to either 9 December or September 12). For consistency, do not use such formats even when the day number is greater than 12.
When dates are used within sentence, spell out the month (12 September 2009 or September 12, 2009); where space is limited, such as in narrow table columns, infoboxes, and footnotes, use the YYYY-MM-DD format (2009-09-12), or abbreviate the month to three letters (12 Sep 2009 or Sep 12, 2009).
- Tables which do not use the YYYY-MM-DD format should use Template:Sort or Template:Dts so that the rows can be sorted correctly.
Because some perceive dates in the YYYY-MM-DD format to be in conformance with the current ISO 8601 standard, that format should only be used for dates in the (proleptic) Gregorian calendar and in the year range 1583 through 9999.

Sssoul (talk) 16:55, 12 September 2009 (UTC)

Trimmed and tweaked version:

Do not use ambiguous date formats such as 09/12/2009 (9 December or September 12), even when the day number is greater than 12.
When dates are used in running prose, spell out the month (12 September 2009, September 12, 2009); where space is limited, such as in narrow table columns, infoboxes, and footnotes, use YYYY-MM-DD (2009-09-12) or abbreviate the month (12 Sep 2009 or Sep 12, 2009). Tables that do not use YYYY-MM-DD should use Template:Sort or Template:Dts so the rows can be sorted correctly.
The YYYY-MM-DD format should be used only for dates in the (proleptic) Gregorian calendar and in the year range 1583–9999, because YYYY-MM-DD is regarded by some editors as conforming with ISO 8601. Tony (talk) 02:17, 13 September 2009 (UTC)

No one is going to get any lip outa me on what Tony wrote with his first bullet point, since it appears to be pretty much what I’ve been doing all along in order to communicate with maximal fluidity and minimal confusion. However, I’d punch up the first two bullet points a notch. I would propose folding in the following:

In regular body text, months should generally be spelled out so they read more naturally and fluidly and there is no ambiguity.
All-numeric dates should generally be limited to tables and other sections that are truly space-constrained and there is no room for spelled-out months.
Where possible, the header to tables and other sections that contain all-numeric dates should include a legend denoting the formatting order; e.g.: (day-month-year) or (month-day-year), or (year-month-day).
Where possible in space-constrained sections, abbreviations for months should be considered as an alternative to all-numeric dates, When doing so, editors should use the progression used by the Associated Press in their Manual of Style, which is as follows:

Jan., Feb., March, April, May, June, July, Aug., Sept., Oct., Nov., Dec.

Or, for especially tight tables or places where alignment between rows is important, a rigid three-letter abbreviation series may be preferable, as follows:

Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec.

Greg L (talk) 02:44, 13 September 2009 (UTC)

I'm not with you on the AP progression Greg, there are other just-as-valid ones, such as Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec. Headbomb {^ταλκ_{κοντριβς} – WP Physics} 03:37, 13 September 2009 (UTC)

Sure there are others. Note, however, that 99% of the English-language newspapers that subscribe to the Associated Press also observe the AP Manual of Style. Thus, that method is exceedingly familiar and natural to readers for whom English is their first language. What you propose would be an alternative for use in tables where space is at a extreme premium. Greg L (talk) 03:48, 13 September 2009 (UTC)
P.S. I folded your suggestion into my proposal. Greg L (talk) 04:06, 13 September 2009 (UTC)

Just say "use abbreviations", I don't see the need to specify the style. If someone adds "Nvmbr" for November, or "8-ber" for October, someone will turn it into something less abhorrent in no time. That being said, I only know of these two styles, so if that's all there is, we might as well add list them. Headbomb {^ταλκ_{κοντριβς} – WP Physics} 06:16, 13 September 2009 (UTC)

“Nvmbr”? Manuals of style exist for publications like the Associated Press, which have all-professional writing staffs with journalism degrees. A manual of style for a publication that *anyone* can contribute to needs to be every bit as thorough as any of the other manuals of style. Yeah, include the recommended list of abbreviations so there is that much less garbage that experienced editors must clean up. Greg L (talk) 20:39, 13 September 2009 (UTC)

All-numeric formats other than YYYY-MM-DD have no place in an international publication except within direct quotes. Not in running text, not in tables, not in citations, not nowhere,n not no how. Anything seen anywhere will be imitated by those who do not read the MOS or the MOSNUM. It will spread to places where no one can decipher what it means. This isn't a style issue, this is an issue of damaging information. --Jc3s5h (talk) 03:22, 13 September 2009 (UTC)

Wikipedia is first and foremost, and English-language publication. The ISO 8601 standard has absolutely nothing to do with manuals of styles, which are adhered to in order to make articles read most naturally. Manuals of style simply follow the practices that are most familiar to the intended readership. If the article is about a general-interest topic, then the date formats most familiar to readers for whom English is their first language is what must be used. Thus, if a table is going into an article on the Boston Red Sox, and if there is truly no room for spelled-out dates—or even abbreviated months—then the date format most familiar to American baseball enthusiasts is most appropriate. Again, we follow the way the world really works and can’t pretend to be *leading by example*. Greg L (talk) 03:35, 13 September 2009 (UTC)

Greg, I saw what you said about legends. My belief is that editors will imitate the numberic formats other than YYYY-MM-DD, but will not imitate the practice of providing a legend to explain what the format means. If the editors disappear and used sources that are expensive to access, it will be very inconvenient to fix the damage. The prohibition of all-numberic non-YYYY-MM-DD date formats is non-negotiable; the whole point of my proposal is to prohibit them and I will not support any change that fail to prohibit them. --Jc3s5h (talk) 16:03, 13 September 2009 (UTC)

BTW, quoting you, Jc3s5h: It [all-numeric formats other than YYYY-MM-DD] will spread to places where no one can decipher what it means. Did you even read what I proposed above? If that wording was adopted, there would be zero confusion because all-numeric dates would rarely be used, and when they are used, there’d be a parenthetical legend explaining the order. Zero confusion. Greg L (talk) 03:42, 13 September 2009 (UTC)

Regarding Jc3s5h's recent edit summary "readers, not editors", I think that more than 95% of all readers will have no idea that an ISO standard about representation of dates and times, let alone expect that dates conform to it. It is some editors (IIRC, User:Gerry Ashton was one of them) who don't like using YYYY-MM-DD for non-Gregorian dates because of what the ISO standard says, even if not everybody would agree that said standard applies to dates displayed in the articles for humans to read them. --___A. di M. 16:35, 13 September 2009 (UTC)

I don't know about readers who do not edit, since I have no interaction with them. From what I've seen with postings all over the English Wikipedia, whenever they mention the YYYY-MM-DD format, they are apt to refer to it as an ISO date, even though they often seem unaware of the exact name or full requirements of the standard. Even the Chicago Manual of Style mentions ISO in connection with this format. Considering how may of our readers seem to be deeply interested in computers, I think that a significant number of those who might adopt that format themselves (at least for some purposes) think there is an ISO standard associated with the YYYY-MM-DD date. Therefore we should not expose them to instances where Wikipedia dates violate the standard. If we had a context where it was made clear to readers that we disavow the ISO standard, and the calendar must be determined from context. that would be fine, but we do not have such a context. --Jc3s5h (talk) 16:48, 13 September 2009 (UTC)

←(outdent)

Arguing over the virtues of this all-numeric date format versus that all-numeric date format is inane. Over the years, much to-do has been made over this silly issue. Editors have wasted literally megabytes of WT:MOSNUM file space to this issue. Editors absolutely convinced that “[insert favorite style here] is good and unambiguous and universally understood by all God’s creatures and should be mandated on en.Wikipedia” have come here beating around the bush, trying to use every argument in the book to promote their favorite practice that they believe is *universally right for all.*
The simple fact is that an article about some relatively small city in England or Australia would be better off having its written-out dates formatted in the way that is most common and natural for readers in their respective countries. In just the same way, an article on Coeur d'Alene, Idaho should rightly use the date format most well-recognized by the disproportionate number of American readers: “November 28, 2007” and “11-28-2007”. General-interest articles are read by a general readership and whatever is most natural to them is what should generally be used on Wikipedia.
Arguments that some alphabet-soup standards organization has proposed “this ‘n’ that” don’t matter one twit here. The simple fact is that ISO 8601 was never intended for general encyclopedic, magazine, or newspaper use and it is profoundly naive to think that newspapers will soon start writing “A man was arrested in a residential burglary at 2009-09-12T19:30:00.” General-interest readers don’t write dates and times like this in real life and it is seldom indeed they are ever exposed to it in daily print of any sort. I’ve had arguments thrown my way in above threads on this page that boil down to nothing more than “well… any confusion will be short-lived and they’ll soon ‘get it’ ” and this simply proves in my mind that these editors don’t have a flying clue what technical writing is all about.
Again… it doesn’t ever matter if there is some standard out there that seems superior. The IEC had a proposal to replace “256 kilobytes” with “256 kibibytes” but we don’t simply adopt some new standard because it is good and holy; we simply follow the way the real world works. That’s just too much common sense. All encyclopedias endeavor to write with minimal confusion and in a manner that causes the fewest interruptions in the train of thought. Introducing unfamiliar ways to write out dates and/or times is the last thing any publication should do.
Now, the simple fact is that en.Wikipedia finds itself in a unique situation because it is electronic and at the top of the heap in the breadth of its content amongst the other-language Wikipedias. Further, because of the inexorable adoption of English as a second language, many readers for whom English is their second language come here to polish up on their knowledge of a given subject as well as their English skills. And, given that Wikipedia is free for anyone to contribute to, there are many would-be authors who have no knowledge whatsoever about technical writing or the customs and practices of other peoples. Accordingly, having a manual of style is the only thing that prevents Wikipedia from degenerating into total crap.
I, for one, am just disgusted by the amount of bickering that has transpired over the years here on dates. My stopping here and reading what some people have written above makes increasingly makes me think it is exceedingly unwise of me to make the effort here. I see a clear tendency in some of the arguments in this thread for editors to assume (or wish) that “[insert favorite numeric format here] is soooooo well recognized it will cause little or no confusion.” Horse hooie. Readers from all walks of life on this globe have been exposed largely to one way of writing all-numeric dates in only one particular way and will quite often be confused when confronted with an unfamiliar format. Now…
My above proposal makes it exceedingly clear that it should be a rare date indeed that is all-numeric. I don’t think they have a place in footnotes (under the pretense there isn’t enough room down there). Of course there is; footnotes are an unconstrained space that grows as you add. My proposal makes it entirely clear that all-numeric dates should be used only where space is at an extreme premium and that abbreviated months should be considered as an alternative to all-numeric dates where possible to minimize confusion. Finally, my proposal introduced some sort of *profound* concept that should have been adopted here long ago: that if editors are going to have a table with all-numeric dates, that header of the table include a parenthetical legend disclosing the format used; e.g., Date inducted into the Hall of Fame (month‑day‑year).
If editors here are still laboring under the misperception that they are going to somehow prevail and convince everyone else that there is some single way to format dates that is universally good for all God’s Internet creatures and Wikipedia ought to exclusively adopt that one single practice, well… wake up and smell the coffee. There are multiple ways of doing so and we should be focusing on crafting guidelines in our manual of style that ensure the most appropriate practices are used for a given article and confusion is minimized to the greatest extent possible. Read my 02:44, 13 September 2009 proposal, above, to see just how amazingly simple that can be. Greg L (talk) 20:09, 13 September 2009 (UTC)

You are right in much of what you say, but any all-numeric date other than with the year first (2009-09-13) is absolutely unacceptable in any article. In fact, I wish we could just ban all-numeric dates altogether, since, as you say, nobody uses them in "real life". But if they are going to be permitted, even an article about Coeur d'Alene, Idaho must not be allowed to use 5-3-2009 because you cannot assume, just because it is an article about an American town, that its only readers will be American. People will simply not be sure whether it means 3 May or 5 March. Such ambiguity cannot be tolerated. (There is no problem about different ways of writing dates when the month is written out as a word, which is a separate issue, because there is no ambiguity there.) -- Alarics (talk) 20:44, 13 September 2009 (UTC)

Quoting you: [anything other than something like 2009-09-13 is] absolutely unacceptable in any article. Well, your absolutism is noted, but you need to get realistic. A “one shoe fits all”-approach isn’t in the cards and what you desire simply won’t ever happen.
Moreover, I really wish people would read my proposal and understand what it actually says, rather than assume it is saying something that it does not. It makes these seven, simple points:

Write out the date to minimize confusion.
Write out the date to minimize confusion.
Write out the date to minimize confusion.
Write out the date to minimize confusion.
Write out the date to minimize confusion.
If you really, genuinely have a high-space-constrained section like a table, try abbreviating the months to minimize confusion.
If you have so darn little space in a table that you must use numeric-only (this should be rare indeed on Wikipedia), include a parenthetical legend in the table’s header to minimize confusion.

Did you notice the “minimize confusion” emphasis yet? It’s a better approach than these highly unrealistic “adopt what some alphabet-soup standards organization proposed for use on airline tickets even though it isn’t universally used in real-life, general-interest publications”-approaches, which only unnecessarily increase confusion. Greg L (talk) 21:01, 13 September 2009 (UTC)

BTW, I note that some I.P. editor edited ISO 8601 to strip out a desperately-needed citation to the ISO themselves (I know; shocking) and introduced a huge swath of un-cited fairy tale. I wonder if an editor here is logging out to do this? A Check User on that I.P. address would be interesting, wouldn’t it? It’s crazy crap like this that highlights how attractive is the practice of the German-version of Wikipedia of locking down all articles and requiring a gate keeper’s approval for edits. I added some {citation needed} tags to the fairly tale stuff. If this mysterious I.P. editor can’t back up what he/she just added with clear citations that speak precisely to the point, it gets tossed. Wikipedia isn’t a forum for POV-pushing and fabrication. Greg L (talk) 21:34, 13 September 2009 (UTC)

I don't know where you got the idea that I am backing anything put out by a standards organisation. I hold no brief for ISO or any such thing, and in a different discussion today (about delimitation of numerals) I have actually agreed with you about that. No, I simply point out that the 3-5-2009 format should never be used anywhere because it is utterly ambiguous and therefore useless. If all-numeric dates have to be used at all (and I wish people wouldn't), the "computer date" one with the year first (2009-09-13) is the only one that can be accepted, not because some organisation has come up with it but simply because it is the only all-numeric date that is not ambiguous. It's not enough merely to "minimise confusion": we have to positively prevent ambiguity. If you read again what I wrote, you will see that I am certainly not trying to push a "one size fits all" policy here, just mentioning one particular format that simply won't do. I don't, incidentally, think people do use it much in Wikipedia, probably because they realise that it's ambiguous in an international setting. I only mentioned it because you included it as one of the styles that might be acceptable at a pinch. It isn't. -- Alarics (talk) 22:11, 13 September 2009 (UTC)

You have strong convictions about what is and isn’t clear to readerships but you seem to lack an understanding how how things work for American users—probably too for readers in England, Australia, Singapore, Canada, and Europe as well. The date 9-13-2009 (today’s date), or 5-11-2006 for that matter, is understood perfectly well by Americans and would be suitable for an article that is closely associated with the U.S.
I also agree with you, however, that these all-numeric date formats are generally unsuitable for use on Wikipedia. The use of all-numeric dates of any sort is fraught with the potential for confusion because the American-style way of numeric dates is just that: the American way of doing it and is not observed in many other English-speaking countries.
This is why what I proposed above couldn’t possibly make it any clearer that all-numeric dates should be looked upon with great disfavor and used only where absolutely necessary (which, I suspect should be nearly never since work-arounds aren’t nearly as elusive as some editors might like to admit). Moreover, where the darned all-numeric dates are used, what I proposed calls for including a legend to always make it perfectly clear what order the temporal data is organized; e.g. Date inducted into the Hall of Fame (month‑day‑year).
To suggest that American audiences would be any less confused by 2009/9/13 than 9/13/2009 because the former is “just sooooo much clearer” seems to be founded upon wishful thinking rather than an understanding of the real-world practices of various peoples.
To those editors here who pretend to promote any one particular custom or practice as being The One and Only True Way™©® and ought to therefore be *universally* adopted on Wikipedia: I hope you would soon realize such efforts are invariably doomed for failure. You might like things to be simple, but, if you dig down into the details, things are a bit more complicated than you might like. Greg L (talk) 23:29, 13 September 2009 (UTC)

I see no reason whatsoever to display any form of all-numeric date at all on WP. Eliminate ISO dates altogether. Where space is tight use three-letter month abbreviations. Sorry Greg but I wouldn't even bother with the Associated Press format. Space is either limited or not, if it is, use three-letter abbreviations, if not, spell the month out.

Dates are given with the month fully spelt out (12 September 2009, September 12, 2009) or, where space is limited, abbreviated to three letters (12 Sep 2009, Sep 12, 2009).
Do not use ambiguous date formats such as 09/12/2009 (9 December or September 12).

Do not use ISO (YYYY-MM-DD) format. Template:dts produces hidden ISO dates for automatic sorting tables.

JIMp_talk·cont 14:08, 14 September 2009 (UTC)

"The date 9-13-2009 (today's date), or 5-11-2006 for that matter, is understood perfectly well by Americans and would be suitable for an article that is closely associated with the U.S." (Greg L)

No, it wouldn't, unless you suppose that only Americans are ever going to read an article about the USA, which is an absurd idea. Of course Americans will not be confused by such a date in an American newspaper. Even a few Americans (those who have discovered that the rest of the world doesn't follow their illogical order) might now have doubts if they come upon it in an international context such as Wikipedia. But more to the point, Brits and others reading it are simply not going to know whether it means 5 November (which is what they are likely to assume) or 11 May (which they might suspect if they have become aware, probably as a result of "9/11", that in the USA they do these things oddly differently). User:Jimp's proposal has my vote. -- Alarics (talk) 14:32, 14 September 2009 (UTC)

Wikipedia editors should acknowledge what the ISO has too much hubris to acknowledge: it is beyond our ability to communicate to our editors and readers that there are particular rules associated with the YYYY-MM-DD date format whenever it is stated or implied that the format is governed by the ISO 8601 format. Even an experienced editor like Jimp does not get it. If the YYYY-MM-DD format is to be forbidden, there should be no mention of ISO. --Jc3s5h (talk) 15:36, 14 September 2009 (UTC)

Jimp's concept makes sense to me (it's clear and easy to explain and apply), but i think we need to keep "even when the day number is greater than 12" for the literal-minded out there who might think it means "don't use all-numeric formats when the day is 12 or under". how about:

Dates are given with the month fully spelt out (12 September 2009, September 12, 2009) or, where space is limited, abbreviated to three letters (12 Sep 2009, Sep 12, 2009).
- Do not use numbers to represent months, as it can be ambiguous (09/12/2009 could be 9 December or September 12); even when the day is greater than 12, spelling out the month minimizes confusion.
- Do not use YYYY-MM-DD format. ~~(When ISO dates are needed for automatic sorting tables, Template:dts produces hidden ISO dates.)~~ (When automatic table sorting is needed, Template:dts produces hidden YYYY-MM-DD dates along with visible dates in an acceptable format.)

something like that? Sssoul (talk) 16:03, 14 September 2009 (UTC)

I am neutral about eliminating visible all-numeric dates altogether, but if it is to be done I would write something like

Do not use YYYY-MM-DD format. (When automatic sorting table sorting is needed, Template:dts produces hidden YYYY-MM-DD dates together with visible dates in an acceptable format.)

The use of "ISO" implies that Wikipedia has formally accepted that standard to govern our use of all-numeric dates, and we have done no such thing. Also, the mention of ISO dates without mention of the YYYY-MM-DD format invites the YYYYMMDD format, which is also a valid ISO 8601 date format. --Jc3s5h (talk) 16:23, 14 September 2009 (UTC)

I support Sssoul's revised proposal. JIMp_talk·cont 17:27, 14 September 2009 (UTC)

I doubt if one in a million people gives a toss whether it is called ISO or not, but anyway I support Sssoul's revised proposal. -- Alarics (talk) 18:06, 14 September 2009 (UTC)

Now I’m seeing some sensible progress and consensus-building. The whole point is to be unambiguous in technical writing. So lose the all-numeric dates entirely. I truly can not imagine a spot in an all-electronic, adjustable-everything encyclopedia that can’t be made to fit either spelled-out months or abbreviations. Greg L (talk) 20:37, 14 September 2009 (UTC)

Second pink-div

P.S. There is no need to be so proscriptive and micromanage how months may be abbreviated. In tables, three-letter abbreviations might certainly be best. But for other uses, the AP method is a natural-looking technique that English-speaking peoples are frequently exposed to in real life and there is no reason that technique can’t also be used here; ergo…

Months in dates. Spell out: 12 September 2009, September 12, 2009.
All-numeric date formats. Do not use (9/12/2009 can mean either "September 12" or "9 December").
YYYY-MM-DD numerical format. Do not use, except as hidden input code in templates such as {{dts}} and {{Birth date}}, which chronologically sort tables.
Limited space. Consider using either short-form (Sept. 6, 2009) or three-letter abbreviated months (Sep 6, 2009).
- Short-form. Use the series in the AP Stylebook, including the accompanying periods (.):
  Jan. - Feb. - March - April - May - June - July - Aug. - Sept. - Oct. - Nov. - Dec.
- Abbreviated months. For fixed-width columns in tables and similar contexts, use the three-letter series:
  Jan - Feb - Mar - Apr - May - Jun - Jul - Aug - Sep - Oct - Nov - Dec
Line-end word wrap. Best practice is to avoid it by using a non-breaking space between month and day (6 May or May 6).

This version last updated 17:46, 16 September 2009 (UTC)

Greg L (talk) 21:17, 14 September 2009 (UTC)

I would be happy to go along with that. Personally, I'd lose the AP version of the month abbreviations and just go with the three-letter ones, for simplicity, but I wouldn't go to the stake over it. I think Jc3s5h will complain about the mention of ISO. I couldn't care less one way or the other, but if it keeps him or her happy, why not just say "Do not use YYYY-MM-DD format ... " -- Alarics (talk) 21:38, 14 September 2009 (UTC)

Good point. I never heard of ISO 8601 until the last few days here here on WT:MOSNUM. Fixed. Thanks. Greg L (talk) 21:46, 14 September 2009 (UTC)

There may be some benefit in addressing Pigsonthewing's concern by indicating that all-numeric dates ought not be visible to readers in articles, and that template input is outside the scope of MOS. Perhaps another second-level bullet
- This guideline applies to dates that are visible to readers; template inputs and the like are dealt with elsewhere.

Also, I think just about everyone agrees that the 9/12/2009 format with no legend or explanation in close proximity is bad, but not everyone agrees that 2009-09-12 is ambiguous. Certainly the latter format has much wider use in Wikipedia than the former. If we are going to recommend avoiding all visible numeric formats, it may be appropriate to advertise this discussion to a wider audience. --Jc3s5h (talk) 22:55, 14 September 2009 (UTC)

When you refer to templates there, can we clarify that you don't mean citation templates? Alarics (talk) 23:20, 14 September 2009 (UTC)

I revised the above pinkbox. I think that addresses the concern of Andy Mabbett (Pigsonthewing), per suggestion of Jc3s5h. I think I also addressed the concern of Alarics. Together, the edits should have clarified that all-numeric input is—of course—entirely fine as template code so long as the resultant output does not appear to the reader in all-numeric form. Thanks. Greg L (talk) 00:05, 15 September 2009 (UTC)

It's longer than one would want, but I support. Tony (talk) 01:57, 15 September 2009 (UTC)

Well… does anyone really think that last bullet point…

• Do not use YYYY-MM-DD format except where required as hidden input code in templates such as {{dts}} and {{Birth date}}, which chronologically sort tables.

…is truly necessary? It would seem that the first bullet point ought to get the point across well enough. Greg L (talk) 02:12, 15 September 2009 (UTC)

P.S. I’m gonna lose that last bullet point. If someone has to have it back, restore it and update the “Version” date with five tildes. (I just made the change with this edit, which synchs the “Version” time-stamp with this autosignature). Greg L (talk) 02:19, 15 September 2009 (UTC)

Yes, I do think we need to be as explicit as possible that we don't want YYYY-MM-DD numerical format, since that is the one that is everywhere on WP at the moment. Some people might not see that as being covered by the injunction not to use all-numeric dates. So I have restored it. (BTW, a minor point, I have changed your curly quotes to straight quotes, as I believe curly quotes are now deprecated?) Also, in the interests of concision I removed the phrase "primarily in the US" (in relation to one of the versions of the all-numeric dates) as I don't think it's really needed. But put it back if you insist. -- Alarics (talk) 08:33, 15 September 2009 (UTC)

If we want to make it shorter, I personally don't really think we need the two indented sub-bullets about short forms. I think it is enough to say "* Where space is limited, consider using either short-form or abbreviated months; e.g., Sept. 6, 2009, 6 Sep 2009." and leave it at that. I know there are concerns about people writing "Nvmbr" and so on, but I don't think I have ever come across it myself. It seems a rather minor problem to devote so much attention to. There isn't in fact usually a space problem so it will not usually come into play, and making so much of a meal of it seems to detract from the main message we want to get across, which is to write dates out in full. But I am not going to the stake over this.

I also personally would lose the bit about putting non-breaking spaces in dates. It's a counsel of perfection which I seriously doubt many people can be bothered to do. I bow to no-one in my pernicketyness but even I can rarely be bothered to do this when writing text. If there were a key on every keyboard for   it might be a different matter. But if people insist on keeping this instruction in, in a spirit of compromise I will not press the point to a dispute. I just think nobody will take any notice of it. -- Alarics (talk) 08:47, 15 September 2009 (UTC)

i agree with Alarics on both these points - as GregL himself said, we shouldn't try to micromanage this, and a] there's no need to give full lists of preferred abbreviations, and b] the nonbreaking-space bit is overcomplicating things. Sssoul (talk) 20:38, 15 September 2009 (UTC)

Were there some other short-forms and abbreviations that legitimately belong here, Sssoul? If so, please provide examples of said alternatives. If you know of no others, and we don’t show editors how to properly do this in an encyclopedia, we will soon all be exposed to a wide variety of novel ways. If your response is along the lines of “let editors use common sense”, then my response will be to ask “just how wide is your social circle there are Harvard where all this ‘common sense’ abounds?” As I mentioned in several places here now, Wikipedia is an encyclopedia to which *anyone* may contribute and for many of these editors, MOS and MOSNUM will be the only manual of style to which they have ever been exposed.
Besides, Alarics compromised and dropped his objection to providing example abbreviations. The flip side? Others—including me—compromised as to the number of bullet points in the above wording because we restored an explicit bullet point proscribing YYYY/MM/DD, which Alarics felt was important. Greg L (talk) 22:03, 15 September 2009 (UTC)

I didn’t know that YYYY/MM/DD was propagating all over Wikipedia; it’s “IEC prefix-time” all over again, where a small group of editors widely and quickly spread unwise practices. There are seventeen “Binary” archives, above, devoted to reversing that one, single practice, which got started with just a couple of highly animated editors. And it took three years to reverse it. Editors sometimes think early adoption of “standards” is a good thing and don’t realize it is a bad thing. We follow familiar and natural practices rather than try to lead by example because some alphabet-soup-named standards organization has a standard for airline tickets and they think the standard is so logical that everyone should adopt it (so Wikipedia will start the ball rolling). So…
By all means, I have no problem with your adding that bullet point back. I was trying to address a concern of others that we need our guidelines to be as succinct as possible. But you raise an important point and it is probably better to be clearer.
We’ve got a single-purpose I.P. editor, which is a Verizon account tracing to Herndon Virginia, who has been putting un-cited fantasy imaginations in our ISO 8601 article that states that IS 8601 “applies to all written communications … [even] hand written … when communicating internationally, nationally, locally, internally, or even privately.” Of course, the ISO themselves say no such thing and their own description of the standard makes it quite clear what the standard is all about.
I recognize that “Herndon VA” I.P. location from before (I can’t quite remember exactly remember the circumstances) and I suspect this individual is a registered editor who logs out for the express purpose of intentionally POV-pushing with un-cited fantasy material. So…
Indeed, debate surrounding how to unambiguously communicate dates sure brings out odd behavior here on Wikipedia.
As for brevity and stripping out things like how to abbreviate months, it’s really important to remember that a lot of would-be editors don’t have any idea as to what is the proper way to do things. WP:MOS and WP:MOSNUM are, for very many volunteer editors, the only manual of style to which they have ever been exposed. The above information takes up very little room, is to-the-point, and will help to improve the readability of Wikipedia and greatly minimize confusion—at least to the extent editors will follow it. One thing is for certain, if we don’t have the information, it is quite difficult for inexperienced editors to do the right thing. Having explicit mention that the YYYY/MM/DD format is now proscribed is important to you; providing clear guidance on how one properly makes abbreviated and short-form months is important to me. Greg L (talk) 19:04, 15 September 2009 (UTC)

Jolly good show. Can we now proceed to incorporate the text as agreed? I very much look forward to being able to point to an authoritative page that says numerical dates of any sort are strongly deprecated. Alarics (talk) 19:49, 15 September 2009 (UTC)

Indeed. I’m happy. This is where the tough part starts. It seems that any reasonable interpretation of the above thread shows a general consensus. The trouble is when there are really strong convictions on one side of a contentious issue. Wikipedia’s membership into the United Federation of Planets would be completely cut off at the knees by, you know… writing dates with months spelled out as words. The result of just posting it may be met with hyperbole and vitriol and RfCs and editwarring and ANIs and WQAs and members tossed out into a central courtyard to be met by an ArbCom firing squad. Still, I’m game since what appears above is pretty much what any English-language newspaper, magazine, and encyclopedia does. We should have had this sooner. Greg L (talk) 20:25, 15 September 2009 (UTC)

←(outdent)

I propose we move forward this way: Tony is a long-time editor on WT:MOSNUM and his opinion is widely respected. Moreover, he has been least active on this thread and this issue. Were you, Alarics, or I to post the above pink‑div verbiage to MOSNUM, the guideline would be prejudiced by the fact that we were most active on this thread and, therefore, the perception of outsiders might be that you and I were overly biased. So…
I will contact Tony on his talk page, direct him to this thread, and ask him to consider this and the related threads and advise as to whether he believes a general consensus exists on this issue. If so (fireworks be damned) we move forward and post. Why? Because we all did a pretty good job, IMO, of consensus-building here and the resulting product is straight out of Technical Writing 101. I think this approach at finalizing it all is the right thing to do. Greg L (talk) 20:59, 15 September 2009 (UTC)

Righty-ho. Alarics (talk) 21:07, 15 September 2009 (UTC)

Support. I like the deprecation of all-numerical ... definitely. And after thinking through that we have plenty of lists in which date abbreviations are poorly treated, I think it's a plus to provide two alternative short systems. You know I've gone on this minimalist kick lately: I do wish it was trimmed down to about 80%. But I hope we can agree that this is basically it. Tony (talk) 17:27, 16 September 2009 (UTC)

Alarics, since we have no gate-keepers here or roving bureaucrats with “God-powers” to declare that a general consensus has been achieved or that the wording has been washed in unicorn tears, we simply have work in good faith and do our best. I think it is quite clear that a general consensus exists on the principles embodied in the above pink-div. Surprisingly, I am not exactly sure where this fits on MOSNUM (I haven’t even looked), so I will give you honors (plague) of posting it to the proper place. Greg L (talk) 17:52, 16 September 2009 (UTC)

Very kind of you! Unfortunately I now discover that only administrators are allowed to edit the page, and I am just a humble pleb. I don't know who here is an administrator. As you say, it's not just a question of slotting it in at the right point, some other bits and pieces will also have to be shuffled around a bit. Who will step in and perform this noble deed? Alarics (talk) 19:36, 16 September 2009 (UTC)

Hmmm… it appears that what I have long been advocating (an admin gatekeeper to maintain order and stop edit-warring) has been put into place. I didn’t know that. Bravo! I’m not sure what formalities are required now. I’ll contact Tony and see what he recommends.
I do see that TheFeds has arrived way-late to the discussion. As I advised him below, “general consensus” does not require that 100% of editors be in full agreement and it never did. Those who wish to have significant input here are expected to fully participate. I see no reason for someone to arrive at a party after the guests are heading out the door to go home, and demand that everyone turn around, go back into the party, and have the discussions all over again with him participating this time. Greg L (talk) 21:04, 16 September 2009 (UTC)

Yellow-div proposal

I still see no need for these AP dates. If there's enough room, spell it out fully. Simplicity and consistancy seem good enough reason to stick only two options. We have enough consistancy strify with day-month vs month-day. ISO is not required for {{dts}} input (but it does accept ISO). It seems that the real point here is that it's not necessary to display ISO in order to get dates to work in autosorting tables (the simplest way of avoiding this being the use of {{dts}}, where it's the ISO output whichis hidden). I'm not sure that we need start talking about template input at the MoS. However, the point about the line wrapping is worth noting. So I'd go with something like

Dates are given with the month fully spelt out (12 September 2009, September 12, 2009) or, where space is limited, abbreviated to three letters (12 Sep 2009, Sep 12, 2009).
Do not use numbers to represent months, as it can be ambiguous (09/12/2009 could be 9 December or September 12); even when the day is greater than 12, spelling out the month minimises confusion.

Do not use YYYY-MM-DD format. (When automatic table sorting is needed, Template:dts produces hidden YYYY-MM-DD dates along with visible dates in an acceptable format.)

Where line-end word wrap can occur, best practice is to insert a non-breaking space between the month and day (6 May or May 6).

JIMp_talk·cont 00:46, 15 September 2009 (UTC)

Perhaps you, Jimp, may chose to personally not use the AP short-form for months. However, that style is widespread throughout the English-speaking world in print and it works very nicely in things like notes and citations, where one can make verbiage less ponderous, such as this: The magazine Science Week (issue 38, Sept. 15, 2007) referenced the first use of…. Rigid three-letter abbreviations certainly have their uses, but this example is precisely the intended use for the A.P. method.
Moreover, given that Wikipedia is “the encyclopedia to which *anyone* may contribute,” we need style-guide advise that helps editors and (very hopefully) leaves less clean-up work for experienced editors. What is in the above pink-div doesn’t take up much room and prevents articles from being mucked up with schemes like four-letter-only abbreviations (“Sept. Octo., Nove.” that might seem sensible to some, but aren’t used in professional publications); that is, until someone who knows better and is willing to fix it stumbles across it. Greg L (talk) 01:08, 15 September 2009 (UTC)
Anyone consulting MOSNUM will have no trouble reading what is in "Second pink-div", and "Dates should appear to readers with the month spelt out" cannot be clearer. When I have occasionally visited MOS/MOSNUM it was to find guidance, and there will be unusual circumstances where all the information in "Second pink-div", including the AP format, is very helpful. Johnuniq (talk) 04:05, 16 September 2009 (UTC)

As much as we can call it my preference not to have AP forms, could we not say that it's your preference to have them, Greg? We've all got our preferences. You note that the pink version "prevents articles from being mucked up with schemes like four-letter-only abbreviations". Yes, if followed strictly, it will. However, out there, I believe it's more of a edit-by-example situation than a follow MOSNUM one. People are likely to see "Sept." and churn out "Octo.", "Nove.", etc. by analogy. Uniform three-letter abbreviations across the board will set a simple rule that will be easy to follow without having to have people come here to find out exactly what it acceptable (Who knows the AP standard off by heart?). In short the yellow version is likely to cause less clean up. You suggest, Greg, that the AP style is common place throughout English. Is it common place throughout WP at the moment? Will the pink version, if implemented, not make it so? Nor do I really understand the need for AP style. Abbreviate if & only if you're short on space. If you are short on space, why not use the most space-saving abbreviation? Note that the three-letter abbreviations are perfectly understandable to anyone with a decent command of written English so settling on these consistently is not going to harm the reader. Let the contributor make the effort to conform. I think it's well accepted that the guidelines are aimed at the benefit of the reader not the writer. JIMp_talk·cont 09:50, 17 September 2009 (UTC)

Comment on requested edit

It's not clear what precise edit is requested here. Could you settle on a specific change with explicit wording you would like to make and get consensus for that please? Given the controversy surrounding MOSNUM, I'd rather not leave room for ambiguity or unilateral alterations. Cheers, Skomorokh 18:51, 12 September 2009 (UTC)

I see that when I copied the replacement text, I forgot to remove some indentation that another editor has used when writing it the talk page. To make it clear, I have struck out the text to remove and underlined the text to add. --Jc3s5h (talk) 00:28, 13 September 2009 (UTC)

Jc3s5h, you said earlier you were going to change the examples from 11 September to 12 September, but your underlined version only changed the first example. i see that modifications are still being discussed, but if you're the one "maintaining" the currently-proposed wording, all the examples should be changed to 12. Sssoul (talk) 06:32, 13 September 2009 (UTC)

It's clear that proposals are still in development. Once you (plural) have settled on something, then request the edit. Regards, Skomorokh 03:20, 13 September 2009 (UTC)

I am very concerned about bloat. The guideline should be as succinct as possible. Tony (talk) 10:26, 13 September 2009 (UTC)

Jc3s5h, is this where you want comments on the latest revision? i think "running prose" will baffle a lot of people - can we stick with "within sentences" instead? Sssoul (talk) 20:15, 13 September 2009 (UTC)

I have no problem with "within sentences", since, even though full sentences might show up in tables, they won't show up in space-constrained tables. --Jc3s5h (talk) 21:25, 13 September 2009 (UTC)

Templates

At the risk of stoking this rather unfortunate fire; the eventual wording needs to allow for the use of YYYY-MM-DD format in templates such as {{Birth date}}, {{Start date}} and their relatives. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:03, 14 September 2009 (UTC)

You mean in their input, or in their output? If it's the input, I don't see the point: wiki source code needn't follow English grammar and it's outside of the scope of the MoS. (You would normally use a space and no capital P in "subatomic particle", but that doesn't mean that Template:SubatomicParticle breaches the MoS or anything else.) If it's their output, I can't see why text generated by templates should be treated any differently than text hard-coded in the wiki source. --___A. di M. 19:16, 14 September 2009 (UTC)

Their input: the proposed wording could be taken as precluding such use, even if that's not what's meant. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 21:06, 14 September 2009 (UTC)

Rephrasing and simplifying

This core of this rephrased version is probably a little more succinct, per Tony's comment. However, I have restored the ordinal suffix part, and just incorporated non-breaking spaces as standard. (If an editor leaves them out, no big deal—the meaning is unchanged, and a bot can deal with it later.) We don't necessarily need to state the full reasons for everything in the MoS, though. (We can add explanations in footnotes using <ref group = "Note">reason</ref>, to reduce clutter in the prescriptive language.)

This offers the YYYY-MM-DD format as an alternative only when space is tight, stating that ISO 8601 compliance is not required or implied:

Within sentences, write dates with the month fully spelt out (e.g. 12 September 2009 or September 12, 2009), instead of using numbers to represent months (to avoid confusion between month-day and day-month interpretations). Do not use ordinal suffixes (i.e. -st, -nd, -rd and -th). Use non-breaking spaces between the month and day (e.g. 6 May or May 6).

In tables, references and template output, and only when space is limited, YYYY-MM-DD format may be used. Use leading zeroes for months and days.

When space is not limited, but a sortable format is required, {{dts}} may be used to generate sortable output, with dates fully spelt-out.

Conformance with ISO 8601 is not required, and is not implied by this format.

Or, there's this, which gives uses abbreviated month names and ignores ISO 8601 and YYYY-MM-DD:

Within sentences, write dates with the month fully spelt out (e.g. 12 September 2009 or September 12, 2009), instead of using numbers to represent months (to avoid confusion between month-day and day-month interpretations). Do not use ordinal suffixes (i.e. -st, -nd, -rd and -th). Use non-breaking spaces between the month and day (e.g. 6 May or May 6).

In tables, references and template output, and only when space is limited, the month may be abbreviated with the first three letters of the month name (optionally followed by a period).

When a sortable format is required, {{dts}} may be used to generate sortable output, with dates fully spelt-out.

I don't think we need the AP list; there's a very limited set of cases where an abbreviation is helpful, but where a 3-letter abbreviation is not sufficient. TheFeds 04:56, 16 September 2009 (UTC)

My two cents, for what they may be worth. (a) I believe that "spelled" will be more commonly understood than "spelt."

(b) While I'm fine with moving away from M/D/Y formulations in American-related articles if that is the consensus, I note that many American primary sources use it (e.g., Major League Baseball ... as in [11]). Also, Americans and Brits in addition of course use various phrases -- subway, football, bomb, boot, bum, corn, "a good screw", rubber, nervy, pissed, "to table", and "to knock up" (and no doubt a host of other phrases) -- in ways that are completely at odds with each other. But I guess that's not reason for us to hesitate about trying to advance matters where we can. Perhaps in the next phase we will have outlawing of each of those phrases on Wikipedia, and instead mandate phrases that are understood on both sides of the pond in the same manner.

(c) Might there be any sense to inline the dates, thus allowing the reader to change the setting on his computer so that dates appear in the manner he prefers? I thought that a pretty nifty trick, which exists now. But perhaps the problem is that readers may not know to set up their computers to address how dates appear.

(d) I agree with the sentiment that carried through most of this thread that YYYY-MM-DD should never be used in references. They never have been allowed in references in this guidance (only in lists and tables where space was limited). And the sentiment above seems strongly to move away from numbers where (as in references ... as well as text that is not a quote) there is no good reason to not spell out the month in some fashion.--VMAsNYC (talk) 08:02, 16 September 2009 (UTC)

VMAsNYC, I think you've misunderstood what is being done here re dates: there is no move away from M/D/Y in American-related articles. We are saying that "September 16, 2009" and "16 September 2009" are both acceptable (though I hope people will realise that we don't want both in the same article). It is only the all-numeric version of M/D/Y that won't do -- and neither will the all-numeric version of D/M/Y.

Re user-set date preferences: I assume you have been away for a while. After squillions of man-hours of argument, this idea has been decisively abandoned on all fronts, not least because people finally realised that the vast majority of WP users don't register and so don't have user preferences. Wikilinking of dates is now officially strongly deprecated.

TheFeds, I strongly feel we need to make it explicit that YYYY-MM-DD (in numerals) is not acceptable, so your second suggested version won't do as far as I am concerned (I think we reached a broad consensus on this yesterday). In your first suggested version, I would rather say "do NOT use YYYY-MM-DD except blah blah", rather than saying it MAY be used in certain circumstances. In particular, it needs to be obvious that we don't want it in references any more than in sentences.

I still think asking people to put non-breaking spaces in dates is peeing into the wind, and if you think it is "no big deal" why include it?, especially if, as you say, a bot can fix it. However, the deal yesterday with Greg L was that I would not press the point, in exchange for having an explicit condemnation of YYYY-MM-DD. -- Alarics (talk) 12:48, 16 September 2009 (UTC)

As for "conformance with ISO 8601 is not required, and is not implied by this format" I beleive it is beyond our ability to communicate to our readers that the YYYY-MM-DD date is only associated with an ISO standard if we declare it is, just as it is beyond the ability of the ISO to get people claiming conformance to the ISO 8601 standard to actually read it and actually conform to it. Remember, readers don't usually read MOSNUM, and we must concern ourselves with how readers interpret what we write. --Jc3s5h (talk) 15:07, 16 September 2009 (UTC)

Alarics, in the second version, wouldn't the "instead of using numbers to represent months" phrasing disallow in-sentence use of YYYY-MM-DD (and all other fully-numerical forms) explicitly enough? (And as for the first version, a change to word order to emphasize the "only" would be fine.)

VMAsNYC: I only retained "spelt" to be consistent with the previous proposal; "spelled" works for me too.

Regarding references, there has been discussion (on other talk pages) of how to make references more concise (in terms of both input and output), because with the multitude of fields that can be filled out, some references become enormous. Some citation templates explicitly recommend using date=YYYY-MM-DD, and others don't. Using the date parameter is an alternative to using the (redundant) year, month and day parameters, and makes for symmetry with the accessdate and publicationdate parameters (which don't have individual year, month and day equivalents). If we want to make a change that could affect the output guidelines for those citations, we should probably ask around before doing so. Also, I think reference lists are a harmless place to use a short-form date, because they're not read as prose.

Jc3s5h, I'm not sure if I'm clear on what you're saying. You said that we can't express the fact that "the YYYY-MM-DD date is only associated with an ISO standard"—but I'm trying to convey the converse, that YYYY-MM-DD isn't necessarily ISO 8601. (This is helpful, because most people have never read ISO 8601, and wouldn't necessarily know that it has special requirements beyond YYYY-MM-DD.) That avoids the need to explain what to do when someone wants to use that format for a non-Gregorian date. TheFeds 16:55, 16 September 2009 (UTC)

Re YYYY-MM-DD: No, I'm afraid I don't think "instead of using numbers to represent months" is explicit enough. I much preferred the version in the pink box further up this page: "Do not use YYYY-MM-DD format except .....".

ISO: I thought we had agreed earlier not to mention ISO at all, anywhere, and that that met Jc3s5h's concerns. It would also fit nicely with the fact that, amongst our generalist encyclopaedia readership, not many people will know what ISO is and many fewer will care what it says about dates. Again, please can we go back to the pink version, on which I thought we had a consensus.

Unwieldy references: as far as input (i.e. the mess in the edit window) is concerned, as I understand it (correct me if I have misunderstood) the main complaint is about citation templates with a large number of unused fields. I am not sure that is a problem that MOSNUM can deal with. Somewhere the other day there was a proposal for a bot to strip out unused fields; I think already people who use citation templates are encouraged to delete the empty fields. Where citation template documentation is encouraging people to use YYYY-MM-DD, the documentation will have to be changed. -- Alarics (talk) 17:36, 16 September 2009 (UTC)

(unindent) TheFeds, you wrote "I'm trying to convey the converse, that YYYY-MM-DD isn't necessarily ISO 8601." I know you would like to convey that, but you can't. The vast majority of the instances I've seen where someone within or without Wikipedia (including the Chicago Manual of Style get some aspect of it wrong. So rather than trying and failing to teach people how to use the format properly (and "properly" depends on whether you claim conformance to the standard or not) we should regard it as permanently contaminated by the ISO, and ban it in any context where it is likely to be misused. It's OK for computer-generated timestamps, but should not be used in articles. The effort to get people to use it as if it were covered by ISO just in case our readers think it is seems to just cause more confusion. Ban it, just like it is illegal to mark a gasoline tanker truck (petrol tanker lorry) "inflammable" in the USA, due to possible confusion. --Jc3s5h (talk) 18:15, 16 September 2009 (UTC)

Jc3s5h, just to be certain, then: you're saying that we should eliminate every all-numeric date format (including YYYY-MM-DD) from prose, tables and references? And that the justification for this is that MM-DD-YYYY and DD-MM-YYYY are mutually confusing, and YYYY-MM-DD could be interpreted to be ISO 8601-compliant by a reader, even if that was not intended by the editor who used it? If you don't think a disclaimer of ISO-compliance will be enough to mitigate that problem, we can certainly stick with something along the lines of the second (turquoise) version.

Alarics, I didn't interpret the removal of mention of ISO to be the only potential solution to Jc3s5h's concern, which is why I offered that alternative. But given that it seems unsatisfactory, let's try this:

Within sentences, write dates with the month fully spelled out (e.g. 12 September 2009 or September 12, 2009). Do not use all-numeric formats (e.g. 2009-09-12). Do not use ordinal suffixes (i.e. -st, -nd, -rd and -th). Use non-breaking spaces between the month and day (e.g. 6 May or May 6).

In tables, references and template output, particularly when space is limited, the month may be abbreviated with the first three letters of the month name (optionally followed by a period).

When a sortable format is required, {{dts}} may be used to generate sortable output, with dates fully spelled-out.

This is largely like what Jimp had, but a little more concise (combining the don't-use-ISO and the don't-use-confusing-formats rules), and incorporates the ordinal suffix rule.

As for the non-breaking spaces, if they are universally desirable, we might as well say so somewhere. If we really don't care one way or another what editors choose (e.g. we can live with lines breaking between "May" and "6"), then we can leave it out. TheFeds 19:51, 16 September 2009 (UTC)

TheFeds: We’ve discussed the pink-div above for quite some time and you stayed out of that discussion for whatever reason (you didn’t see it, didn’t want to participate, were too busy, whatever). Unfortunately, it takes an exceedingly great deal of time to explain the blow-by-blow how how we ended up with what we did. In a nutshell: parties compromised on both sides. There is now a general consensus in support of the pink-div. It appears you feel the pink-div has shortcomings you don’t fully support; it would have been much better if you had participated the entire time in the discussion since influence requires full and proper participation; otherwise, you just waste everyone’s time here. Please also note that “general consensus” does not mean that 100% of editors are in agreement—and it never did. If you can later develop a general consensus in support of this version of yours, great. Good luck—it’s hard to do. Greg L (talk) 20:36, 16 September 2009 (UTC)

TheFeds, you asked "Jc3s5h, just to be certain, then: you're saying that we should eliminate every all-numeric date format (including YYYY-MM-DD) from prose, tables and references?" Yes. Originally, I was neutral. But after seeing during this discussion how much difficulty it causes for careful, experienced editors, I have to conclude the format is permanently contaminated and cannot be rehabilitated. The closest analogy I can thing of is getting people to write noon or midnight in an unambiguous way in countries where am and pm are in general use. It's just hopeless. The usual solution is to just not use noon, midnight, 12:00 am, or 12:00 pm, and instead use 12:01 am or 11:59 pm. --Jc3s5h (talk) 20:54, 16 September 2009 (UTC)

The Feds, your new green version still doesn't explicitly condemn YYYY-MM-DD, whether in text or references. We need to do that, because so many people have now got the idea that it's OK to use that form, in references at least. Can we please now agree on the pink version, further up. Alarics (talk) 21:26, 16 September 2009 (UTC)

Given that so many people (me included) use the YYYY-MM-DD format in references, we should probably advertise this discussion more widely if we're going to discourage it entirely. Dabomb87 (talk) 21:32, 16 September 2009 (UTC)

I think you're all tap dancing around the fact that YYYY-MM-DD is the international standard date format.

ISO 8601 - International Standard
EN 28601 - European Norm
BS EN 28601 - British Standard
IS/EN 28601 - Irish Standard
ANSI X3.30 - American National Standard
CSA-Z243.4 - Canadian Standard
AS 3802 - Australian Standard
IS 7900 - Indian Standard
JIS X 0301 - Japanese Standard
GB/T 7408 - Chinese Standard

You may not like it, but it is the only date format that is unambiguous and universally recognized around the world. If you're writing international documents, or international software, it's your first choice of date formats. It would help enormously if more people knew more about it.RockyMtnGuy (talk) 02:03, 17 September 2009 (UTC)

RocyMtnGuy, your posting illustrates the problem. There is one format, at least 10 standards, plus the informal assumptions of the people who have not read the standards. Have you read all those standards? Are they all equivalent? I'm quite certain that the ISO 8601 standard does does not match the informal interpretation that many of our readers would have. ISO tried to get the world to adopt a standard, but they found out they were not the Vatican and it isn't the 16th century. --Jc3s5h (talk) 02:11, 17 September 2009 (UTC)

RocyMtnGuy, you can’t seriously be suggesting that anything more than 0.001% of the readers visiting the typical Wikipedia page (or reading the typical Encyclopedia Britannica page or The New York Times page, etc.) has ever heard of these standards—or could possibly give a rip. So then let’s cut to the chase: The New York Times and Encyclopedia Britannica and Aviation Week & Space Technology, and Scientific American and pretty much any half-way professional publication writes out the month of the day because that’s proper writing style. Moreover, they couldn’t give a rip about your “standards” either because all-numeric dates have no place in anything but airline tickets, computer punch cards, and nine-mile-long mailing lists from a database company.
All you’ve demonstrated with the above list (which must have taken quite some time to dredge up), is that you are fascinated with “standards.” Either that, or you think that by throwing standards organization alphabet soup on this talk page that you might get your way. The trouble is, all-numeric dates don’t look professional and have no place whatsoever in an encyclopedia. I suggest you place greater credence in looking towards the manuals of styles of professional publications that are inhabited by professional writers; you know—those individuals who have journalism degrees because they went to college to get an advanced education on how to write clearly.
The practical effect of TheFeds’ blue-box above (“Conformance with ISO 8601 is not required, and is not implied by this format”) is that all-numeric dates would still be permitted. Just pardon me all over the place for thinking that might have been the intended effect. That’s also precisely why most everyone else here doesn’t want that wording and understands that the practice of any all-numeric date must now be tossed out on its ear with an explicit proscription, as is shown in the pink-div, above.
Your bold emphasis, above “international standard” is entirely beside the point. There’s an international standard to write out “75%” as “75 %” (with a space before the percent symbol). Perhaps I should start railing here about how Wikipedia is flouting the rule of SI with regard to the percent symbol. After all, it is an international standard that is backed by the BIPM, which is the god of all standards organizations and that means the standard has been washed in unicorn tears (yadda yadda yadda). No one gives a rip about standards if they are being misapplied to such an extent that common sense good-writing practices have to be abandoned. Wikipedia follows the way the real world works and is always at great peril of looking like it’s been hijacked by another “IEC prefix” crowd whenever we depart from well-known manuals of style. Greg L (talk) 05:13, 17 September 2009 (UTC)

My fixation on standards arose because I used to design software which had to be usable in a number of countries, including the U.S., the U.K., France, Canada, and various places in Latin America. Fortunately, all of the users understood English to varying degrees, but you could never assume they understood the same measurement units or date formats as you do. A lot of the things you might consider to be "standard" vary from country to country. As a result, I've read a lot of standards manuals, not all of them in English.
I agree that in the English Wikipedia, in most instances you would want to spell out the month in full, e.g. 16 September 2009 or September 16, 2009. It doesn't really matter which one you use because they are both understandable. The problems arise if you use a numeric date format, e.g. 2009-09-16, which you might want to do in a table or in a template - either because you are short of space, or because you want to sort it, or because the date arithmetic is much simpler.
The exact details of ISO 8601, which covers much more than calendar dates, are not really relevant to this discussion. Only the date format YYYY-MM-DD is - although ISO 8601 also does resolve other ambiguities such as whether 12:00 is AM or PM. The particular national standards bodies basically reiterate ISO 8601, and the detail differences are immaterial.
The fundamental problem with the American date format (MM/DD/YYYY) and the British date format (DD/MM/YYYY) is that they are mutually ambiguous. Another problem is that Americans don't consistently use the MM/DD/YYYY format and the British don't consistently use the DD/MM/YYYY format. The advantage of the ISO format, which I consider to be the "international" format, is that it is unambiguous, and is neither American nor British. It's also been around a lot longer than you think, under various different names.
Now, you may consider whatever date format you use to be "natural", but I'm in Canada, where this is a more complicated issue. Canada has two official languages and the English language population has both British and American roots, in addition to which much of the population has other national origins. As a result there are at least four different date formats in common use. That being the case, the style manuals commonly recommend the ISO date format because it is the only one which is unambiguous and readable in any language.RockyMtnGuy (talk) 06:38, 17 September 2009 (UTC)

What you are saying applies to stuff like technical documentation. However, this is an encyclopedia for general readers so instead of quoting standards bodies, you should be looking for examples of how dates are written by quality publishers for general readers. To put it another way, any time one of your standards bodies cares to publish an encyclopedia, they can show the dates however they want. Likewise, we can show the dates however we want. Johnuniq (talk) 07:30, 17 September 2009 (UTC)

RockyMtnGuy has got hold of the wrong end of the stick here, I think. We're not talking about producing technical documentation for a foreign-language audience, we're talking about an English-language encyclopaedia. In his bullet point beginning "the fundamental problem with ...." everything he says is correct, and it is what we have all been saying at various points on this now very confusing and jumbled page. He sets out clearly the reasons for not using either the American or the British/international numerical form. We are all agreed on that now, I think. But the point is, we have already gone further than that and said that we also don't want dates in any numerical form, including YYYY-MM-DD. Most of us agree that YYYY-MM-DD is the least likely numerical form to be ambiguous, and is probably fairly widely understood, but not everybody agrees that it is absolutely unambiguous, and anyone not familiar with it has to stop and work out what it means. In any case, all that is eclipsed by the other arguments against using any numerical date form except in special, tightly-defined circumstances (tables perhaps, but NOT footnote references). -- Alarics (talk) 10:16, 17 September 2009 (UTC)

Just my three farthings worth.

Spelt vs spelled will boil down to conformance with the dialect used on the page which happens to be US Eng (if I'm not wrong) so those who would have it spelt spelled are in luck.
I'm not sure what the need is for full stops after three-letter month abbreviations. Note that you would never dot May so there's a speck of inconsistency. Note also that Jun., and Jul. save no characters & Apr. & Mar. save but one each.
MOSNUM can declare this and can declare that about how YYYY-MM-DD dates are to be interpreted. What percentage of readers come here for instructions on how to read WP? We could even try to allow the format only when no problems of interpretation may arise but this is likely to be doomed (How many standards have we got listed above?). No, YYYY-MM-DD dates are best eliminated along with all other numeric dates.
{{dts}} can produce abbreviated output e.g. {{dts|15 Mar 1944}} → "15 Mar 1944".

JIMp_talk·cont 10:26, 17 September 2009 (UTC)

In fact, I just checked, and both Oxford and Collins accept either "spelled" or "spelt" for British English, so at least that's one problem that needn't detain us further. Alarics (talk) 11:25, 17 September 2009 (UTC)

Oppose all proposals which forbid the use of 2009-09-17 format, as this is the standard international way of writing dates, and should be allowed on Wikipedia. Offliner (talk) 14:33, 17 September 2009 (UTC)

I see that Offliner opposes ISO-8601, which forbids the YYYY-MM-DD format for dates in the Julian calendar, and (in the absence of agreement with our readers) years outside the range 1583 through 9999. So Offliner, since you oppose ISO 8601, what specification of this format do you support? --Jc3s5h (talk) 14:54, 17 September 2009 (UTC)

Nope, I do not have an opinion about ISO-8601; I'm only saying that I support the usage of YYYY-MM-DD format in Wikipedia, since it's the clearest and most unambigous format and used as a de facto standard in international communications. Offliner (talk) 15:07, 17 September 2009 (UTC)

Whatever our opinion about ISO-8601 as long as the MOS permits YYYY-MM-DD we'll have to say something about what to do with Julian dates. Do we allow Julian YYYY-MM-DD dates against the ISO advice? Do we conform strictly to ISO? Do we attempt to limit YYYY-MM-DD dates to a 1583–9999 range? Is it worth all this bother to preserve a format which is not that familiar and not really usual in written English? Certainly we might argue that of all all-numeric formats only year-month-day is unambiguous with respect to the day-month problem. However, this problem is easily enough fixed by using letters for months. Letters-for-months format is unambiguous and clearer than YYYY-MM-DD. YYYY-MM-DD may be a de facto standard in international communications but not in English. JIMp_talk·cont 17:22, 17 September 2009 (UTC)

Offliner, did you read any of the preceding discussion? Nobody disputes that YYYY-MM-DD is the least ambiguous form of numerical date. But the point is, and we have pretty well all agreed on this, any numerical date is inappropriate in an encyclopaedia. We want people to spell the month out in words, as would be done in any encyclopaedia, book or newspaper. Alarics (talk) 17:27, 17 September 2009 (UTC)

Not sure how this works, but is it ok for me to suggest that since there is not controversy over the use of dates with months spelled out (but is controversy over the use of numerical dates) that we use the months spelled out format?

To be honest, one point that hasn't been made (but that I, due to my ignorance, can make) is that I (and, I wager, wide swaths of the public) had no idea until now that there is no YYYY-DD-MM format. Thus, upon seeing a date 2009-02-03 for the first time, I would have no idea whether it was Feb. 3 or March 2. Therefore, for those as ignorant as I was, the all-numerical format needlessly introduces ambiguity.

I also find it off-putting, non user friendly, at odds with publications I read, and agree with the above sentiments that it should not now for the first time be permitted to be introduced into references.

I also would favor the deletion of the proposed "non-breaking spaces" paragraph as impractical and unnecessary.--VMAsNYC (talk) 18:33, 17 September 2009 (UTC)

Alarics, the green version says "Do not use all-numeric formats (e.g. 2009-09-12)." That's a succinct way of banning all such constructions, YYYY-MM-DD clearly included. I don't think that YYYY-MM-DD merits special mention, mainly because it also invites us to also explain why MM-DD-YYYY and DD-MM-YYYY are prohibited. We don't really need a full explanation in the guideline text itself (though a footnote wouldn't be a problem). Maybe we could strengthen the wording from "do not" to "never"?

Jimp, the full stops/periods are to accomodate those who don't like seeing abbreviations without a terminating period, but I wouldn't be opposed to doing away with that entirely. (British English in particular deprecates period usage in most abbreviations.) The AP style with three-to-five-character forms is common as well, but not particularly useful on Wikipedia if we're only interested in using the abbreviations in tight spaces, because the three-letter forms are obviously more compact, and very clearly understandable. (If we were to recommend month abbreviations in prose, then the AP style would be appropriate.) As for consistency about May, that's why I said "abbreviate": you're not abbreviating "May" by writing it with three letters, so the proposal did not permit the use a period in that case. (Maybe that that logic could have been clearer.)

VMAsNYC, is it that you disfavour any non-breaking space guidance, or just that you'd rather see it as an item in WP:NBSP? (I don't think editors will regularly use non-breaking spaces when they write a date, regardless of what the guideline says. But if there's consensus that a non-breaking space is appropriate there, by codifying that consensus into a guideline somewhere, an AWB user or a bot will be able to make the change without lengthy discussion about the propriety of that action.)

With regard to the reservations expressed by RockyMtnGuy and Offliner, I'm a little torn. On one hand, the YYYY-MM-DD format is common, less ambiguous with regard to months and days, and compact. But it looks like there's a feeling here that the distinction between YYYY-MM-DD and ISO 8601 (mostly related to ISO 8601 being unsuitable for expressing Julian dates) will cause enough confusion to warrant eliminating it entirely. I proposed a version where ISO 8601 was disavowed, with YYYY-MM-DD still permitted, however, since many felt that insufficient in comparison to the risk of misinterpretation, I dropped it. (Also, note that I'm not a fan of YYYY-MM-DD in prose: I have no reservations toward disallowing it in sentences.)

I still think we need to ask around the places where citation-oriented people congregate, whether it's a good idea to force them to change the output of the accessdate and other parameters. And I definitely think that for template input, ISO 8601 is perfectly appropriate, and the text of any proposal is not intended to prohibit this. TheFeds 18:45, 17 September 2009 (UTC)

Those citation templates now just give back what you put in. If we were to have them convert ISO to spelt-out month, then we'll have the day-month vs month-day problem. Probably the best thing to do is fix things on the input end. Of course we're not dotting May but it won't look good next to a bunch of dotted abbreviations. We stipulate no dots with abbreviations/symbols for units of measurement. Some like these specks, some don't; we can't please everyone. I agree with you about the relative uselessness of AP on WP. JIMp_talk·cont 20:02, 17 September 2009 (UTC)

Reply to Jimp. I guess I disfavour any guidance here (which is the only place I'm focusing) that I can't understand (and imagine most people won't understand) and which I (and apparently others) expect most people won't follow. I think its bad form to have a law that we expect will be ignored largely. If there is a consensus of those who know what it means, and it can be said elsewhere, and said in a way that allows a bot to do the work (rather than one that mandates that editors do the work), I guess that would not trouble me. Just my two cents.--VMAsNYC (talk) 20:20, 17 September 2009 (UTC)