Wikipedia talk:Manual of Style/Dates and numbers/Archive 131

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Consensus needed to allow lighthouse ranges in km

Comments moved from Lightmouse talk page:

I notice that you are adding kilometers to distances expressed in nautical miles, as at Mount Desert Light. I think it deserves a little discussion. I wrote many of the articles you are are changing and deliberately used only nautical miles, as that is the measure used by virtually all mariners. I can't argue with the premise that we ought to use landlubbers' units, at least for statements like,

"...Mount Desert Rock, a small island about 18 nautical miles (33 km) south of Mount Desert Island."

I think, though, that if we're going that way, then it should be both statute miles and kilometers -- after all most of the readers of articles on US lighthouses are going to be US residents and therefore more or less metric illiterate. A nautical mile is no more likely to be understood by an American that by a European, unless they are water people.

I might even go farther and suggest that we eliminate nautical miles entirely from locations -- and keep them only for the range of the light, without conversion. . . Jim - Jameslwoodward (talk to mecontribs) 10:50, 25 October 2010 (UTC)

You make some interesting points. The place for a house-style discussion like this is Wikipedia talk:Manual of Style (dates and numbers). Would you mind reposting your comments there? Lightmouse (talk) 11:23, 25 October 2010 (UTC)
I don't think it needs discussion there -- what I am proposing is closer to existing policy, which suggests using both statute miles and kilometers in articles. Note that this usage is already implicit in in {{convert}}, where we have
{{convert|19|nmi}} which yields 19 nautical miles (35 km; 22 mi) or
{{convert|19|nmi|abbr=none}} which yields 19 nautical miles (35 kilometres; 22 miles)
That is, you can force {{convert}} to show only km, as you did, but it naturally shows both km and miles. If, however, you really want to discuss it, I suggest that Wikipedia talk:WikiProject Lighthouses is a better forum.
. . Jim - Jameslwoodward (talk to mecontribs) 11:50, 26 October 2010 (UTC)

The more people ensuring conversions to km are provided, the better. It would be great if you guys at the lighthouse project could go through the articles and ensure that km is provided (my main concern) in whatever way suits you. I'm less keen on the suggestion that km would not be provided for the range of the light. Lightmouse (talk) 12:37, 26 October 2010 (UTC)

I trust you know that no one -- American, British, European, Chinese, whatever, uses km at sea -- the nautical mile is the only unit used, for good reason. The International Regulations for Preventing Collisions at Sea 1972, with approximately 170 signatories, uses only nautical miles for long distances (short distances -- the lengths of vessels and tows -- are in meters). Hence our choice of only nautical miles for the range of lights, an essentially nautical measurement.. . Jim - Jameslwoodward (talk to mecontribs) 22:33, 26 October 2010 (UTC)

I could give you my opinion but it's more a matter for the Wikipedia talk:Manual of Style (dates and numbers). It currently allows or even encourages editors to provide a 'km' value for the range of lights. If this is to be forbidden in future, then it would be best to discuss it there. Would you like to make a posting there along the lines you suggest or would you prefer it if I started it? Lightmouse (talk) 14:55, 27 October 2010 (UTC)

Well, first I would quote the MOS
"use the units in most widespread use worldwide for the type of measurement in question."
That is, as I point out above, clearly the Nautical Mile and nothing else. Km are not used at sea by anyone.
However, I have no problem with adding km to provided that you also add statute miles. I have the impression from your comments above that you believe that somehow the nautical mile is a US/UK creature that is properly balanced by the km and only the km. That is not correct -- the nautical mile is a generally accepted international unit. I have no problem with stating light ranges in nautical miles, km, and statue miles, but I think it is within the spirit of the MOS to favor metric system landlubbers with distances in km while leaving American and British landlubbers in the dark. If you insist on adding land distance to sea measurements, then please add both metric and customary, as required by the MOS. . . Jim - Jameslwoodward (talk to mecontribs) 16:21, 29 October 2010 (UTC)

Can we please have comments from others on this issue? Lightmouse (talk) 16:51, 29 October 2010 (UTC)

I think giving the distance in nautical miles is important, since that's the standard unit for sea travel and people working in that area need the distance in this format. Many readers (like me) however don't know anything about nautical miles, so a conversion to the international standard unit km is necessary. We should additionally consider including a conversion to anglophone miles, but only if the article is about a country which uses these units. Offliner (talk) 17:20, 29 October 2010 (UTC)
Or, of course, the other Wikipedia solution for words which not everybody knows but which are standard in a given context: a link to nautical mile. The immediately available information that a nmi is a bit shy of 2 km is really all most readers will need. Septentrionalis PMAnderson 17:37, 29 October 2010 (UTC)
(ec)The easy solution is to let the {{convert}} deal with it(BTW, it provides km and mi by default). Don't force output units! Let those working on that tool address the problems if any. I think the problem comes in when an automated tool starts forcing output units especially when they differ from the default. To do that is overriding what editors have actually selected for that article. Vegaswikian (talk) 17:48, 29 October 2010 (UTC)

I'm much the same as you, Offliner. It's too much of a constraint to say that a metric editor cannot add metric units if they don't know, and implement, non-metric units that are applicable for the domain, region, and era. The addition of a conversion into km does not prevent lighthouse specialists from adding more conversions. Lightmouse (talk) 17:44, 29 October 2010 (UTC)

Converting just to kilometres does little to help a landlubber like me in the U.S., so if the conversion template is giving statute miles just let it do so, so long as it also gives metric conversions (km or kilometres) for non-nautical readers everywhere else. I don't think anyone would dispute a universal specialist convention that gives Nautical Miles first, especially since there's a safety question involved (I'm sure that most readers of such articles are doing so in buildings planted very firmly on dry land, but I guess one should always think of someone afloat who might be using Wikipedia —or something derived from a Wikipedia article— as a practical, immediate aid to navigation or plotting.) —— Shakescene (talk) 18:13, 29 October 2010 (UTC)

Can't see why nautical miles can't just be converted to both miles and km. Those not familiar with nautical miles will generally be familiar with either miles or km, and I can't see a good reason for boxing out one or the other. We are not here to try and force people to use units we happen to like, but to make WP accessible for everyone. Similarly of course knots, mph and km/h. Richard New Forest (talk) 18:41, 29 October 2010 (UTC)

That is the default. An example output from using nmi with defaults on the conversion is 25 nautical miles (46 km; 29 mi). The problem as I see it, is Lightmouse not allowing convert to select the default units. Take the defaults and the problems go away. Vegaswikian (talk) 18:52, 29 October 2010 (UTC)

Indeed. Richard New Forest (talk) 20:46, 29 October 2010 (UTC)

  • I don’t see any problems. The article, ‘Mount Desert Light’ currently reads as follows:
Mount Desert Light is a lighthouse on Mount Desert Rock, a small island about 18 nautical miles (33 kilometres; 21 miles) south of Mount Desert Island, which is part of the state of Maine in the United States. It was first established in 1830.
I’m not noticing anything that gets my panties in a bunch. It is an island that is reachable by modest pleasure craft, so it seems unremarkable and natural for the distance over water to be expressed in nautical miles. The primary unit of measure was fully spelled out on its first occurrence (kudos). Since the fact that nautical miles are only a tad bigger than a statute mile is not generally appreciated by a general-interest readership and since the island has a strong connection with the United States, a parenthetical to statute miles (“miles”) is provided for an American readership. Again, kudos. Moreover, the “miles” wasn’t encumbered with excess specificity like “statute miles” or “barbarian American miles”, so “triple kudos.” Since Wikipedia is an on-line encyclopedia read by many other English-speaking peoples who think in terms of kilometers, it provides a parenthetical in kilometers. Kudos on that too. I think I personally would have used the unit symbol for kilometers (km) in the conversion, but in this instance, where it would be juxtaposed next to a fully spelled out “mile”, that is a matter of editorial discretion that doesn’t jump out at me one way or another. I am truly not seeing what the problem is here.
The only thing I noted—while everyone is worrying about units of measure—is the lede entirely omitted mention of where on this pale blue dot the island is located, so I just now added that it is Maine, part of the U.S. (oh… that little detail). Greg L (talk) 22:24, 29 October 2010 (UTC)
The issue is this change. It forces a specific conversion which is not correct. All of these changes require someone to find them and fix them. Vegaswikian (talk) 02:07, 30 October 2010 (UTC)
Surely a full conversion of nautical miles on first appearance, to both km and landlubber miles; then just nautical miles? PMA, the articles on units typically don't provide the conversion rate quickly and conveniently (some of them go into etymology and history first). Much better for the readers to give it in full just once, I think. Tony (talk) 02:20, 30 October 2010 (UTC)
Links are never a good way to offer conversions. They can be read quickly only by registered users who've enabled both JavaScript and pop-ups, who are some fraction (I think a minority) of all registered users, who are a minority of editors, who are in turn a small minority of readers. Few people will load a new page just to find out the meaning of nautical mile. They will either make uninformed assumptions or plough on in ignorance of something they'd prefer to know. —— Shakescene (talk) 04:17, 30 October 2010 (UTC)
I checked Nautical mile before posting; it gives the conversion (1,852 m) in the first paragraph, and loads quickly, although I have neither of these mod. cons. For my part, having an interruption of thought at every measurement is a recipe for bad writing. Serving a minority at the expense of all readers is, as usual, a net harm. Septentrionalis PMAnderson 06:54, 30 October 2010 (UTC)
But if it just happened on first occurrence ... would that be a better compromise than linking? It would seem to provide the conversion up-front; BTW, nmi in metres would probably be betters as nmi in km (1.9 km), or 1.85 ... not sure how many decimal places are appropriate. Tony (talk) 07:08, 30 October 2010 (UTC)
A nmi is defined as an exact number of meters, as inches are an exact number of mm (this may be Vegaswikian's beef above). It may be best to convert once in an article like Mount Desert Light, but it depends on the article; is the exact value of some other distance important? What I would like to avoid is the 10 nmi (18 km) from X, 20 nmi (37 km) from Y clutter. It also depends on the context: it is false precision to convert about 18 nmi away to 33,336 m, but it is true precision to convert a ten (nautical) mile exclusion zone to 18,520 m. (Is the precision worth the trouble? That's why we have editors.) Septentrionalis PMAnderson 07:28, 30 October 2010 (UTC)

←I was thinking more like this (assuming first appearance in an article):

10 nmi (11 mi; 18 km) from X, 20 nmi from Y.

That is, after you get the idea of the proportions, it's nmi only. What happens with knots? Tony (talk) 07:42, 30 October 2010 (UTC)

Presumably the same thing as with nmi, adding per hour. (I might spell per hour out, or allow others to do so; some readers do not understand that knots are a unit of velocity.) Septentrionalis PMAnderson 15:36, 30 October 2010 (UTC)
  • I still don’t understand the concern. Vegaswikian wrote that the issue is this change, which now resolves to this body text. That text read 18 nautical miles (33 km). He further wrote All of these changes require someone to find them and fix them. A quantity of 18 international nautical miles is 33.336 kilometers so the conversion to “33 km” is correct and does not suffer from excess or insufficient precision. And if we use the U.S. nautical mile, one obtains the same value. So I don’t understand why there is so much wikidrama over this conversion (other than it would be nice to also have statute miles, which is now there). What I especially liked was that someone was bold enough to not use “nmi”, which would be a house style that draws attention to itself since there is no single, universally adopted unit symbol for the nautical mile and—even if there were—few of our general-interest readership would recognize it. So, quite correctly, not only is the primary unit of measure fully spelled out, but so too are the ones in the parenthetical conversion (although I think it would have been entirely satisfactory to use “km” in the parenthetical, as the bot correctly defaults to). Greg L (talk) 18:56, 30 October 2010}

Leading zeros in 12-hour clock times?

MOS:TIME doesn't state explicitly whether or not one should use leading zeros for 12 hour clock times, eg 1:38:09 pm or 01:38:09 pm, but I think it should. Currently it only implies by example. For 24-hour clocks, it states explicitly that "discretion may be used as to whether the hour has a leading zero". I suggest that MOS:TIME should state an explict guideline for leading zeros in 12-hour clock times. Presumably that would be either "no leading zero" or "leading zero is optional". Mitch Ames (talk) 01:48, 30 October 2010 (UTC)

There is never a leading zero in 12 hour clock times. Headbomb {talk / contribs / physics / books} 04:55, 30 October 2010 (UTC)
OK. If I can get someone to second that, I'll update the MOS page - unless someone else does it first. (It's not that I don't believe you, but I want to ensure we have consensus.} Mitch Ames (talk) 06:21, 30 October 2010 (UTC)
Leading zeroes are not utterly unknown (some authors are simply copying digital clocks), but they certainly are not customary; in fact, some people use leading zeroes to differentiate between 12-hour and 24-hour clocks. I suppose if you wanted to make some point involving the leading zero (like the time 09:09:09 AM, September 09, 2009), you might use it, but that's IAR. Septentrionalis PMAnderson 07:01, 30 October 2010 (UTC)
Per PMA. Tony (talk) 07:04, 30 October 2010 (UTC)
Are there so many people using the 12-hour clock with a leading zero on Wikipedia and being reverted? If not, explicitly stating that there's no leading zero looks like a solution searching for a problem (see also WP:BEANS). A. di M. (talk) 12:33, 30 October 2010 (UTC)
It's not necessarily solely that there may be few instances right at the moment (I believe they have been cleaned up regularly); it's that there's a steady trickle of "anyone can edits" adding them—not a lot, but some. There is no reasons our gnomes should not be backed up by a short addition to the style guide. Tony (talk) 13:28, 30 October 2010 (UTC)
  • In all articles of an encyclopedia directed to a general-interest readership, there should never be a leading zero in 12-hour clock times. Why? Because 7:30 AM is 1) unambiguous, and 2) most natural and common to our readership. Bravo to Mitch Ames. I see he is active in all-things-ISO (his contributions) and nevertheless saw fit to suggest that MOSNUM should be explicit one way or another when he wrote I suggest that MOS:TIME should state an explict guideline for leading zeros in 12-hour clock times. Presumably that would be either "no leading zero" or "leading zero is optional". I agree with him; MOSNUM should not be silent.

    I might add that military-related articles should be no exception. We don’t have military personnel coming to our articles to find out at what time their duty begins to man the guardhouse; common sense tells us our military-related articles are overwhelming read by civilians; ergo, it would be unwise to start using unnatural looking time formats just because the subject matter is of a military nature. Quite correctly, our article here on ‘Death of Linda Norgrove’ wisely states that the rescue attempt began at 3:30 AM and not 03:30. One should never write in a fashion that unnecessarily draws attention to the writing style; it interferes with the the objective of transparently communicating thought. As always, exceptions would be direct quotes and articles discussing time formats. Greg L (talk) 17:38, 30 October 2010 (UTC)

That's a different issue. 03:30 is a 24-hour time; if it were 12-hour it would have to specify (perhaps implicitly) AM or PM. Septentrionalis PMAnderson 02:18, 31 October 2010 (UTC)
Uhm… Yeah; good point (ever-diligent in keeping me honest, aren’t you?). But the same goes for 12-hour time IMO: MOSNUM should explicitly advise to lose the leading zero and use AM & PM; that is the way time is routinely expressed in real life and print in newspapers and magazines and is what our readership is most accustomed to. For editors to do otherwise causes *!* brain-interrupt neurons to fire when text unnecessarily calls attention to itself. Greg L (talk) 03:13, 31 October 2010 (UTC)
If it were a problem, as was suggested above, then it would make sense to explicitly indicate a position. I have not seen this to be a problem so why do we need to explicitly state it? --Walter Görlitz (talk) 03:20, 31 October 2010 (UTC)
I assume Mitch Ames has seen a problem that you and I haven’t seen. He is *into* ISO-related articles. The ISO is a standards body that touches upon—among other things—temporal-related standards. I suspect Mitch has a keen eye on the issue of how time is expressed on Wikipedia’s articles. How about it Mitch? Is your concern founded upon an observation that some articles have had time expressed in an inappropriate format and/or there is editwarring over this issue? Or is your concern purely theoretical? Greg L (talk) 06:17, 31 October 2010 (UTC)
My question was triggered by this edit, which prompted my removal of the leading zero. Personally I don't like the leading zero - but I know that's not sufficient grounds to change something, so I checked MOS:TIME, where I noticed the explicit mention of leading zero for 24 hour time but for 12. I'd like to see consistency in the guidelines as well as the articles, hence this discussion.
(I am a big fan of standards, especially ISO standards, and - as you've noticed - I take an interest in a few of them. However - aside from reflecting my pedantry - that's not related to this issue of whether MOS should mention the leading zero. ISO 8601 (which is on my watchlist) does cover date/time representation, but has only 24-hour time, and I believe consensus is that we need not be required to follow it.)
Mitch Ames (talk) 06:57, 31 October 2010 (UTC)
Leading zeros are pretty rare in the 24-hour clock, too, despite the impression one could get from our article 24-hour clock. A. di M. (talk) 11:46, 31 October 2010 (UTC)
Yeah? I see that the Coordinated Universal Time article says, "... TAI instant 1 January 1961 00:00:01.422818 exactly was identified as UTC instant 1 January 1961 00:00:00.000000 exactly". On the 24 hour clock, your local time, what time is it one second after 2400? Lately, I've had my alarm set to oh-five-hundred because of something I want to catch on TV. Sometimes, though, I still get up at oh-dark-thirty. Wtmitchell (talk) (earlier Boracay Bill) 12:41, 31 October 2010 (UTC)
Outside technical contexts and the like, 0:01 is much more common than 00:01 in my experience. A. di M. (talk) 12:50, 31 October 2010 (UTC)

My observations are that in the UK, leading zeros do not appear on 12 hour times. Moreover, in the UK (unlike the US), the 24 hour clock is used for railway and airline timetables so a Brit would ask whether "1:00" was "am or pm", but would accept that "01:00" was an hour after midnight. The Continent does not have these problems - the terms "am" and "pm" (or the foreign equivalent thereof) are hardly ever used there. Martinvl (talk) 13:14, 31 October 2010 (UTC)

  • I think we can examine expressions of time in the same vein as the delimiting of numbers. Swiss (as I recall) schoolchildren are taught three or four ways to delimit numbers, (one of the Swiss ways being 1 234 567,89). But one of the techniques taught to Swiss schoolchildren is the American technique (1,234,567.89) where the decimal marker is a fullstop and commas delimit to the left of the marker. Americans are taught only one way to delimit number and have not even been exposed to other ways in school. Since American-delimited numbers are familiar to all, Wikipedia adopted that style since it causes the least confusion.† Expressions of time are the same; 3:30 AM does not confuse Europeans in the least but 03:30 is quite unfamiliar to very many Americans and 03:30 AM is a bizarre and useless hybrid of military and customary practices. The best practice for expressing time in a transparent manner that causes the least confusion in our readership and draws the least attention to itself is obvious. Greg L (talk) 17:28, 31 October 2010 (UTC)
    Yes, for science-related articles, MOSNUM allows a number-delimiting style used in international scientific papers (1 234 567.89). However, since Wikipedia’s articles are far & wide directed to a general-interest readership and our articles are not papers in science journals (replete with scientific notation and two-factor units of measure with negative exponents, like a mass flow rate of 4.85625×103 kg·s-1 instead of 4,856.25 kilograms per second), that exception is often seized upon by editors for articles that are the least bit *sciencey* so Wikipedia can advertise and promote abstruse number-delimiting practices complete with hyper-advanced ways of expressing measures that couldn’t possibly be less accessible to a general-interest readership. But that’s a separate issue.

    I think too many young, would-be technical writers think they’ve done a good job by ralphing abstruse stuff like this onto pages because it looks advanced and they have mastered the techniques and feel proud. MOSNUM would do well to help steer such inexperienced editors towards technical writing techniques that promote clarity and make our articles as accessible as possible to a general-interest readership. Greg L (talk) 17:49, 31 October 2010 (UTC)

  • Well, the 12-hour clock does confuse some Europeans. (Maybe "3:30 AM" doesn't, but "3:30 PM" might, and "12:15 AM" definitely does, and for many "12:00 PM" is not merely confusing but actually unintelligible. In a train station in Paris a couple of years ago I asked a guy at the info desk what trains were leaving that evening at seven and I had to tell him "No, I mean 7 pm, that is 19!" because I had seen he was searching his computer for trains leaving the following morning.) I think it should be treated as an ENGVAR issue (i.e., use the format mainly used in written English in the country you're talking about, or just pick a format and do not change it without a style-independent reason if both formats are widespread in that country). A. di M. (talk) 18:12, 31 October 2010 (UTC)
  • I suspect the mild confusion some Europeans experience with 12-hour AM/PM pales by comparison to the mass confusion a 24-hour clock causes in most Americans. Writing occurred at 15:00 is exceedingly abstruse to Americans. Writing occurred at 3:00 PM confuses, I think, the smallest percentage of our readership.

    For instance, in American cinema, even in movies showing a military setting, like in A Few Good Men, the screenwriter will, for the benefit of the audience, find a way to have someone cluelessly paraphrase an expression like “catch an oh-six hundred flight” by asking “That flight to Cuba, was that 0600 in the morning? Sir?” That’s because 24-hour time is pretty much used in the American civilian world only be emergency workers and at hospitals. Back when Western Electric actually made stuff in America (before cheap Chinese, battery-operated clocks) 24-hour clocks were surprisingly ubiquitous in hospitals. Now, even there, 12-hour clocks are now the norm; medical workers just do the math mentally.

    I certainly see no point confusing 1% of Europeans at a tradeoff of confusing most Americans. Greg L (talk) 22:39, 31 October 2010 (UTC)

I'm glad to see this question (whether we should be using 24-hour time at all) discussed; this section is much clearer now. If Greg wants to make a subsection on which clock to use - as opposed to how to represent it - I will support the refactoring. The claim that 24 hour time is unintelligible seems exaggerated (after all, our own sigs use it); but it may be an unnecessary barrier. Septentrionalis PMAnderson 18:58, 1 November 2010 (UTC)
  • Thank you, PMA; I can get started.

    I’d hate to lose inertia on this, but my first step would be to suggest we turn over that slimy rock that nobody wants to turn over for fear of what might be found underneath: Whether it should be 3:30 p.m. or 3:30 pm or 3:30 PM. I would propose that we look towards the practices of the Associated Press since their manual of style is observed by virtually every subscribing newspaper in the English-speaking world and is what our English-speaking readership have become most accustomed to. The AP’s on-line version of their manual of style is $19 and I’ve never had an occasion to need it. I’ve gotten by whenever I had a question (such as how the AP abbreviates the months) by simply calling my local newspaper; the editors there seem happy to refer to their print copy (you can hear a metal drawer slide open) and look things up for me.

    Does anyone think we can resolve this formatting question by looking towards the AP(?), or might the resultant answer produce Turkish butt-stabbings and ArbCom inquisitions where tongues are torn out those dragged before the court? If we can agree to look towards the AP, then the next question is “what does it say?” If someone has a copy and can post what it says below, great. Failing that, I’ll call my local paper tomorrow. Greg L (talk) 01:06, 2 November 2010 (UTC)

    My impression is that capitals prevail in Leftpondia and lowercase prevails in Rightpondia, and that on both sides of the pond the periods are getting rarer than they used to be. A. di M. (talk) 01:34, 2 November 2010 (UTC)
  • The dots and the caps are detestable. Tony (talk) 07:30, 2 November 2010 (UTC)
    The point is not whether something is detestable, but – as Greg says – whether the readers are used to see it. (And in the rare instances when what the readers are used to depends on which side of the pond they are from, WP:ENGVAR applies, and if readers are accustomed to several different things WP:MOS#Stability of articles applies.) A. di M. (talk) 11:00, 2 November 2010 (UTC)
    Quite so. If the non-emotional content of "detestable" is "not Australian usage". however, we already have policy on this: leave it alone unless it violates strong national ties; this is a strong argument never to have a bot on the subject.Septentrionalis PMAnderson 16:04, 2 November 2010 (UTC)

When we update MOS:TIME, we should probably ensure consistency between it and the lead sentence in our 12-hour clock article, which says: ante meridiem ("a.m., English: "before midday") and post meridiem (p.m., English: "after midday"). Also in the Typography section it says It is technically correct to write "a.m." and "p.m." .... Mitch Ames (talk) 10:15, 2 November 2010 (UTC)

Searching this thing (see this) for "AM PM" appears that "9 a.m." is what the greatest number of style guides recommend. A. di M. (talk) 11:00, 2 November 2010 (UTC)
The [Reuters Style Guide] states: When referring to times first give the local time by the 12-hour clock (without using the words local time) and follow it with a bracketed conversion to a 24-hour clock time for a specified time zone, e.g. ... will meet at 10 a.m. (1600 GMT). Use figures except for noon and midnight. Use a colon to separate hours and minutes, 3:15 p.m. The Reuters style guide also makes the observation that "AM" is an abreviation for "amplitude modulation", while common sense dictates that "am" is a pronoun. This is good enough reason to use "a.m." to denote "ante meridies".
May I also take this opportunity to remind editors that a significant proportion of Wikipedia readers are British, Australian, Canadian, Indian, Pakistani, South African, New Zealander etc while a large number of readers whose native toungue is not English refer to the English-language version becasuse their own language version is not as extensive. Martinvl (talk) 13:30, 2 November 2010 (UTC)
Thanks for the info on Reuters’ practices. I’ll check this morning with my local paper as to AP practices, which are observed by very many English-speaking newspapers (no matter what country they are in).

I suspect too that Reuters’ advise on a parenthetical to GMT 1600 GMT is to ensure that newspaper editors in other time zones—even in other countries—can properly calculate the timing of an event and the expectation is that GMT will be stripped out before publication in a local paper. When I check with my local paper on AP practices, I’ll look into this too. Clearly, providing accurate GMT information to the editors of subscribing newspapers can be crucial when it is an article mentioning, say, the time at which a lunar eclipse will reach totality.

Practices are all over the map on Wikipedia: I see that one article, ‘2002 Bali bombings’ mentions local time using a 24-hour clock and gives GMT 00:15 Local time (17:15 GMT). That article appears to have been influenced by an editor who deeply believes in all-things-ISO. But, as exemplified by our ‘September 11 attacks’ article, the timing of an event in local time is all that is sufficient to convey the nature of the event; the one in Bali was shortly after midnight and 9/11 was the product of early morning flights. Unless one is a military officer planning flight arrival times of cruise missiles crossing times zones, I’m not so sure that time-stamping a placed article with GMT adds anything at all meaningful for a general-interest readership. Clearly, we as Wikipedians interested in time nuances like date delinking, UTC, GMT, ISO, etc. understand this stuff. But we also understand HTML and wikimarkup; a general-interest readership does not have nearly as much facility with such things.

As for en.Wikipedia being read by others for whom English is not their first language (like Pakistanis), I note that the the Početna-language Wikipedia article on 9/11 has the time that a flight crashed into a tower as U 9:03:11. While that is interesting, I see no need to shortchange our English-speaking readership by writing in a fashion where the writing style unnecessarily calls attention to itself. There is certainly some good that can come of exposing other cultures to the full measure of English-speaking practices. You may take comfort in that I am quite familiar with the fact that en.Wikipedia’s readership that speaks English as its first language includes New Zealand, British, Australian, and Canadian (and those kooky Brits, for whom the language is named after). That is, after all, why I proposed looking towards the Associated Press: subscribing papers are found all over the world.

All this discussion is good because it reminds us that to write a good guideline requires thoughtful attention to the exceptions that it ought to mention. For instance, if the article is on the time of a supernova, like ‘SN 1987A’ and one is writing about the time the neutrinos flashed clean through the earth and were detected by three neutrino detectors dispersed around the world, mentioning GMT might be more-than appropriate. Greg L (talk) 15:23, 2 November 2010 (UTC)

There are also world-wide political and military events - for example, V-E Day, which was May 7, except in the Pacific, where it is May 8; using GMT is one way to explain this. But these are not much more common than supernovas. Septentrionalis PMAnderson 15:53, 2 November 2010 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────

  • I spoke with my local paper, which is an Associated Press subscriber. The AP’s manual of style addresses expression of time of day just like that of Reuters, which is 12-hour clock, no leading zeros for the hour, a space separating the numeric value and the symbol, and the symbols are a.m. and p.m. I did not know this; I had always used “AM” and “PM” myself, but will abandon that and from hereon it will be 3:30 p.m. Given that English-language newspapers throughout the world use 12-hour time, this is what I propose for general use on Wikipedia as it is the writing style for expressions of time of day that least calls attention to itself.

    As I wrote above, articles that are of a military nature are no exception; the readership of our military-related articles are not military personnel going online to see what time they are due to report to their duty station. Quite correctly, our article here on ‘Death of Linda Norgrove’ states that the rescue attempt began at 3:30 AM and not 03:30 (though it should now correctly be a.m. and not the uppercase version).

    I also spoke with the editor about Reuters’ practice of mentioning (1600 GMT). He hasn’t seen such a thing in AP feed but agreed that it is likely to allow unambiguous conversion to local time for a subscribing publication. We also discussed the ambiguity—or at least confusing nature—of 12:00 p.m. He said their practice is to use noon and midnight.

    I’ll write up a guideline later today that addresses the fundamentals as well as some or all of the above-mentioned exceptions and inclusions (like time-critical astronomical events not closely associated with a particular spot on earth such as ‘SN 1987A’). I don’t know whether I will post it here first as a draft for talking points, or post it as a guideline with a copy here for continued discussion; either way, if there isn’t a community consensus for something, nothing sticks. And I often do best when I sleep on something and shave once before acting, so it may be some number of hours. Greg L (talk) 18:56, 2 November 2010 (UTC)

It’s UTC now, and has been since 1970. Don’t use GMT. It’s generally preferred just from what I’ve seen in talk pages anyway, and it takes a global view more into account. Wikipedia is global, too! I’ve always capitalised am. and pm myself. Maybe a petition to change this? I see am and pm capitalised often in Britain, too and it is the de facto convention in the US if I’m not mistaken.
By the way, it is most impressive that you actually spoke with your paper just for a silly quarrel on Wikipedia. You are that committed, wow. You might just deserve a barnstar for that, but I’ll see how this coversation plays out first.
Anyway, some people here think American readers have never seen 24-hour clocks before (Windows and Macintosh OS, anyone?), but I very much doubt that. The same goes for... people of certain ethnicities around the world who are used to the 24-hour clock. So yeah, I’m in favour of keeping both to avoid confusion but I don’t mind a 24-hour clock-only policy, without altering sources, of course. I think that would cause too much butthurt to a bunch of people in this thread, so forget it, I don’t want to start a flame war.
How abour recommending the use of neither clock, and “enforce” it for events at midnight? I’d assume most people would think in a respected encyclopaedia like this one, the am or pm would always follow the time given, unless the 24-hour clock was used. Therefore, I have no problem with the current “policy” and honestly don't see why anyone would.
Here is my proposition:
  • Usage of both clock formats is acceptable. No preference given in any article to either format.
  • For times at midnight, always use the 24-hour clock.
  • When using the 24-hour clock, have a leading zero (09:15, not 9:15) as it avoids some ambiguity over which time format is used in the unlikely event there is some confusion.
  • The am or pm ... symbol must always follow the time if time is given in the 12-hour clock format.
  • Because of this do not write time in prose, as in at five o’ clock, but use numerals (at 5 am) instead.
    When quoting from time written in words, clarify which clock format is used (We began advancing at half-past ten. as opposed to We began advancing at half-past ten [am].)
What do you think of my proposition? I tried to be as fair to everyone as possible, while still taking into account common sense and the people's views. --Γιάννης Α. | 19:51, 2 November 2010 (UTC)
At the very least, you should swap the order of the last two examples ("we began advancing at half-past ten [am]" as opposed to "we began advancing at half-past ten"), because otherwise it would confuse the hell out of people unable to distinguish red from green. Even I, who can tell them apart, was confused for a moment because I thought you were saying a thing and showing another, before noticing the colours.
As for "times at midnight", what would be wrong with just using midnight (provided the context already makes clear whether it's at the beginning or at the end of the day)?
BTW, FWIW, I've been in Ireland for nearly two months now and, while the 12-hour clock is pretty common here, I can't recall ever seeing AM or PM written with capital letters (except in text in all-caps and books written by Americans). A. di M. (talk) 20:12, 2 November 2010 (UTC)
Articles should be consistent. It wouldn't be appropriate to use a.m. and p.m. for most times, but switch to 00:00 for midnight. "Noon" and "midnight" should be used if needed in articles that use a.m./p.m. and "midnight" should be accompanied by whatever phrase is needed to show whether it is at the beginning or end of the day. Jc3s5h (talk) 20:19, 2 November 2010 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────

  • I agree with Jc3s5h: articles should be consistent.

    Γιάννης Α., you speak of using “common sense” in suggesting that any editor who wants to use 24-hour time should be free to do so. But it couldn’t be more well established now that 24-hour format is not used in daily life by either Americans or Canadians; clocks on their walls at home and at work and on time/temp digital displays in front of banks are virtually all 12-hour clocks (no leading zero). All regular digital-display clocks one would buy at a housegoods store say 3:00 in the afternoon and not 15:00. It is a simple fact that the 24-hour clock is unnatural to those who use only 12-hour in their daily life. Certainly many can *figure out* what 15:25 means, but many others would be thoroughly stumped as they stared at such an expression.

    That underlies why both Reuters and the Associated Press use 12-hour clock. Consequently, if someone abroad picks up an English-language newspaper in, say Morocco, and if that newspaper is a subscriber to AP or Reuters news feed, time is in 12-hour format. These practices followed by so many major English-language newspapers throughout the world are by no accident. Moreover, our readership for whom English is their first language but who are not Americans are perfectly fluent in a.m./p.m. 12-hour clocks. So…

    It makes zero sense for some of Wikipedia’s articles to mention that something happened at 15:30 while still other articles—totally randomly—use 12-hour time. You mentioned “people of certain ethnicities around the world who are used to the 24-hour clock” but they invariably have their own-language wikipedias and the last thing we need is for en.Wikipedia to have articles randomly using two different time formats depending on editor preference just because en.Wikipedia is so expansive, non-English-speaking peoples visit it.

    So if you believe 24-hour time is such a clear and natural way to express time in a manner that isn’t confusing and doesn’t draw attention to itself, I expect you to explain why all our articles shouldn’t be using 24-hour time. But an argument of “just let editors and ISO-fans do what they want” sounds like more of an accommodation to make certain editors happy and less so about concern for what is best for our readership.

    In my view, an attitude of “whichever way makes an editor happy” makes just as much sense as when Wikipedia—only a few years ago—had some computer-related articles using expressions like the computer came with 512 mebibytes of RAM while still other articles spoke of 512 megabytes of RAM. I find inconsistent practices to be bad ‘cess and a complete violation of Technical Writing 101. In the end, such nonsense proved to be the product of an extremely small cabal of editors who wanted Wikipedia to go contrary to the way the real world worked. I think Wikipedia has had enough of “either of two entirely different ways may be used depending on the editor’s personal preference.” Greg L (talk) 21:20, 2 November 2010 (UTC)

Two points - firstly us Brits use the 24 hour clock for rail and for air timetables. To prove this, look up the rail timetables between London and Manchester. I will let you find the URL - in that way you can't accuse me of cheating. You will almost certainly get a 24 hour timetable. Also, the BBC domestic TV service introduced a new weather map system a year or two ago. After experimenting with the 12 hour and 24 hour clock variants, they have settled on the 24 hour system because that is what the British viewers were happiest with.
Regarding the readership - I came across an interesting fact this evening - the wrold's largest English language newspaper is The Times of India - it publishes more copies than there are people in India who use English as their mother tongue. In India (as in South Africa), English is the language of mass communications hence many people will write in English even though it is not their mother tongue as they will get a larger readership that way. That is certainly the case in respect of South African articles.
This leaves the question - when should one use the 24 hour clock? I think that it is up to the editors concerned with the proviso that each article should be consistent and should reflect local usage. This would be consistent with both the use of language (UK or US English) and the use of units of measure (metric, imperial or customary). The definitve rules that I think should be applied across the board are:
  • No leading zeros on 12-hour times, leading zeros mandatory on 24 hour times
  • Ante meridies and Post meridies should be abbreviated "a.m." and "p.m." on grounds that "AM" means "amplitude modulation" and "am" is a pronoun verb.
  • When using the 12-hour clock, the terms "noon" and "midnight" should be used rather than "12:00 a.m." or "12:00 p.m.".
Martinvl (talk) 22:12, 2 November 2010 (UTC)
I have no doubt that American rail and air tickets also use 24-hour time. The fact that British airplane tickets do the same simply validates that Brits use 12-hour time in real life and is what Brits are far & wide most accustomed to at home and at work (which is something I already knew having spend a month in the British Isles). That rail and air tickets use 24-hour time is hardly a ringing endorsement for using it in prose in encyclopedias. You haven’t explained how using either format at random in Wikipedia’s articles best serves our readership. Greg L (talk) 22:40, 2 November 2010 (UTC)
A good reason to use both 12 and 24 hour style is to divide the pain of having to convert more equally to our readers. Why should we give in to one particular style, especially if it is not a world standard? When I see a time with a.m./p.m. I always have do to a double take and try to remember what each of them mean. When Americans see a time like 19:00 they have to think twice. We can share the burden by using both, depending on the subject or established style in the article. −Woodstone (talk) 04:55, 3 November 2010 (UTC)
Woodstone, you write of “divide the pain of having to convert more equally to our readers”. As far as I know, there is not a single English-speaking country where the digital clocks in homes and businesses read “15:00” at 3:00 p.m. The use of 24-hour time across the English-speaking nations is limited to things like time cards when shift workers clock-in, airline tickets, and the like. Could it be that you just like 24-hour time and think Wikipedia should expose the virtues of the system to our readers in hopes the grow more accustomed to it? Greg L (talk) 13:35, 3 November 2010 (UTC)
The clock on the bottom right corner of the screen on this computer at UCD right now reads "14:15". A. di M. (talk) 14:15, 3 November 2010 (UTC)
I have two digital clocks in my house (in English-speaking Australia) that show 15:00 - one on my oven and one on my video recorder. And another one on the remote control for my air-conditioner. None of these have the option to display 12 hour time. Mitch Ames (talk) 11:57, 4 November 2010 (UTC)
The article Cumbria shootings which uses the 24 hour clock is very much in line with British reporting. Please also visit [interactive map]. I therefore agree with User:Woodstone that the use of 12 or 24 hour clock depends on the context or article style. Martinvl (talk) 06:33, 3 November 2010 (UTC)
Well, you can point out internet web sites that date-stamp articles with UTC but I’m not going to play such silly games. You and I both know that ordinary digital clocks sold in housewares stores in England, Australia, Canada, the US, New Zealand, etc. do not display “15:30” in the afternoon and that 12-hour time is what all English-speaking peoples are most familiar with in their daily lives. Everyone understands “3:30 p.m.” but many people do not understand “15:30”. That is a simple fact. Period. Full stop. Airline tickets have been in 24-hour time since at least the 1960s but it is clear that 24-hour time is no closer to catching on in daily use now than it was in the 60s. Allowing advocates of 24-hour time do get their way is the same as caving to the crowd that advocated “kibibytes, mebibytes, and gibibytes” on any computer-related article they pleased. Though the practice made some of our futurist editors happy because they could do as they pleased, in the end, our readership suffered. Greg L (talk) 14:12, 3 November 2010 (UTC)
I am afraid this discussion may be suffering from the same misconception that some people have with metrication. As far as I know no culture uses the 24-hour clock in as radical a way as some seem to believe. In German and French, just like in British English, it is mostly absent from colloquial speech. In some cases such as appointments with a physician or buying a train ticket it's more likely to be used, but even in that context I have occasionally not used it in Germany. In fact, I have not seen any differences in use between German, French and British English except that Germans are more likely to write "15:00" and say "3 in the afternoon", where Brits would say and write "3 p.m."
The key thing in German is that TV schedules use the 24-hour clock. The article says the Radio Times tried this in 1934, but had to reverse the decision after half a year.
As to "everyone understands", I happen to have learned "12 p.m." as a special case, but I am sure most non-native speakers will get this wrong because it's such an odd exception. And I simply can't remember whether the exception also applies to "12:30 p.m." Hans Adler 14:49, 3 November 2010 (UTC)

←Hans, I know you're from Germany, but I think you significantly understate the usage of the 24-hour clock on the continent (and to a lesser extent in the UK). IMO, both 24- and 12-hour clocks should continue to be mandated on WP. Tony (talk) 15:29, 3 November 2010 (UTC)

OK, let's put it this way: My daughter is six. At the moment she only knows the 12-hour clock and is beginning to get proficient with that. At some point she will learn about the 24-hour clock and likely go through a short stage of using it exclusively. And then she will be fully proficient, using the 24-hour clock in writing almost all of the time, and the 12-hour clock in speech most of the time. The difference between German and British English is that the Brits write the same way they speak, i.e. they mostly use the 12-hour clock even in writing. It's the same mechanism as for weight units. The German standard unit in writing has long been the kilogramme, even at markets. But until recently older people selling or shopping at a market usually thought and spoke in terms of a (metric) pound of half a kilogramme. I believe my mother still speaks of "half a pound of butter" when referring to the standard pack of 250 g. And some people still use the (metric) hundredweight (Zentner) of 50 kg in speech, though usually not in writing.
Of course none of this is very relevant to the MOS. My main point was that we shouldn't think about it in terms of some polarisation or even evangelising. The changes on the continent happened very gradually with a large number of intermediate steps. Even in France, where they started by changing everything overnight, then reversed it and introduced some of the innovations again in a piecemeal fashion. Hans Adler 16:00, 3 November 2010 (UTC)
  • Never mind Hans; you’re going to do just as you want to do no matter what. I can see I am wasting my time here. We all know that virtually every English-speaking individual understands 12-hour clock perfectly well. We all know that a significant percentage of people don't understand 24-hour clock and that many find that 24-hour time to be unusual and a writing style that calls attention to itself. We all know those simple statements to be absolutely true. But we will just *teach our readership to grow comfortable with 24-hour time*, won’t we?

    This simple understanding of what confuses least is why both AP and Reuters require “3:30 p.m.”

    We could solve this just by looking to The New York Times practices as well as the Chicago Manual of Style as well as the AP’s manual of style. But Wikipedia is going to educate the masses to a Better Way™©® because we have believers who hide their reasoning behind absurd statements like how other cultures (who, by the way, have their own-language Wikipedias) somehow magically have such standing that we can now confuse a significant fraction of our English-speaking readership—or at least make give their brains a pause as they stumble across a rare time format they seldom use in real life. BTW, the parenthetical Reuters (1600 GMT) that they add to their articles is for the benefit of subscribing editors in various time zones so they can convert to 12-hour time in their time zone.

    I can see that a fundamental consideration like universal understandability and following the practices of all the leading manuals of style doesn’t matter. We all know why: because having a consistent practice that confuses absolutely no one takes a backseat to “but… 24-hour time is logical and superior and I like it and I want to use it in articles I write and readers will understand it after they stumble and stare at it and somewhere someone is gonna have to be taught 24-hour time and, by God, I’m gonna teach them on any article of mine they land on.” It’s most unfortunate that Wikipedia’s MOS and MOSNUM suffers from this phenomenon. I see that I am utterly wasting my time and this is just another “mebibyte” thing where making some editors happy takes a back seat to the needs of our readership.

    I was tempted to start a straw man poll asking whether editors agreed with the premise that 12-hour time is universally understood by all peoples for whom English is their first language and that 24-hour time is not used in daily life by most English-speaking people. I’m not so sure everyone could have seen past their personal biases to have answered that honestly. Since intellectual dishonesty disgusts me and so too does imposing personal biases in the overwhelming face of how all English-language manuals of style call for 12-hour time in order to communicate clearly, I’m dropping this. Greg L (talk) 16:10, 3 November 2010 (UTC)

You are concentrating so strongly on polarising this trivial matter that you didn't notice that I mostly agree with you. But I guess it was at least in part my fault for being too verbose. I'll try to be shorter and clearer:
  • In practice there is very little difference between the various variants of English. In the vast majority of cases only the 12-hour clock really makes sense.
  • Some specific contexts such as timetables, military or police may prefer the 24-hour clock in some variants of English. [1]
  • As the example of Germany shows, the situation is unlikely to change any time soon, so it wouldn't make much sense to push the 24-hour clock into contexts where it is not common.
  • Every time of the form "12:xy" in the 12-hour clock is potentially confusing to a non-negligible fraction of our readers, who do not know how "a.m." and "p.m." are used in those cases. In the few cases when the question "day or night?" is not answered by context, we need a flexible, unobtrusive approach such as substituting by "noon" or "midnight" where appropriate, or adding some redundancy. Hans Adler 16:32, 3 November 2010 (UTC)
  • There are two vastly different cases for time of the form 12:j a.m. or p.m., where j is an integer. The first case is 0<j<60. Granted, some people slept through elementary school classes on how to tell time, or don't speak English as their first language, but most people should understand this, especially considering the many documents that use 12:01 a.m. as a mechanism to avoid the ambiguity associated with midnight. The second case is j=0, in which case both NIST and the National Maritime Museum state 12:00 a.m. and 12:00 p.m. are ambiguous.
Furthermore, there is no grammatical way to combine a time such as 12:45 a.m. with "midnight" in a misguided effort to make it less confusing to those who don't quite know how to tell time in English. That is, it is ungrammatical and contradictory to write something like "12:45 a.m. midnight". If one wishes to avoid 12:45 a.m. and still be consistent in an article, 12 hour times would have to take a form such as "12:45 in the morning" or "12:45 in the afternoon". Thus, for articles that use 12 hour times I would advocate forms such as:
  • 12:45 a.m.
  • noon
  • midnight November 3/4, 2010
Jc3s5h (talk) 17:10, 3 November 2010 (UTC) modified 19:34 UTC.
Well, then Jc3s5h, you just go ahead and use your 24-hour time in your articles. I could not possibly care less what you do. But with regard to your inane there is no grammatical way to combine a time such as 12:45 a.m. with "midnight", as if that is somehow a compelling reason to use a format that pretty much every professional English-language publication directed to a general-interest readership sees fit to eschew: you are entitled to your opinion. It’s clear to me that you hope that Wikipedia might in some small fashion Lead the Way To a New and Brighter Future of Better, More Logical Practices©™® that has somehow escaped the pros. Isn’t that amazing how you figured out a better way to communicate that escaped the pros? Are you proud? But please spare me your justifications for doing as you do because I find them to be inane facades to masquerade the fact that you want to use Wikipedia to educate our readership to a better way of telling time, one awkward instance after the next. Greg L (talk) 18:51, 3 November 2010 (UTC)
Are there any good-faith differences of opinion on any subject? Art LaPella (talk) 19:20, 3 November 2010 (UTC)
Greg L wrote You and I both know that ordinary digital clocks sold in housewares stores in England, Australia, Canada, the US, New Zealand, etc. do not display “15:30” in the afternoon. I did a quick check around my home this evening (I live in the south of England). I looked at the digital clocks in my car, on my video recorder, on my wife's digital radio, on the central heating timer, on the cooker, on the telephone and on my bedside cupboard. The bedside clock had a switch to select one format or the other, the central heating timer had a 12 hour display - all other gadgets had 24 hour displays. I must therefore reject your assertion. Martinvl (talk) 19:31, 3 November 2010 (UTC)
That is very surprising. Very well, I accept that things have changed since I’ve been in the British Isle. So let me ask you this Martinv: Besides looking archaic or quaint to you, does 3:30 p.m. confuse anyone there in England for even a half-second? Because I absolutely guarantee you that occurred at 15:30 is not even understood by many Americans and Canadians (setting aside the issue of what looks momentarily awkward to the rest). And for that percentage that do *get* the concept, they still have to do mental math to relate to the expression so they can understand that the expression means 3:30 and not 2:30 in the afternoon. Clearly, there is a ‘people factor’ at play here that must underlie why the AP and Reuters both specify that their reporters and editors use 12-hour. My hunch is that the prevailing view of the experts is that whereas not every English-speaking country uses 12-hour clocks, all English-speaking countries understand it. Greg L (talk) 21:55, 3 November 2010 (UTC)
It makes perfect sense if you consider that the UK and Ireland aren't so big compared to the rest of Europe. Just one version of these goods arrives in Rotterdam or London in large numbers, and gets distributed to all of Europe. Hans Adler 23:10, 3 November 2010 (UTC)
All English-speaking countries can understand 24-hour clocks, too. All the Americans (not Americans in Britain) I know,&emdash;off the internet, this is&emdash;are well aware of 24-hour clocks. There is no confusion with those. It is pure logic that there are 12 hours in a day and I am sure all Americans can understand that. Your forgot to state in your “point” is that the Associated Press is an American news service and that Reuters mainly markets its news to Americans. This is why I support letting editors do whatever on earth they want. Editors are the people we are looking at here, as readers will be familiar with both clocks, I believe. Editors, however, may be used to one system or another and will write articles using one system that someone like Greg may get really pissed off(sorry, but it’s true) about. So let’s let everyone do whatever they want, except for using a.m. and p.m. for midnight events. I still don’t know how that works, but then again, I’m not used to using the 12-hour clock very much myself as I prefer writing times in the 12-hour clocks format(BTW, I live in England). This isn’t important enough to actually make someone change time formats because they would think readers are uncomfortable with them. Finally, Greg is implying that editors might completely ignore the a.m. and p.m. when writing times. This would be the only source of confusion for anyone in my view. Would any editors forget the symbol at the end? No, they would not, because even Americans, when writing time in numerals add a.m. and p.m. at the end of time presented in the 12-hour clock format. Oh, and sorry for repeating myself a few times there, which I think I did. --Γιάννης Α. | 22:50, 3 November 2010 (UTC)
“Can” understand 24-hour time is not the point. It never has been and never will be in proper technical writing. Greg L (talk) 23:01, 3 November 2010 (UTC)
Never will, in a bazillion years? Oh, we might. Something like stardates might be better. Art LaPella (talk) 23:25, 3 November 2010 (UTC)
I’m saying that whether or not Americans can learn 24-hour time by using it here in an “Oh… didn’tch know??”-fashion (the same thing we did with “mebibytes” for three years) is an improper basis upon which to formulate best technical writing practices. That’s exactly how proponents of “kibibyte and mebibyte” thought: we will just use it and readers will figure it out. The basis for both the AP and Reuters using 12-hour time is it *is* understood in England and elsewhere (they understand 3:30 p.m.) and it is the only time format universally understood by Americans and Canadians (many of whom have no flying idea what 15:30 means). For Wikipedia to march off into the bushes all fat dumb & happy with a practice that is 180 degrees out of phase with what the professionals do when writing for an international, English-speaking readership is unwise. I won’t address your “stardate” point directly since you either didn’t understand my point or you are being absurd for some weird reason. If you can’t be serious, then you best tune out. Greg L (talk) 23:52, 3 November 2010 (UTC)
The serious point is that you clarified the obscure words "never will", or more exactly you didn't repeat them. Art LaPella (talk) 00:49, 4 November 2010 (UTC)
  • I'll avoid an overly long comment, to this already painfully long thread, and simply voice my support for adherence to the AP and Reuters approach, using 12-hour time, for the reasons articulated by some above.--Epeefleche (talk) 04:15, 4 November 2010 (UTC)
  • A large portion of the US general public is unfamiliar with 24 hour time. People with experience in science or the military understand 1730 but it is not commonly used in everyday life. (I even learned to tell time with bells in the navy.) Over half the readers of English Wikipedia are from the United States (Wikipedia page views by country.). In general readership articles, the time format should be one that most Americans readily understand. I noticed that the British have not updated Big Ben to 24 hour time, it is still using Roman Numerals. -- SWTPC6800 (talk) First watch 4 bells, 4 November 2010 (UTC)

This debate is all very interesting, but can I please remind everyone that my original question - as per the section heading - was about the leading zero in 12 hour clock times, not about whether we should use 12 or 24 hour clocks. Can I suggest that we at least temporarily defer (or, better, move to a different section) the question of whether to use a 12 or 24 hour clock, and address the original question. Given that some articles do use a 12 hour clock - and we apparently have no clear consensus to change that - should we mandate "no leading zero", or should we state explicitly "leading zero is optional"? (I assume that not many people will want to mandate a leading zero.) As per my initial post, I still think that the MOS should state one way or the other, as it does for 24 hour clock. Thanks, Mitch Ames (talk) 12:10, 4 November 2010 (UTC)

Sifting through the entire discussion, I have not seen any convincing arguments in favour of having leading zeros on the 12 hour clock and a number against. In particular, neither Reuters nor AP use leading zeros; furthermore the absence of leading zeros helps to distinguish the 12 hour clock from the 24 hour clock. This suggests to me that leading zeros should not be used when displaying time using a 12-hour clock.
May I also propose that the rule for 24 hour clocks be modified to make leading zeros mandatory. The rationale is that this will simplify differentiation between 12 and 24 hour clock times. Martinvl (talk) 12:37, 4 November 2010 (UTC)
Does anybody have any objection if I amend WP:MOS and WP:MOSNUM to reflect a preference for prefixing single digit hours with a leading zero when expressing time using the 24-hour clock but to council against doing so when using the 12 hour clock? I plan to use the word "should" rather than "must". Martinvl (talk) 08:55, 6 November 2010 (UTC)
I agree with Martinvl's proposal to recommend leading zero for 24h time, no leading zero for 12h time. Mitch Ames (talk) 11:42, 6 November 2010 (UTC)

Proposed clarification to Calendars section

I propose a slight change to the wording of the Calendars section, as follows, with my additions underlined:

  • Dates before the adoption of the Gregorian calendar (on 15 October 1582, or the relevant adoption date for the country of interest) are normally given in the Julian calendar...
  • Dates of events in countries using the Gregorian calendar at the time of the event are given in the Gregorian calendar...

Without these additions the literal meaning of the sentences doesn't necessarily match what I believe our intentions are. Mitch Ames (talk) 11:24, 7 November 2010 (UTC)

I think the de-facto standard for events happening after the Gregorian calendar was introduced but before it was adopted in a particular country is to give both Old Style and New Style dates in the lead and state in a footnote which calendar will be used in the rest of the article, as in the current revision of Isaac Newton. A. di M. (talk) 12:20, 7 November 2010 (UTC)
I think A. de M. has summarized what is usually done for articles about America and Europe. I wouldn't venture to guess what is done when the subject is other parts of the world. Jc3s5h (talk) 14:45, 7 November 2010 (UTC)
Just because something is a de facto standard doesn't mean that we shouldn't include it in the MOS. Mitch Ames (talk) 10:14, 8 November 2010 (UTC)
Real life isn't simple. Just look at this weekend, which celebrated the 305th anniversary of Guy Fawkes' Gunpowder Plot on 5 November 1605 [O.S.] last Friday (5 November 2010 N.S.), but today celebrates the Great October Socialist Revolution of 7 November 1917 whose participants and opponents were observing the Julian month of October. Similarly, the calendar on George Washington's Birthday (always celebrated, before the invention of Presidents' Day, as February 22, 1732) read 11 February 1731. I certainly support the spirit of Mitch Ames' suggestion, so long as it's a recommendation rather than a binding rule, but writing it might be difficult. —— Shakescene (talk) 18:49, 7 November 2010 (UTC)
Not to mention the Oktoberfest, which is almost completely in September. A. di M. (talk) 01:28, 8 November 2010 (UTC)
... and the Queen's Birthday - which next year they are changing in Western Australia, so we can sing Happy Birthday to her when she comes to visit :-) Mitch Ames (talk) 10:04, 8 November 2010 (UTC)
Seriously though, when people celebrate or commemorate the "anniversary" of a date is irrelevent from our perspective. (In Australia it is common to move many "anniversaries" to the nearest Monday, so we can have a long weekend.) We're talking about describing the date than an event took place, (usually) independently of when its anniversaries are celebrated. Mitch Ames (talk) 10:11, 8 November 2010 (UTC)
Well, it's not the parties that concerned me; it's just that in post-1582 events whose participants were still observing the Julian calendar, it would be odd (and confusing to readers) if we wrote about the events leading to the Gunpowder Plot in relation to the Gregorian date being observed on the Continent rather than the Fifth of November, whereas the article on (say) Alexander Kerensky describes the events surrounding the October revolution in relation to 7 November 1917 (N.S.) Similar questions come up with people whose lives straddle a calendar change, e.g. William Shakespeare (christened in late April 1564, died 23 April 1616 both O.S.), George Washington (see above), Kerensky, or, indeed, Pope Gregory himself. These difficulties are hardly insurmountable and the different approaches are probably covered by your proposed word "normally"; but while tedious, I think an article about something after 1582 and before a country's final adoption of Gregorian dates should indicate at least once (and early) the calendar being used (a quick scan of the Gunpowder Plot article showed no such flag to me, although I surmise from all kinds of circumstantial reasons that it's using Julian dates). —— Shakescene (talk) 10:31, 8 November 2010 (UTC)

Discourage use of the ordinal suffix (e.g. th) in fractions

Should we make it clear in WP:Mosnum#Fractions that the ordinal suffix should not be used in numerically written fractions ( e.g. 1/100 is correct - not 1/100th)? Clearly if someone is spelling out one one-hundredth they will use the suffix, but I think we need to give clear guidance on whether they should use it when the fraction is written as a number. It is made clear in WP:DATESNO that we don't use it for dates.

I suggest we add the following to the Fractions section:

* Use of the ordinal suffix (e.g., th) in fractions expressed in numerals is discouraged (e.g. 1/100, not 1/100th)

Make sense, or not needed?  7  04:09, 4 November 2010 (UTC)

Thank you very much for the courtesy of notifying me about this discussion 7 but I am not an MOS expert, so I'll defer to the opinions of the local experts. The suffix is also used in article titles such as: Rule of 1/1000th common ancestry so we have to make sure we cover all occurences one way or the other. Take care. Dr.K. λogosπraxis 04:42, 4 November 2010 (UTC)
Perhaps the language should make it clear it's talking about fractions expressed in numerals, such as 1/100, as opposed to words, as in one one-hundredth or seven one-hundredths or five sixths of the world's population. —— Shakescene (talk) 20:29, 4 November 2010 (UTC)
Agreed. Added expressed in numerals above.  7  23:37, 4 November 2010 (UTC)
It should also be noted that there are cases where, e.g., 1/6th and 1/7th, not expressing fractions, are not incorrect, unless they go against some properly-agreed consensus and convention of the Wikipedia:WikiProject Military History. See Duke of Wellington's Regiment#6th and 7th Battalions and Duke of Wellington's Regiment#7th Battalion. So any recommendation, no matter how forceful, should not be so unqualified as to justify someone unleashing an automated unmanned 'bot. —— Shakescene (talk) 02:53, 5 November 2010 (UTC)
It makes sense to me. The suffix is not needed as it's just the way or reading out fractions. 1/2 is "one half", 1/4 is "one quarter", 1/5 is "one fifth", 2/5 is "two fifths", i.e. two 1/5s. All apart from 1/2 and 1/4 use the same word as the ordinal ("third", "fifth", "seventh", etc.) but they are quite different things. I would say is should be stronger than discouraged, though it might be good to get a reference for that from e.g. a style guide.
And I had a look at Rule of 1/1000th common ancestry. Going by both the text of the article and the reference it should be Rule of 1/1000 common ancestry so I moved it to there.--JohnBlackburnewordsdeeds 21:06, 4 November 2010 (UTC)
Yes check.svg Done I've gone ahead and implemented the change. If concerns arise feel free to undo.  7  03:33, 10 November 2010 (UTC)

Problem with current era notation (Anno Domini/Common Era) guideline

While I understand that this issue has been visited more than once before, and that some may be irritated with digging it back up, but I feel that the indecisiveness regarding whether to favor either AD or CE on Wikipedia needs to end. Or, at the very least, I think we need to form a compromise that would dictate the use of one era notation for certain topics only (ex. Common Era for non-Christian religion-related articles), and the other notation elsewhere. The current recommendation—to use whichever notation you'd like as long as you are consistent within an article and do not arbitrarily change from one notation to another—isn't working. There are a couple of problems with it that I'm coming across quite a bit lately:

  • As seen with this diff, editors will often make a contribution using the B/CE notation when the article they are editing uses the BC/AD notation throughout. Often, these edits are not caught or reverted and we end up having an article with inconsistent era notations. This could easily confuse the average reader who might be unfamiliar with the Common Era notation, and who might then understandably assume that 500 AD and 500 CE do not indicate the same year, since they are both being used inconsistently in the same article, without explanation.
  • Needless edit wars over which notation to use, as well as completely arbitrary conversion of random articles from BCE to BC or vice versa that may go unnoticed for months. Here are just some recent example diffs where I was reverting arbitrary era notation changes and then informing the user about the WP:ERA guideline:
Blombos Caveoriginal edit, my reversion, second edit 2 days later, my second reversion
List of solar eclipses for [X] century BC1st c., 2nd c., 3rd c.... 20th c.
Notifications to users[2], [3]
  • Ugly compromises, as with Jesus, where both notations are used simultaneously at each instance to alleviate constant edit warring over which to use. An argument could be made that the use of both notations in Jesus helps to educate readers about the CE notation, but it's more cluttered and confusing than anything else.

I'm hoping to gather some responses and comments about these issues, and perhaps we will be able to flesh out our thoughts on what type of solution we want to go ahead with, if any.

My proposed solutions would be, either: (1) AD/BC ubiquitously; (2) CE/BCE ubiquitously [this would require moving all BC year articles, ex. 1 BC, 1st century BC]; or (3) One notation for specifically outlined articles, and the other notation for the remainder of articles. Thank you for reading and replying. — CIS (talk | stalk) 16:11, 5 November 2010 (UTC)

The project page says "No preference is given to either style.", and I wouldn't expect that to change in the forseeable future. As a practical matter, noting the existence of the {{US English}} and {{British English}} templates, I'll suggest that a template something like {{Year style}} might take e.g., AD, BC, CE, and BCE as alternative parameters to define the consensus style for an article, could produce an informative hatnote, could appropriately categorize an article, and could provide a hint for bot-edits or assisted-edits to conform nonconformant articles to the expressed consensus style preference. Wtmitchell (talk) (earlier Boracay Bill) 01:47, 7 November 2010 (UTC)

Discussion

This isn't going to find wp-wide concensus in any form, that's why we leave it up to individual articles. However: the idea that the articles about aspects of one religion or sect should be written only by and for proponents of that belief are not only deeply POV but also fail to address their full audience. If anything, I would say those editors should avoid using AD/BC most carefully in the religion-based articles. CE/BCE serves just as well without alienating any thoughtful readers. Save the AD/BC for use in the massive majority of non-religious topics such as movies, television, language, archeology etc. There hardly anybody will be bothered by them.
The issue of inconsistent choices by editors won't go away, but a talkpage header akin to {{British-English}} and {{American English}} can limit disputes by putting the chosen option for a given article on the record. LeadSongDog come howl! 18:18, 5 November 2010 (UTC)
I agree that it would be a better idea to use CE/BCE for all religion-related articles — Christianity, Judaism or otherwise. I've edited my original post to reflect that. What I had suggested in my original post was just an example of a possible compromise, not the only compromise I would support. I'm simply trying to strike up conversation here about what other editors think of these issues. As for the talkpage header you suggested, that could be a good idea, but I doubt any of the editors making the notation changes bother to visit the talk page beforehand. — CIS (talk | stalk) 18:26, 5 November 2010 (UTC)
I doubt that this will be of much use or relevance, but I subconsciously read [B]CE as [Before] Christian Era before making (when necessary) the mental adjustment to [Before] Common Era (and I don't happen to be a Christian; it's just that "Before Christ" to "Before Christian Era" is an easy conversion for my mind to make). There's nothing to stop Christian readers and editors from doing the same. That's not in itself an argument for replacing BC/AD with BCE/CE in any particular article or class of articles, but it might lessen the feeling by some Christian editors and readers that such a change need be necessarily positively hostile to, or dismissive of, their beliefs and world-view. —— Shakescene (talk) 18:59, 5 November 2010 (UTC)
I suppose that's a fair point, and for the moment I'm not strongly in favor of either notation personally, although I've traditionally been in favor of AD/BC. My arguments for that notation would include (1) they are the original notations; (2) as with Wednesday, Thursday or January, most modern terms in everyday use have religious origins. The only reason AD/BC tend to be singled out as being potentially "offensive" or "sectarian" while the religious names of the days of the week are not, is because Christianity is still a living religion; (3) BCE/CE do nothing to remove the era's bias, they only mask the religious reference; the dividing epoch of human history in the international calendar is still a 6th-century monk's estimate of the year of Jesus’ birth; (4) BCE has a bulky three letters rather than the simpler two, and it sounds confusingly similar to CE when spoken, while AD and BC look and sound comfortably different.
But all arguments for AD/BC aside, I would still prefer a ubiquitous use of BCE/CE over continuing with the current guideline. — CIS (talk | stalk) 19:53, 5 November 2010 (UTC)
  • In writing, using the politically-correct BCE isn’t unusual. But I’ve seen that it is often avoided in narrated TV documentaries and science shows, “Bee Cee EEE” out-loud draws attention to itself, is distracting, and takes the viewer’s mind off the message point and momentarily focuses it on the way information is being conveyed. Since there is nothing wrong with AD/BC, I prefer to use the BC myself as I prefer to use a writing style that is most transparent and least draws attention to itself—which is a hallmark of all good technical writing. Greg L (talk) 17:49, 11 November 2010 (UTC)
  • Here are my thoughts: Firstly, using BCE/CE on Christian articles seems completely illogical. BC/AD are the Christian form of BCE/CE, so if Christians can't use BC/AD, nobody should. Secondly, Since Wikipedia is a global project, using BC/AD could be offensive to non-Christians. Thirdly, BC stands for Before Christ because people used to believe Jesus was born in the Year 0 or AD 1. Modern historians do not recognise a Year 0 and now think Jesus was born around 5 BC (if he actually existed). So according to the notation, Jesus was born 5 years before Jesus was born. The idea that Jesus was more likely born before AD 1 was one of the reasons historians introduced BCE and CE. BCE and CE are more secular and make literal sense without having to change the year we live in. McLerristarr | Mclay1 20:36, 11 November 2010 (UTC)
    • Yup. That’s the reasoning underlying why BCE was invented alright. Arguments about who Jesus really was, when he was really born, and whether he rode a Schwinn bicycle or a donkey to work, or meets Zeus and Wiley Coyote in the morning while punching in on the ol’ time clock are all immaterial in my book and has no bearing on whether “10 BC” is somehow more or less ambiguous than “10 BCE” since they both mean precisely the same thing.

      The only difference between the two is BCE is the form for elementary schools and those who want to trip all over themselves to be as inoffensive as possible so that even Oprah and Jerry Springer couldn’t find someone crying all over their Hostess Twinkies at 4:00 p.m. over how they are so deeply offended for [yadda-yadda] reason.

      I couldn’t really care less one way or another except for noting that BCE hasn’t caught on real well in the real world and that’s why the spoken version of it sounds so odd that it is seldom, if ever, used in narration on TV documentaries—no matter how modern they might be. I suppose those Jerry Springer whiners can’t fuss over TV science documentaries using “BC” because they don’t watch those types of shows (busy watching Dukes of Hazard). IMO, use of BCE appears awkward in written form as well and unnecessarily draws attention to itself and that’s why I would never use it myself as I find it interferes with transparent communication of thought.

      Since everybody has an opinion on this and Wikipedia’s MOS is in a near continual state of chaos as different volunteers here spout about how Their Way Is the Right Way©™®, it’s unrealistic to expect any change to the current guidelines (everyone has a half-baked reason to do as they do). So it would be best if we all just dropped it since, as in previous occasions this issue has come up, it’s not going to go anywhere. You are free to do as you like. Greg L (talk) 22:59, 11 November 2010 (UTC)


      P.S. I’d normally suggest we just do as the Associated Press, Reuters, and the Chicago Manual of Style suggest. However, that would no-doubt be met with the observation that en.Wikipedia—unlike all the other-language Wikipedias that suck by comparison—is an especially International encyclopedia and how things are done way differently above 12,000 feet of altitude in Tibet. Greg L (talk) 23:17, 11 November 2010 (UTC)

      • I use BCE and CE, not because I don't want to offend people – I'm proudly politically incorrect, but because AD and BC offends me. We've been using out current dating system for a very long time now so changing the numbering would be stupid but the fact that we still use something like "Before Christ" as an international standard in our modern world just seems silly. McLerristarr | Mclay1 05:58, 15 November 2010 (UTC)
Honestly? As an atheist, BC/AD offend you? Do Tuesday and Wednesday offend you too because they're named after pagan deities? I'm an atheist, and BC/AD come nowhere near "offending" me, nor do the nonsecular names of the days of the week. If you are truly offended by BC/AD, I'd hate to see how offended you would be by something that is actually offensive. — CIS (talk | stalk) 13:49, 15 November 2010 (UTC)
There are different levels of offensiveness. On a scale of 1 to 10 with 0 being inoffensive and 10 being "I now have to kill you", BC and AD are a 2. There are no alternatives to the days of the week but there is a perfectly usable alternative to BC and AD. McLerristarr | Mclay1 13:58, 15 November 2010 (UTC)
So if there were secular alternatives for the days of the week, such as Firstday, Secondday, and so on, you would support using them? Because I don't see any inherent problem in the religious connotations of these types of everyday usages. If we secularize and sanitize everything that has a religious meaning, nothing will have any cultural significance anymore. We're getting a bit off topic and personal, but I just thought I'd put it out there — that I am indeed an atheist, but I see no problem in religiously-influenced commonalities. This is the Western world, and like it or not, Christianity and Roman paganism are a large part of our history. I don't think we should erase references to them in our culture for no reason other than that people are "offended" by the reference. But I suppose that if things like this are all we're worried about today, that's a good thing. You say that you consider yourself to be "politically incorrect", but what is the purpose of BCE/CE other than to be politically correct and unnecessarily euphemistic? Common Era and Anno Domini both refer to the same 6th century attempt to date Jesus’ birth; one just doesn't acknowledge that. — CIS (talk | stalk) 14:45, 15 November 2010 (UTC)
Yes, well as you say, these things are part of our life and unavoidable, so I'd rather live in the year 2010 and disambiguate it from the other 2010 by using abbreviations that recognise it's a historical numbering system rather than still a religious numbering thing. I suppose the German pagan days of the week don't bother me because nobody is a pagan anymore so I don't feel I have to compete against it. The fact that we're having this conversation proves that everyone has a different point of view and the MOS guidelines should not be changed. McLerristarr | Mclay1 01:40, 16 November 2010 (UTC)

Wishful thinking about standards - not a serious suggestion!

This isn't a serious suggestion - because I know that it won't be accepted - but I sometimes think life would be much simpler if we could just use ISO 8601's expanded representation. Ie, don't specify AD/CE or BC/BCE, just use +YYYY or -YYYY. Despite all the recent arguments about how most of our readers couldn't cope with a "standard" way or referring to dates/times, one can't help but feel that a consistent, unambiguous way of refering to dates/times could only be good thing. Of course to be strictly compliant with 8601 (clause 4.1.2.4) we'd have to get our readers to agree first! As I said, this isn't a serious suggestion, but one can dream ... Mitch Ames (talk) 00:59, 6 November 2010 (UTC)
You could have your own religious vision, in which you dream that the Creator, Almighty or Supreme Being has decreed that the only divinely-accepted calendar, imperative on all pious true believers who trust in Salvation, is ISO 8601, and that any other method of recording the passage of time is the work of the Devil's agents or of proud, foolish mortals who know not the limits of their own puny powers. Then, of course, the followers of all other religious beliefs or unbeliefs would suddenly undergo a Damascene conversion and abandon their foolish calendar wars to adopt The One True Calendar.—— Shakescene (talk) 01:28, 6 November 2010 (UTC)
I wish proposing ISO 8601 notation was like poker, where there is an ante. The ante ought to be reading the standard before making the proposal. Jc3s5h (talk) 01:29, 6 November 2010 (UTC)
ante meridiem, perhaps?—— Shakescene (talk) 01:34, 6 November 2010 (UTC)
Actually I have a copy of, and have read, ISO 8601:2004(E). OK, I haven't read the whole thing from beginning to end, but I have browsed through it, and read some parts of it in detail, as I needed to. Jc3s5h, is there some particular part of it that you think is applicable in this context and that I haven't, but should read? Some part that makes the standard completely inappropriate here? Apart, of course, from the requirement for our readers to agree to us using it – (clauses 3.2.1 "... dates preceding the introduction of the Gregorian calendar ... should only be by agreement of the partners in information interchange", 3.5 "By mutual agreement ... expand the ... year, which is otherwise limited to four digits. This enables reference to dates ... before the start of the year [0000] ...", 4.1.2.1 "Values in the range [0000] through [1582] shall only be used by mutual agreement ...", 4.1.2.4 "If, by agreement, expanded representations are used ..., [eg] +-YYYYY" – which I did mention in my "proposal". As I said, my proposal wasn't serious, but I am interested in standardisation, and I would like to know if there is any other (than "mutual agreement") reason why ISO 8601 would not be appropriate. Mitch Ames (talk) 02:55, 6 November 2010 (UTC)
The signed integer convention is very unfamiliar to the general public; I think very few people would realize that −45 is not 45 BC but 46 BC (due to the lack of a year 0 in the BC/AD system). A. di M. (talk) 12:35, 6 November 2010 (UTC)
I wouldn't have guessed that −45 is a year at all. Is that standard for computer data or something? It surely wouldn't work for communicating information to the general public. Art LaPella (talk) 15:45, 6 November 2010 (UTC)
As mentioned earlier, ISO 8601 is an international standard covering the exchange of date and time-related data. It includes negative years, instead of BC or BCE. However, you wouldn't be expected to "guess" this, because 8601 explicitly requires mutual agreement between sender and receiver before using negative years. Hence my note that we couldn't really use 8601 because to do so would require all our readers' prior agreement to use negative years - which is clearly not practical. Mitch Ames (talk) 01:05, 7 November 2010 (UTC)

I've just moved this sub-discussion about ISO 8601 and negative years into a separate sub-section, so as to separate it from any serious proposal. As I said in the first line of my original post: This [use of ISO 8601] isn't a serious suggestion. (Nonetheless, it is interesting - well I think so, anyway.)

In addition to the agreement requirement there are also
  1. the requirement to always use the Gregorian calendar, even when stating dates related to events that occurred in countries where the Julian calender was in force, such as the Gunpowder plot in England, 5 November 1605 Julian
  2. the fact that 2009-10 means October 2009, not 2009 through 2010. Jc3s5h (talk) 01:20, 7 November 2010 (UTC)
The requirement to use the Gregorian calendar for events that occurred where that calendar was not used - even after it had been introduced elsewhere - is an interesting complication; I hadn't thought of that. ISO 8601 says that use of the standard for years up to 1582 - ie before the Gregorian calendar was introduced - require prior agreement, however it doesn't appear to require explicit separate agreement about denoting dates after the "introduction" of the Gregorian calendar but before its "acceptance" in any particular place.
2009-10 means October 2009, not 2009 through 2010 isn't a particular problem in itself - other than that it may differ from existing conventions. The reverse is also true. If someone had written 2009/10, would it mean 2009 through 2010 (interpreted according to ISO 8601), or October 2009. What about "07/09"? Years 2007 to 2009, or years 1907 to 1909? July 9th or 7th of September? Years 7 to 9, or (per ISO 8601) the 8th to 10th century? Actually its the expiry date of an old credit card - July 2009. Context is everything.
Two things that do become obvious:
  1. Introducing a new standard or convention for expressing dates - even if that standard is explicitly intended to remove confusion - can actually increase confusion unless we state explicitly that we are using that standard or we use it exclusively (which is clearly not feasible in Wikipedia).
  2. Calendar reform makes life difficult! Mitch Ames (talk) 09:45, 7 November 2010 (UTC)
Thinking further about the "... the requirement to always use the Gregorian calendar, even when stating dates related to events that occurred in countries where the Julian calender was in force ...". I presume you list this as a reason not to use ISO 8601, because it seems reasonable that the article is written using the calendar in force at the time. By this logic we would never use BC at all, because with very few exceptions no-one knew in advance when He would be born! Mitch Ames (talk) 11:10, 7 November 2010 (UTC)
Actually there are many instances where my logic is compatible with BC, since the Julian calendar was never in effect in many parts of Asia, Australia, and Africa. Even in the Roman Empire, it didn't go into effect until the fourth consulship of Iulius Caesar (which we would designate 45 BC). Jc3s5h (talk) 14:53, 7 November 2010 (UTC)
I think you may have missed my point. What event happened in a place where "BC" was being used as part of the calendar at the time that the event took place? Mitch Ames (talk) 10:22, 8 November 2010 (UTC)
  • I’m glad Shakescene’s wasn’t a serious suggestion. Let me know when ISO-format date and time is regularly used in real life so expressions of it in writing don’t draw attention to themselves. Greg L (talk) 17:45, 11 November 2010 (UTC)

symbols for metric units

There's a continuing drip feed into Wikipedia of SI prefix and symbol errors e.g. 'Kw' instead of 'kW', 'Kms' instead of 'km', 'centigrade' instead of 'Celsius', 'mHz' instead of 'MHz', 'cm' to indicate cubic metre, etc. Several gnoming editors keep correcting these. From time to time, somebody will disagree with the corrections and say (correctly) that we have no documented consensus for a particular SI symbol or format. We saw this with the claim that 'km/h' wasn't acceptable. This could have been resolved quickly and without the swearing by global guidance in mosnum.

Can we reprint SI guidance within mosnum? Lightmouse (talk) 23:49, 15 November 2010 (UTC)

Clearly non-SI metric units are important, or even pervasive in some or all English-speaking countries, for some applications. Examples include the energy unit kW·h and the pressure unit mmHg.
Furthermore, some, notably Greg L, insist that nonstandard abbreviations for SI units are used in the literature of some fields, and should be used in articles about those fields. While I could go along with this in principle, I find that the examples that have been advanced to date (such as c.c. for engine displacements of selected automobiles) are really too narrow to constitute a proper field deserving of recognition in an encyclopedia of the putative field's idiosyncratic abbreviations. Jc3s5h (talk) 00:08, 16 November 2010 (UTC)


OK. Thank you, Jc3s5h, for alerting me that you mentioned my views here. But you didn’t characterize my views as accurately as I would prefer. I am saying that the existing MOSNUM wording is proper and is generally sufficient. But Lightmouse’s above frustrations show that there are editors (volunteers with no technical writing common sense whatsoever) doing things that are horribly incorrect so we might well have to expand a bit on the current guidelines. Currently they suffice just fine but aren’t fool-proof against industrial-strength fools. They read as follows:
  • 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 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 be best 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).
There has been too much “Rah-rah SI and ISO and IEC” going on here on Wikipedia in the past. The current guidelines essentially amount to this message point:
“SI is more than fine; it is preferred. But don’t try to lead by example and show the world a new and better way if a given discipline routinely has a different way of doing things.”
For astronomy, different units will often be used besides meters and their prefixed multiples. For Honda motorcycles it is not a “1200 mL” engine but a “1200 cc engine”. For infrared astronomy and likely many other disciplines it is “microns” rather than “micrometers.” I didn’t make this stuff up; that’s just the way the world works in certain disciplines. We had our try with space cadets (bless their dear hearts) pushing “mebibytes” instead of “megabytes” for three years on Wikipedia. And the BIPM and their SI endorses that bit. But just because it’s logical doesn’t mean it’s a good idea to stand apart from the way the entire rest of the planet works when communicating to a general-interest readership. It’s real simple: If you find that a given discipline—as evidenced by the most reliable, current literature on that subject—is at odds with the SI, then don’t worry about and stop trying to change the world. Now…
Lightmouse: I am not in the least bit defending the asinine practices that are frustrating you (cited above): 'Kw' instead of 'kW', 'Kms' instead of 'km', 'centigrade' instead of 'Celsius', 'mHz' instead of 'MHz', 'cm' to indicate cubic metre, etc. You shouldn’t have to put up with that crap. There is absolutely zero (zip, nada) excuse for bastardizing unit symbols that are part of the SI. The rule of SI in that regard must be followed to the letter. The solution to that can be a rather focused bullet points that would read roughly like this:
  • Unit symbols for units of measure that are part of the SI shall rigorously follow the rule of the SI. It is kW for kilowatt and is never Kw.
  • Do not refer to temperatures as “centigrade” or “degrees centigrade. It is “Celsius” or “degrees Celsius.”
  • One never adds a plural “s” to a unit symbol that is part of the SI. It is 25 km away and never 25 kms away.
Printing the entire rule of SI is not only unnecessary, it is unwise because parts of it are contradictory to existing guidelines. For instance, it would have us writing “mebibyte”, “kibibyte”, and 75 % rather than the 75% like the rest of the world uses.
I suggest we hurry and get these added ASAP. I’ll leave the honor of implementing them to whomever is more active than I on editing MOSNUM. Greg L (talk) 00:53, 16 November 2010 (UTC)
The degree Celsius is not an SI unit. The SI unit for temperature is the Kelvin. Centigrade is an acceptable alternative (for non-scientific uses) as it is still commonly used, especially in Britain. McLerristarr | Mclay1 01:03, 16 November 2010 (UTC)
Your statement lacks that necessary element of *truthiness*. That’s a fancy way of saying you could not possibly be more incorrect. Note SI brochure, section 2.1.1.5. And note too CIPM, 1948 and 9th CGPM, 1948: Adoption of "degree Celsius". The word “centigrade” was obsoleted 62-some years ago and no reliable source would ever use it. We won’t be having brain-damaged things like “centigrade” being slapped all over Wikipedia just because some editor crawls out of the woodwork and digs in his heels because he can point to retarded practices from a break-away province of the former Yugoslavia. Centigrade is not appropriate for anything on Wikipedia except to write about it in historical contexts. If we caved to stuff like you just spewed here, we might as well not have any manual of style on Wikipedia because all someone has to do is invoke the ENGVAR mantra to make a case to regurgitate whatever they want, wherever they want on Wikipedia’s pages. Greg L (talk) 01:14, 16 November 2010 (UTC)

P.S. As our Stone (mass) article states: “the stone remains widely used within the United Kingdom”. Having stood on a bathroom scale in England, looked down, and had my eyes bug out, I can attest to the truthfulness of that statement. I couldn’t care less what moronic practice goes on over there on those islands; it doesn’t have to blight Wikipedia (you sure as heck won’t find stones used in our Obesity article) just because England invented the dog-gone language used on en.Wikipedia. The rest of the English-speaking world moves on. Greg L (talk) 01:26, 16 November 2010 (UTC)

But those who use "stone" have reason to disagree. You don't have to use it, and you are free to argue against it; but you don't get to outlaw it.
Perhaps not “outlaw” it. But a Manual of Style governs its use. Stones for weight have precious little place for being used on Wikipedia. Greg L (talk) 01:40, 16 November 2010 (UTC)
It might be desirable to use only Celsius, instead of what Celsius himself called his system, but it is not yet true that only brain-dead translators from the Serbo-Croat use "centigrade". Calm down; English does not have an Academy; the CIPM isn't one; they may be more reputable than the UN, but they can no more make centigrade obsolete than the General Assembly can make The Former Yugoslav Republic of Macedonia (alphabetized under T) English usage; general acceptation would do it - and one mark of that is that we cease to argue about it here. Septentrionalis PMAnderson 01:35, 16 November 2010 (UTC)
WP:CIVILITY. I'm not going to say anymore. McLerristarr | Mclay1 01:29, 16 November 2010 (UTC)
I’m not attacking you as a person; I’m pointing out how fallacious your statement was and, ergo, how your resultant conclusion has zero foundation. Greg L (talk) 01:40, 16 November 2010 (UTC)
Yeah, I noticed you were careful that way. But even so, phrases like "what moronic practice goes on" [in the UK] sounds almost like calling Britons morons, for no good reason; people who remain unaware of our obscure regulations have more reason to call us morons than vice versa. Can't you just call them "wrong", and save the insults for those who deserve it, like say terrorists? Art LaPella (talk) 01:50, 16 November 2010 (UTC)
Indeed, there's nothing intrinsically moronic with the stone. It's just that other people are not familiar with it. (And there'd be nothing wrong in using stones for the weight of a Briton in a pre-1985 context on Wikipedia, provided there's a conversion both to kilos and to pounds.) A. di M. (talk) 15:42, 16 November 2010 (UTC)

I'm afraid this suggested guideline won't have the intended effect:

  • Unit symbols for units of measure that are part of the SI shall rigorously follow the rule of the SI. It is kW for kilowatt and is never Kw.

The cubic centimeter is part of SI so the rule of the SI must be rigorously followed: the symbol is cm3. I can't think of any way to reword this proposal that will allow the nonstandard abbreviations some of us like (cc) and disallow the nonstandard abbreviations all of us frown upon (Kms). Jc3s5h (talk) 01:29, 16 November 2010 (UTC)

  • Easy: Add an explicit exception. New green-div coming… 01:40, 16 November 2010 (UTC)
Should follow the rule of SI, where that is normally followed by reliable sources? That should also cover completely non-SI units, like hp. I doubt reliable sources will use Kms. Septentrionalis PMAnderson 01:35, 16 November 2010 (UTC)
No RS would ever use “kms”. And I like your suggested catch-all wording, PMA. Love it, really. Let’s see who slings mud at it before we try to run with it. Greg L (talk) 01:40, 16 November 2010 (UTC)
(edit conflict) Just do whatever the hell reliable sources do, for heaven's sake. Saying that the equator is 40 megametres long would fully comply with the SI, but if no-one does that (and indeed not only does "megametre" get a red underline by Firefox's spell checker, but there are no occurrences of either it or its plural in the BNC, and the only three occurrences of "megameter"/"megameters" in the COCA are from science fiction works), we shouldn't do that either. A. di M. (talk) 01:50, 16 November 2010 (UTC)
The less common units will normally appear in scientific publications, and those publications usually use symbols rather than spelled-out units. The Corpus of Contemporary American English does not seem to provide the ability to perform case sensitive searches, so apparently one cannot do a search that finds "Mm" but excludes "MM" and "mm". So COCA seems of limited use for the topic at hand. Jc3s5h (talk) 01:58, 16 November 2010 (UTC)
(edit conflict) Search the COCA for "Mm", click on "1000" near "SAMPLE" on the top of the results (it will show you 1000 results chosen at random), set your browser's "Find" feature to case-sensitive, and search for "0 Mm", "1 Mm", "2 Mm", ..., and "9 Mm". I can't remember anyone ever actually using megametres, even in contexts where it would make perfect sense. People just use metres or kilometres or (for larger lengths) astronomical units, with exponential notation for the number if necessary. A. di M. (talk) 02:15, 16 November 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Well, A. di M., I like your attitude but I think your proposal relies too greatly upon common sense. Given the collaborative writing environment here and the spectacular range of skills in contributing editors, it would be good to have what you just wrote and some added specificity. Let’s try this one on for size and see where it goes. As Jc3s5h alluded to above, the second and third bullet points could still be seen as being in conflict if one wikilawyers a bit. But I think a common-sense interpretation based on the provided examples avoids any real conflict or problem in grokking the mix. The green-div would be as follows:

  • 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 most-reliable sources relevant to those disciplines use such units, articles should follow this (e.g., using cc in automotive articles and not cm3 and Cashmere wool must have a fiber diameter not greater than 19 microns). Such non-standard units are always linked on first use.
  • Unit symbols for units of measure that are part of the SI must rigorously follow the rule of the SI. It is kW for kilowatt and is never Kw and it is MHz when one means megahertz and never mHz, which means a millihertz, or one oscillation ever thousand seconds.
  • 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 be best 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).
  • One never adds a plural “s” to a unit symbol that is part of the SI. It is 25 km away and never 25 kms away.
  • Do not refer to temperatures as “centigrade” or “degrees centigrade. It is “Celsius” or “degrees Celsius.”

This ought to resolve what poor ol’ Lightmouse is being frustrated over right now. Poor guy. Greg L (talk) 02:04, 16 November 2010 (UTC)

I think the second point is too strong and would prefer Some disciplines often use non-SI units or write SI units differently from BIPM-prescribed format. When the reliable sources in a field normally use SI units, articles should do so; when they do not, articles should follow reliable sources with the same examples. The chief effect here is to avoid some cm fan counting a pile of automotive trade journals and seeing whether the pile with cm or c.c. is higher; I doubt any field is really in a mixed condition - and if one is, everybody will understand both systems. Septentrionalis PMAnderson 02:26, 16 November 2010 (UTC)
I consider the above green-div to be a live document for anyone who is truly trying to be constructive here, i.e. is mostly on my side ;-) So have at it PMA; please revise as you see fit or copy and past your own, below. But I was hoping to address the issue you raised by writing When a clear majority of most-reliable sources… I wouldn’t want to second-guess anyone on any subject if a clear majority of most-reliable sources use this or that unit of measure. If slaughter houses use the length of goat entrails for length and the clear majority of most-reliable sources on that subject follow the practice, far be it me to second-guess a fan-boy. Greg L (talk) 03:14, 16 November 2010 (UTC)
This is all bloat. First the pluralization of symbols is already mentioned in the MOSNUM. The sentence about symbols implies things that are not meant to be implied, for example the symbols for the kilobyte is very often written KB and not kB, even though the "proper symbol" is kB, there is also the issue of cc vs cm3, and plenty of other cases. We do not need to encode every little detail about the MOSNUM in the MOSNUM. There's also no reason to single out "Celcius" and "Centigrade", plenty of units have multiple name. Have Lightmouse's bot run through the bot approval process, explain the changes it will make there, the reasons for them, and then when people bitch about it, which should be once in a blue moon, then point them to the bot approval page. Of if this does not concern a bot, then have the page in your userspace. No need to clutter the MOSNUM because someone complains every six month about something as silly as writting kw·h properly, instead of "kwh". Headbomb {talk / contribs / physics / books} 06:07, 16 November 2010 (UTC)
Maybe we can just add a bullet point to "Unit symbols" or to "Units and symbols often written incorrectly", reading "Note that symbols for prefixes and units are case sensitive: for example, use MHz for the megahertz, not mHz (which would be the symbol of the millihertz)." (And maybe also "Use capital K when it means 1024 and lowercase k when it means 1000.") Then, if we really need to, we can add "Do not use the old-fashioned name degree centigrade for the degree Celsius except in quotations and historical contexts" at the end of the bullet about °C, °F and K, but I think the latest discussion concluded that people do not write that in the first place often enough for it to be worth a mention in the MOS. All the rest seems to be already covered in MOSNUM. A. di M. (talk) 12:40, 16 November 2010 (UTC)

I think we're making progress. Mosnum has grown like coral and has a lot of fuzziness and too much trivia. It's time for a rewrite. There are three principles that seem to have agreement:
1. Wikipedia uses official SI guidance.
2. Wikipedia has a list of acceptable deviations from SI guidance.
3. Wikipedia has a method of changing the list of acceptable deviations.
Unfortunately the current mosnum text suggests that SI guidance is merely an option. I think this opt-in system is why we have such long rambling mosnum text and lengthy discussion outside mosnum. Anyone can demand proof of consensus that normal metric formats apply for any edit. We need to state that SI guidance applies by default, and the list of acceptable deviations are available as opt-outs. We can then purge a lot of the metric trivia. It'll be shorter, clearer, and more usable in article talk pages. Lightmouse (talk) 15:02, 16 November 2010 (UTC)

It's a reasonable approach, provided that no bot remove a questionable unit without an active consensus that it is not used in the field. (I agree that the consensus may consist of "anyone heard of this unit being used", and hearing no reply in a reasonable amount of time, no less than a week.) Bots of that sort can cause too much damage before they are stopped, even when run by operators who generally abide by Wikipedia principles. — Arthur Rubin (talk) 15:08, 16 November 2010 (UTC)
There is one principle that has universal agreement:
  1. Do what the sources do.
We may reasonably note that we want reliable sources on the particular subject of the article; we may reasonably note that in many fields the consensus of the sources is to use SI units - although there are exceptions - and to abbreviate them as SI does. But this still involves editing by reading the sources - which is how Wikipedia is supposed to develop. Septentrionalis PMAnderson 16:23, 16 November 2010 (UTC)
I disagree with PMAnderson's comment "we want reliable sources on the particular subject of the article". We want the reader to understand. An encyclopedia reader tends to be more of a generalist than the audience of sources focused on a narrow topic. We must make a judgment whether many of our readers will be comfortable with the units peculiar to a particular field. Our audience is not the same as the audience of the sources supporting a particular article.
The whole world allows, and most of the world requires, SI in most situations because the world has judged that people should be able to move from field to field without having to learn a new set of units. We should respect the world's judgment. Jc3s5h (talk) 16:32, 16 November 2010 (UTC)
I should prefer to respect the world's judgment by doing what the world does. Septentrionalis PMAnderson 16:49, 16 November 2010 (UTC)
Holy smokes, Jc3s5h; give it up. Your mission to push adoption of the SI will go nowhere. The tide of realization has turned for some number of years now and your mantra is now tedious. Wikipedia does not exist to promote adoption of the SI by using it in a “Oh… didn’t-cha know??”-fashion in articles covering disciplines that routinely use non-SI units of measure. The unit of gravity, the µgal is a non-SI unit formally sanctioned by the BIPM as being acceptable for use with the SI. Encyclopedias exist to educate its readership on topics and properly prepare them for their continuing studies on the subject. Everyone else here understands that we would do our readers a disservice by routinely talking about gravity gradients of 3.1×10−6 s–2 when books on the subject speak of “3.1 µGal per centimeter.” The latter form is what one would be taught in school and use in the field if they got a job in the discipline. The same goes for those units in other fields that aren’t approved for use with the SI, like parsecs in astronomy. Most-reliable literature on certain astronomy topics routinely use that unit of measure. Wikipedia follows that practice where applicable. Your argument that the world has judged that people should be able to move from field to field without having to learn a new set of units could not possibly be more fallacious. Flouting the most-reliable current literature in favor of SI promotion—as with ISO-time promotion and IEC-prefix promotion—has no place on Wikipedia. All we end up with is poor gullible readers who don’t know any better walking into a computer store and telling the sales clerk that they “Want a computer with 512 mebibytes of RAM.” They are either met with suppressed smirks or outright dismissive laughter to their faces because no computer packaging on the planet (or advertisement in a magazine or sales brochure or sales-page Web site) has a bullet point on the outside that says • Features 512 MiB of RAM. Greg L (talk) 17:52, 16 November 2010 (UTC)
If one can actually enroll in a university or other major education program today and receive instruction that predominantly uses an abbreviation for an SI unit of measure that departs from the official BIPM brochure, by all means go ahead and use the non-standard abbreviation in Wikipedia. An example is the mcg, used in medicine (at least in the USA) because the correct symbol, μg, is too easy to confuse with mg, especially when written by doctors with their infamous handwriting, or an EMT in the back of a moving ambulance.
On the other hand, abbreviations used only by obscure groups of enthusiasts, such as collectors of antique automobiles, have no place in Wikipedia. Jc3s5h (talk) 18:19, 16 November 2010 (UTC)
The matter of deciding which unit to use is addressed by the second bullet point in the green-div below: When a clear majority of most-reliable sources relevant to those disciplines use such units… I’m not going to be drawn into wikidrama based on arguments of “obscure groups of enthusiasts” because that’s not what we’re discussing and it is an absurd stretch. If you mean “Just because ‘micron’ is used in infrared astronomy doesn’t mean it has to be used instead of ‘micrometer’ elsewhere on Wikipedia,” then you won’t get any disagreement from anyone around here. Greg L (talk) 18:34, 16 November 2010 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────

  • PMA is absolutely correct. Once again, I find we are shirking our responsibilities and caving to a few with well-worn heels and {{I DON’T LIKE IT}} tags rather than help Lightmouse out here. In starting this thread, he wrote as follows: There's a continuing drip feed into Wikipedia of SI prefix and symbol errors e.g. 'Kw' instead of 'kW', 'Kms' instead of 'km', 'centigrade' instead of 'Celsius', 'mHz' instead of 'MHz', 'cm' to indicate cubic metre, etc. Headbomb pointed out that MOSNUM already addresses the pluralization of unit symbols. So Lightmouse could have referred such editwarring editors to MOSNUM with regard to “Kms”.

    (New stand-alone paragraph here): But other issues that are flat incorrect still exist and if MOSNUM remains silent on those other issues, Lightmouse will no-doubt experience continuing problems on them.

    It simply doesn’t matter, Headbomb, if “plenty of units have multiple names”; just because “micron” jumps off a bridge all the time is no reason for “centigrade” to do so. The proper term has been “Celsius” since 1948 and the vast majority of most-reliable sources use “Celsius. That simple fact drives a silver spike through the heart of “centigrade” and settles the issue. Finito for that one. There are no-doubt other sources in the U.S. that still use “centigrade” (the Alfa Romeo collectors’ car repair shop down the road?) but that doesn’t make the practice right nor a “be politically correct to ignorant sources”-issue. Manuals of style exist for a reason and having MOSNUM here in a collaborative writing environment requires less timidity—not more—in the face of stupid writing practices that passed into history along with the DC‑3. Any example of Hexane’s boiling point is a range spanning 20 degrees centigrade needs to be corrected the instant it is encountered. Arguments that the editor who greatly expand the Hexane article hails from Zamia and en.Wikipedia is so-very *international* so the “other name” ought to stick due to ENGVAR would be absurd. Greg L (talk) 17:06, 16 November 2010 (UTC)


Given that MOSNUM supposedly already addresses plural unit symbols (I haven’t seen it personally but I’ll take Headbomb’s word for it), I’ve trimmed that point out and arrive with the following:

  • 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.
  • Write units symbols properly. The symbol for kilowatt is kW, not Kw for example.
  • Some disciplines often use non-SI units or write SI units differently from BIPM-prescribed format. When the reliable sources in a field normally use SI units, articles should do so; when they do not, articles should follow reliable sources. For instance, it is cc in an article on Honda motorcycles engines and not cm3; the term "micron" (rather than micrometre) is also still in widespread use in certain disciplines. Such non-standard units or unit names 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 be best 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); likewise, most articles should use Celsius or Fahrenheit, not the SI kelvin, for ambient temperatures on Earth.

I’m just not seeing “bloat” here since three of the above bullet points are copied over from what’s currently on MOSNUM. It’s two additional bullet points to help Lightmouse out with legitimate concerns on which he is absolutely correct. Greg L (talk) 17:14, 16 November 2010 (UTC)

If we are going to claim that the examples in the proposed change are used by a clear majority of most-reliable sources relevant to those disciplines, I think we should provide a citation to show this is indeed the case. Who says "cc" is the predominant abbreviation for the automotive dicipline? Who says "micron" is the predominant unit of measure for the fiber diameter of cashmere wool?
Determining if a clear majority exists is a difficult burden, one best undertaken by the lexicographers, not Wikipedia editors. Perhaps a more feasible standard should be used, such as "listed as the preferred term by one or more reliable sources that provide general coverage of the dicipline (especially specialized dictionaries or style guides of relevant scholarly societies)". Jc3s5h (talk) 18:55, 16 November 2010 (UTC)
(edit conflict) I think a better place for the bullet about symbols would be the "Unit symbols" section, immediately after the current fourth (not counting sub-bullets) bullet. I'd also be more explicit that it is capitalization we're talking about, e.g. "Note that symbols for prefixes and units are case sensitive: for example, use MHz for the megahertz, not mHz (which would be the symbol of the millihertz)." (And maybe also "Use capital K when it means 1024 and lowercase k when it means 1000.") As for Celsius vs centigrade, that could be added at the end of the third bullet in "Units and symbols often written incorrectly", rather than as a new bullet. A. di M. (talk) 19:01, 16 November 2010 (UTC)
  • Very well. As you wish A. di M. since you understand the structure of MOSNUM much better than I.

    Responding to you, Jc3s5h: “We” (the volunteer editors cybersquatting at WT:MOSNUM this week) don’t resolve issues of what units are used in any particular discipline; that is left up to the editors specializing in their respective articles if a disagreement over fact arrises. We set only global principals and stay out of day-to-day editwarring. The bedrock principle (When a clear majority of most-reliable sources…) is perfectly clear and there is no need to change it. Your suggestion—listed as the preferred term by one or more reliable sources that provide general coverage of the dicipline (especially specialized dictionaries or style guides of relevant scholarly societies)—is just a prescription for editwarring because all an editor has to do is point to a single source and whine “But my source uses this unit so I want my waaaaaay.” We need none of that. Greg L (talk) 19:12, 16 November 2010 (UTC)

  • One of the best example that I know of for explaining the need to use the correct case for metric symbols is to explain that 1 mW is sufficient to power a hearing aid, whereas 1 MW is needed for the air conditioning of a medium-sized office block. Martinvl (talk) 19:31, 16 November 2010 (UTC)
This box is better, but it now has the problem of endorsing microns as a "sane" unit of measure. It's not. And it also have the problem of that horrible "* Unit symbols for units of measure that are part of the SI must rigorously follow the rule of the SI." rule which means use cm3, not cc, and so on. So I rewrote the thing, removed the problem parts, and I would support this version of it. Headbomb {talk / contribs / physics / books} 19:43, 16 November 2010 (UTC)
Perhaps we should explain that the official BIPM brochure may mention certain units and symbols that BIPM has no control over, such as "%" or "kibibyte" and that these mentions have no more weight in Wikipedia than other technical publications. Jc3s5h (talk) 20:25, 16 November 2010 (UTC)
What you have is mostly fine with me, Headbomb. But case-by-case basis can be interpreted as “come whine to Mommy here at MOSNUM” for each exception. It would be nice if you could find wording that makes one point clear regarding “micron”. The message point would be this: “You and your Turkish butt-stabbing opponents can go look at most-reliable literature and resolve matters of fact yourself, so don’t expect someone here to hold your thingy every time you take a whiz.” Greg L (talk) 01:23, 17 November 2010 (UTC)<
Reworded, how's this? (Also I removed the "most" in the "clear majority of most-reliable sources" because that's grammatically unsound and doesn't add anything to the sentence).Headbomb {talk / contribs / physics / books} 01:32, 17 November 2010 (UTC)
That’s fine. I realized that the principle of looking towards most-reliable literature with regard to whether or not "micron" is appropriate made it best to consolidate it into the bullet point covering ‘which unit or unit symbol to use’. So I took the liberty of doing so, above. It’s now getting rather tidy and logically tied together. Also, “centigrade” was never capitalized; so the “don’t” example was overkill with bad case layered upon wrong word. Greg L (talk) 02:23, 17 November 2010 (UTC)
  • I see your edit, Headbomb, where your edit summary was then this one can disappear, since it is contained in the previous bullet. But I’m not seeing an explicit mentioning of not using centigrade in any other bullet point. I wouldn’t be surprised if Lightmouse hasn’t been experiencing WTFs from other editors besides User:McLerristarr | Mclay1. I’m sure you can find an appropriate place for *the centigrade / Celcius*-thing if you don’t think it is best grouped with these. But I urge that MOSNUM be explicit about that one since other editors will invariably raise ENGVAR once RS-based reasoning forbids its use. I also corrected the spelling; it is “Celsius”, not “Celcius” (not sure who did that, but it’s correct now). Greg L (talk) 04:04, 17 November 2010 (UTC)
    • It doesn't need to be explicit. Is "degrees centigrade" used by "a clear majority of reliable sources relevant to the field"? No they aren't, therefore out the door they go. Therefore it's covered by the "sometimes people do things differently, if the SI says X, but field Y does Z more often, then it's acceptable for our articles on field Y to use Z." bullet of the greenbox. Headbomb {talk / contribs / physics / books} 04:12, 17 November 2010 (UTC)
      • Are you telling me that will stop editwarring when editors think this: “Centigrade is an acceptable alternative (for non-scientific uses) as it is still commonly used, especially in Britain.”(??) Greg L (talk) 04:18, 17 November 2010 (UTC)

        P.S. …or the entire quote, which is this: “The degree Celsius is not an SI unit. The SI unit for temperature is the Kelvin. Centigrade is an acceptable alternative (for non-scientific uses) as it is still commonly used, especially in Britain.”

          • And if that claim is true - and there appears to be some evidence in support of it - we are not helping the encyclopedia by edit-warring to fet rid of a familiar unit. I have included the strongest recommendation I think to be actual consensus. Septentrionalis PMAnderson 04:31, 17 November 2010 (UTC)
            • As for the Celsius, the obvious solution in about 99.99% of the cases is to just use the symbol and a conversion to Fahrenheit, without spelling the names out. Pretty much everyone able to read English knows that 24 °C (75 °F) is too warm for a room but too cold for a bath, don't they. A. di M. (talk) 17:06, 17 November 2010 (UTC)

As for "clear majority of reliable sources" (this thread has become so huge I won't bother to locate the right way to place this post), IMO the point should be that if an unit is called "A" by more than about 80% of the sources and "B" by less than about 20%, we should call it "A"; if it's called "A" by less than about 20% and "B" by more than about 80% we should call it "B", and if the numbers are less different than that we should just pick one for each article and then stick with it unless a really compelling reason for changing it exists; and if A and B are different units rather than different names for the same unit, in the third of the cases above whichever one we choose there should be a conversion to the other, e.g. 25 parsecs (82 ly). (By "about 20%" rather than a specific fraction I mean "common enough that most people familiar with the field will likely have encountered it before. As a rule of thumb, if two English-language mainstream reliable sources independent of each other and published in the last decade use the unit B, it shouldn't considered to be marginal unless there's some evidence of the contrary.) Whether A, B or both are approved or not by the BIPM should not by itself have any relevance. A. di M. (talk) 16:56, 17 November 2010 (UTC)

BTW, does this solve the problems? A. di M. (talk) 08:14, 18 November 2010 (UTC)
24 °C (75 °F) isn't too warm for a room Nil Einne (talk) 21:00, 18 November 2010 (UTC)
If you are wearing short sleeves, and either you heat it using something else than fossil fuels or live in a planet where the use of fossils fuels hasn't done that much damage (and also have shitloads of money for the fuels you use), then it isn't too warm. A. di M. (talk) 10:27, 19 November 2010 (UTC)

Example of non-standard abbreviation for SI unit

Is "cc" a good example of a nonstandard abbreviation for an SI unit, when explaining in Wikipedia's Manual of Style (dates and numbers) that in some fields nonstandard abbreviations should be preferred over the standard symbol? Jc3s5h (talk) 12:20, 17 November 2010 (UTC)



NOTICE: The above RfC is regarding wording that no longer exists. This RfC speaks of text that read it is cc in automotive articles and not cm3. That text no longer exists and is no longer under consideration. Greg L (talk) 03:33, 18 November 2010 (UTC)
It still exists in MOSNUM as of now. A. di M. (talk) 08:12, 18 November 2010 (UTC)
Indeed. It’s been there for years. But what precipitated it was repeating it here while considering other text. So the example of a common and appropriate use of “cc” has been changed to Honda motorcycle engines and he’s objecting to that now. Such extremism in the face of concrete reality is not helpful. Greg L (talk) 15:09, 18 November 2010 (UTC)

It appears no one is able to suggest an example that has the merit of "cc" (it is well known to a wide readership) without the disadvantage of being punctuated in several different ways. I think if anyone had a better example in mind, it would have surfaced by now. Is there any objection to closing the RfC? Jc3s5h (talk) 18:48, 19 November 2010 (UTC)
I agree 100% with your proposed conclusion but not the self-serving premiss upon which it was supposedly based. The below straw poll shows a clear general consensus to adopt wording that includes Honda motorcycle engines as an example of a subject that uses “cc”; ergo, your statement would be accurate if you had written “no one is able to suggest an example I like.” Greg L (talk) 19:16, 19 November 2010 (UTC)

Challenge if statement is factual

I challenge the statement "it is cc in automotive articles and not cm3". In the absence of qualification, "automotive articles" includes everything from antique cars to 2011 models, and all aspects from economic impact of the auto industry to mechanical engineering aspects of autos.

It would be better to use an example that will not require extensive qualification and for which strong evidence can presented. I would suggest the abbreviation "mcg" for microgram in medicine, but I do not have access to a wide range of reliable medical sources to determine whether a clear majority of them use "mcg" and frown upon "μg".Jc3s5h (talk) 12:12, 17 November 2010 (UTC)

I googled both "mcg" and "µg" with a filter of ".nhs.uk". The number of hits for mcg outnumbered µg by about 10 to 1. Since the letter µ does not appear on the UK keyboard, there fewer hits are expected. However, when I applied the filter ".thelancet.com", the numbers were reversed. Thus far from mediacl circles dispaproving of the use of "µg" being, it appears to me that medical academic cirles prefer "µg" to "mcg". Martinvl (talk) 12:44, 17 November 2010 (UTC)
What about mAh (rather than the standard SI mA h or mA·h) for the milliampere-hour on rechargeable batteries? A significant fraction of people able to use a computer will be familiar with that. A. di M. (talk) 17:00, 17 November 2010 (UTC)
I don't think a clear majority of the best sources related to rechargeable batteries would use mAh instead of mA h or mA·h so I don't think there is any article topic for which we would recommend mAh. This does bring up a different point though: h·mA is just as allowable if we blindly follow the official BIPM brochure, but it is not used in the literature. Jc3s5h (talk) 18:09, 17 November 2010 (UTC)
Are you sure? A. di M. (talk) 18:20, 17 November 2010 (UTC)
Your point eludes me. Jc3s5h (talk) 23:26, 17 November 2010 (UTC)
Of the first ten results on Google Books for "mA h batteries", nine use "mAh" without a space or a dot. A. di M. (talk) 23:36, 17 November 2010 (UTC)
I have no idea how Google selects books; in my mind best sources for rechargeable batteries would be scholarly journals. Of the 10 sources you gave, when I actually looked at the previews, I found that #9 and #10 didn't use any form of mA h, at least on the page Google claimed they did. Of the other 8, 2 that were about batteries or closely related subjects used no space, 3 so-so sources that were only peripherally about batteries used no space, and three books about batteries used a space. So that is 3 to 2 in favor of a space, or 5 to 3 in favor of no space, depending on how you want to count. Jc3s5h (talk) 00:12, 18 November 2010 (UTC)
  • Jc3s5h push to make Wikipedia follow the rule of the SI rather than follow the way the world really works in certain disciplines is not the consensus view. His making waves with formal RfCs over a minor point is unfortunate. I changed the example to List of Honda motorcycles so this point is now moot. Time for the next wikidrama. Greg L (talk) 17:59, 17 November 2010 (UTC)

    P.S. Since the point of contention over “automotive” applications (Honda motorcycle engines is an irrefutable example), I removed the RfC tag. I don’t know if only admins can do that, so if this was an error, please revert. Greg L (talk) 18:02, 17 November 2010 (UTC)

Revert removal of RFC tag because the discussion is over when there is consensus it is over or, by default, after 25 days. While Honda motorcycle engines might be easier to defend than automotive articles in general as a topic where "cc" is appropriate, no concise defense that could easily be added to the manual as a footnote has been offered. Perhaps such a defense can be found, or a different example that can be defended more easily can be found. Jc3s5h (talk) 18:18, 17 November 2010 (UTC)
P.S. a Wikipedia list is not even one reliable source, much less a clear majority of the best sources (and the one reference provided for that list is a dead link). Furthermore, whether Honda motorcycle engines, as described to consumers, form a discipline is debatable. Jc3s5h (talk) 18:23, 17 November 2010 (UTC)
  • Well, you clearly don’t understand much about Honda motorcycle engines and the standard practices in that field. Your “rah-rah SI” crusade now has you denying reality to the point where you are implying that Wikipedia’s going with the flow on Honda motorcycles is some sort of contrived fabrication by some CGS fans. Good luck with that tact. And just because you started a (now-moot) RfC is no reason the community can’t just start ignoring you from hereon; you are bordering on being tendentious by restoring an RfC tag predicated upon a point over “automotive articles” when I addressed your concern and made it Hondas in the green-div. So I’ve deleted the RfC tag again. If you must slap {I DON'T LIKE IT} tags all over the place instead of discussing things here without your hair on fire, then start a NEW RfC below that makes a half-way rational case speaking to the CURRENT green-div and how the new wording too is somehow not true. Greg L (talk) 18:39, 17 November 2010 (UTC)

Symbol-itis on the brain

Do people here think MOSNUM is clear enough for those contributors who think they are being *scientificy* and are actually writing clearly when they make the first sentence of the lede like this?

SN 1979C was a supernova about 50 Mlys away in Messier 100, a spiral galaxy in the constellation Coma Berenices.

What’s that unit of measure? Milkies? That was from this version of ‘SN 1979C’.

MOSNUM currently says this:

Where space is limited, such as in tables, infoboxes, and parenthetical notes, and in mathematical formulas, unit symbols are preferable. In prose it is usually better to spell out unit names, but symbols may also be used when a unit (especially one with a very long name) is used many times in an article.

Maybe that point needs to be punched up a bit. Maybe a “Hey! LISTEN UP:” added to the start of it. Seriously, I suggest only this:

It is generally best to write out the full names of units of measure in regular body text. Unit symbols may be used when a unit—especially one with a very long name—is used many times in an article, but the symbol should be introduced parenthetically at least once, particularly if it is a unit symbol not used or encountered in daily life by a general-interest readership. Where space is limited, such as in tables, infoboxes, and parenthetical notes, and in mathematical formulas, unit symbols are preferable.

P.S. I’ve been *into* astronomy for forty years and subscribe to Sky & Telescope. I was being disingenuous when I feigned about “milkies.” The point is that that SN 1979C was recently the subject of an AP article picked up by many of their subscribers and Wikipedia should have a bot running about fixing this sort of stuff. Greg L (talk) 22:24, 18 November 2010 (UTC)

I'm confident that neither that language, nor "Hey! LISTEN UP:" would get people to read our endless rules. A much better way to get people to consider paying attention to us, would be to remove rules like that one, because even we don't use them. That rule goes on to say: "However, spell out the first instance of each unit in an article (for example, the typical batch is 250 kilograms ... and then 15 kg of emulsifier is added), except for unit names which are hardly ever spelled out even in publications for general audiences (e.g. the degree Celsius)." Try Googling "kg" in Wikipedia, and try to find an example where that rule is followed! (Articles about the kilogram as a unit don't count.) Art LaPella (talk) 22:54, 18 November 2010 (UTC)
Ahhh. “Endles rules.” I see. That is just too funny. Greg L (talk) 00:25, 19 November 2010 (UTC)
Or more precisely, about 500K of MOS rules, last time I counted. Art LaPella (talk) 01:03, 19 November 2010 (UTC)
500 k? With an all-volunteer army of contributors, some of whom couldn’t tell the difference between an em-dash and a hyphen to save their lives, that’s quite some suggestion of yours that we reduce what’s on MOSNUM. Perhaps we should leave those silly manuals of style to the pros at the Associated Press and The New York Times who barely need them? Maybe one rule:

“Wikipedia is the encyclopedia anyone can edit. So let’s *celebrate diversity* and show how eighth-graders really can teach the 50-year-old professional technical writers a thing or two. All technical writing practices are equally good. Whatever you want is just fiiiine. M‘kay?

Seriously, all MOSNUM needs is a wholesale reorganization since it suffers dearly from “too many cooks in the kitchen-itis.” MOSNUM is even worse than AutoCAD 10. I’d volunteer to do it but I’d have acid stomach from volunteers with as much say as I have where they write how they wished they had a time machine so they could prevent me from being born. And why be so frustrated with me? Because I am so wrong for thinking that Celsius is part of the SI. You know, that sort of silly stuff from me; I can’t help myself. Greg L (talk) 01:28, 19 November 2010 (UTC)

As we reduce MOS, we increase the chances that editors will actually read it. We don't have to write the technical manual for English; there are plenty of them (and they disagree); we need to remind people of points that are actual problems, settle the stuff that's not in other manuals because they're not about wikis (and this shrinks steadily as they print new editions), and indicate what to do "when doctors disagree". Under the last head, we could get by with a lot more of follow sources, be consistent within articles, and there are two ways; either is acceptable than we have. Septentrionalis PMAnderson 20:55, 19 November 2010 (UTC)
It's not that that point is not clear, it's that whoever wrote that supernova article most likely hadn't ever read it in the first place; so, rewording it wouldn't help. The only short-term solution to that problem is to just fit it (as you did); then a better solution would be restructuring MOSNUM so that people can actually find what they need, but that would require much more time. A. di M. (talk) 02:38, 19 November 2010 (UTC)
Actually, I was more frustrated with the clarifying sentence I quoted about kilograms, which all of us disregard in practice if not in theory. Removing disregarded rules would result in more attention to our remaining rules and a more professional-looking result, not less. Art LaPella (talk) 04:39, 19 November 2010 (UTC)
Maybe we should replace kilograms in the example with a less ‘everyday’ measurement unit such as the light year. A. di M. (talk) 10:11, 19 November 2010 (UTC)
Yes, either that or start spelling out the first mention of kilograms in articles throughout Wikipedia; others are more likely to learn by example than by repeating rules 10 times in red, thus further obscuring the other rules. If we change "kilograms" to "light years", then the rest of the guideline should be changed to match – "obscure unit names", not just "unit names". Art LaPella (talk) 16:30, 19 November 2010 (UTC)
OK, I should say I have no objection to GregL's change if combined with A. di M.'s change, and with "units of measure in regular body text" qualified with a word like "unusual units ...". Art LaPella (talk) 16:52, 19 November 2010 (UTC)

An important point that I feel gets overlooked in some of these discussions is that MOS guidance is not necessary for the vast majority of editors. Does the average editor need to know the difference between "K", "k", "kb", "Kb", "kB" or "kilobyte" if they are trying to improve an article? No; and no one should make someone who is adding content feel less welcome because they get it "wrong". However the MOS is very useful for the people who come along after the edits in the attempt to standardize the articles (to improve the reading experience for all visitors to WP). Therefore, I feel the MOS can lean towards being more complex and providing more instruction and examples (to assist the tidy-up editors). I like the strategy for each point in the MOS of providing a simple summary followed by description and examples. Perhaps the further description and examples could go in a collapsed division after each point in the MOS (if people want a more straightforward-looking MOS)?  GFHandel.   03:42, 19 November 2010 (UTC)

Yes, I have often advocated some kind of hierarchy to help prevent missing the forest for the trees. In the case of the kilogram rule, I don't enforce it because I don't see examples of anyone else doing it; it looks like people just making rules without thinking about how to put them into practice. Art LaPella (talk) 16:30, 19 November 2010 (UTC)

What does "Write unit symbols properly." mean?

The proposed text contains:

  • "Write unit symbols properly. The symbol for kilowatt is kW, not Kw for example."

If the dispute is about 'kW', that is fine. But that wouldn't resolve the recent dispute where one editor claimed 'kph' was proper and 'km/h' improper. Other editors think 'mtr' is proper instead of 'm'. The issue of 'mHz' and 'Mhz' rather than 'MHz' is the same. All disputes about writing metric units are because of competing definitions of 'properly'. It should be relatively easy for mosnum to help resolve such disputes.

  1. It implies the prefix in 'kW' must be as defined in the SI prefix page. Is this true?
  2. It implies the symbol in 'kW' must be as defined in the SI table 3. Is this true?
  3. It implies the combination (instead of 'k-W' or 'k W') must be as defined in the SI prefix page. Is this true?
  4. Unless otherwise specifed in mosnum text, which other prefixes should be as defined in the SI prefix page?
  5. Unless otherwise specifed in mosnum text, which other symbols should be as defined in SI tables?

Still seeking clarity. Lightmouse (talk) 09:58, 19 November 2010 (UTC)

The seventh bullet in WP:MOSNUM#Unit symbols implies that km/h is the correct symbol, the fourth bullet in the same section makes implies that m is the correct symbol, and the fourth bullet in WP:MOSNUM#Units and symbols often written incorrectly (which I recently added) makes clear that MHz is the correct symbols. The rules are already there; the problem is that they're hard to find, and duplicating them wouldn't help. Incorporating International System of Units#Units or International System of Units#SI writing style in the MOSNUM would also be overkill (though linking to there could be an idea). Finally, we shouldn't give excess weight to the SI recommendation because being a SI unit doesn't make it automatically suitable for Wikipedia, if reliable sources don't use it and hence readers are unfamiliar with it (e.g. with the megametre, see my comments above). As for editors who think mtr or mHz are proper, you should just tell them to have a look at how those units are usually symbolized in the real world. A. di M. (talk) 10:22, 19 November 2010 (UTC)

I agree, the guidance is scattered into a maze of twisty little passages. The proposed text just needs to make explicit the definition of 'properly'. Thus:

Lightmouse (talk) 11:00, 19 November 2010 (UTC)

Lightmouse makes good points. I think it's easy for us MOSNUM editors to be unaware of the job at the coal face. Lightmouse faces niggly, and sometimes not so niggly, complaints from editors who frankly could use more organised and confident guidance here. Probably we need to build detail into an appendix. At the moment, it is most unsatisfactory. Tony (talk) 14:19, 19 November 2010 (UTC)
  • It wasn’t my idea. Headbomb stripped it down. I agree, we need to have more explicit and specific wording that doesn’t read so much like “Write good.” I propose we get the above wording, straw poll finished off and adopted and then make a quick fix to address this concern. Greg L (talk) 17:50, 19 November 2010 (UTC)


  • The results of the above straw poll show a general consensus exists for wording to address what should be uncontroversial points regarding following the rule of the SI. I added the contents of the above green‑div to MOSNUM but have left the straw poll active in the off-chance that consensus could change. In addressing Lightmouse’s concern over the extreme brevity of the second bullet point and its clear ambiguity, I propose the following:
  • Unit symbols that are—or appear to be—part of the International System of Units (SI) must rigorously follow the rule of the SI. It is kW for kilowatt and never Kw (uppercase K is the symbol for the Kelvin in the SI and is not the prefix denoting 1000). And it is MHz when one means one-million hertz and never mHz unless one truly means one-thousandth of a hertz. The only exception to this pertains to omitting the space before the percent symbol (as addressed below at "Percentages"). For guidance, see "SI prefix", "SI base unit", "SI derived unit", and "SI writing style".
The guideline itself comprises only 24 words. When taken in context with the other bullet points in that section (where “cc” and “micron” are clearly not part of the SI nor intended to be so), I think the intent and scope of this bullet point is perfectly clear.

The examples provided in this bullet point illustrate the intent for those who are not as familiar with this issue as we are. The links are an acknowledgment to a point in the previous thread raised by GFHandel and Art LaPella pointing out the simple reality that Wikipedia receives its share of novice contributors who can foul things up quickly and could use assistance to help them learn how to do things correctly. Such guidance can only help minimize the followup clean-up required by the rest of us. Greg L (talk) 18:14, 19 November 2010 (UTC)

  • Not seeing how this should be controversial, I took the liberty of replacing the second bullet point with this version. Greg L (talk) 19:08, 19 November 2010 (UTC)
  • I note this edit by Wjemather to remove “appears to be”. Logically, Kw is not part of the SI and would technically be exempt. So I was trying to circumvent that argument and be über-logical by saying if “If it is part of the SI or appears to be so.” I’m not going to edit war over it since what I had may be unnecessary verbosity. I sure hope the truncation doesn’t open the door to wikilawyering. Greg L (talk) 19:29, 19 November 2010 (UTC)
    • An occasion for "follow reliable sources"? Unless the world's copyediting has gotten so bad than Kw is endemic, that should answer any such claims of exemption. Septentrionalis PMAnderson 21:03, 19 November 2010 (UTC)

Percent symbol and SI

Lightmouse just made an edit that I had thought about making myself. I decided not to because a full explanation would have been too wordy, and I wasen't sure how a really short explanation would go over. Of course % is not an SI symbol, but the BIPM's SI brochure claims that when it is used with SI units, there should be a space before it; this manual and some external style manuals disagree. I think we could justify our decision in two ways:

  1. The encyclopedia as a whole is not using % with SI because we also use traditional units, many of which are not accepted for use with SI. We are not SI compliant and can pick and choose which parts of SI we will use.
  2. BIPM exceeded its authority by attempting to make rules about a non-SI, non-metric unit that existed before they did. We should reject their rule because they had no power to make the rule.

So should we leave Lightmouse's edit as it is, or is there a short way to touch on the full explanation that I have not thought of? Jc3s5h (talk) 16:09, 20 November 2010 (UTC)

Or the third alternative: BIPM is a recommended standard; it can be no more - it doesn't employ censors. It was and is largely sensible, and has been almost entirely adopted in practice. Where it has not been adopted (as with %), it is not usage - and we should not follow it. Septentrionalis PMAnderson 16:41, 20 November 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I appreciate you all raising and discussing this issue here. The community voted on proposed wording above and now there shouldn’t be an endless stream of edits that seriously affect its scope or implications. I fear things might degrade into editwars where undiscussed “minor” edits to the placed text has fairly substantial change in meaning.

With regard to the percent symbol, indeed, the percent symbol is not an SI unit. However, the percent symbol represents a dimensionless quantity (as with “parts per million”, the percent symbol means parts per hundred) and the percent symbol is approved for use with the SI. More to the point, the BIPM specifically addresses how to format the percent symbol via their SI brochure, Section 5.3.7 Stating values of dimensionless quantities, or quantities of dimension one. It states as follows:

When it is used, a space separates the number and the symbol %.

It is this specific aspect of SI writing style that MOSNUM (and the rest of the planet except perhaps for scientific papers) specifically flouts. The first bullet point in the version of MOSNUM as I write this requires conformance with SI writing style. That is just fabulous as far as its scope applies as indicated by the provided examples; Lightmouse should not have had to put up with flack from the community when his bot corrects stuff like “Kw”. That bullet point is sweeping and explicit. However, MOSNUM—like the rest of the general-interest world—has long ignored the rule of the SI with regard to writing style as it applies to the percent symbol. So it is important that MOSNUM not appear to be internally in conflict where some bullet points contradict others.

I might add that I disagree with Jc3s5h where he wrote how the BIPM exceeded its authority by attempting to make rules about a non-SI, non-metric unit that existed before they did. The full scope of the BIPM’s rules are just fine for the intended audience: readers of peer-reviewed scientific papers that are published internationally. For the rest of us (a general-interest readership), most of the SI’s rules have been adopted, but not all. In acknowledgement of this simple reality, even NASA ignores some aspects of the SI (using the name “micron”) when communicating to a general-interest readership.

The BIPM also has such advise as how there are certain countries that still use the long number system for named numbers (like billion). To address this ambiguity, they call for writing A carbon monoxide concentration of 35×10−6 instead of 35 ppm. Quite wisely, our ‘Carbon monoxide’ article, here flouts SI writing style with regard to not using “ppm.” Why? Because MOSNUM declares that named numbers on en.Wikipedia are from the short-scale, expressions like “ppm” and “ppb” are consistent with this convention, and that’s how we least confuse a general-interest readership that is not accustomed to reading internationally distributed, peer-reviewed scientific papers. Greg L (talk) 17:45, 20 November 2010 (UTC)

Actually, not even peer-reviewed scientific papers that are published internationally usually use a space before %, when they are written in English. A. di M. (talk) 19:28, 20 November 2010 (UTC)
Are you sure? That sure makes the BIPM look like intransigent nerds. Greg L (talk) 19:47, 20 November 2010 (UTC)
The only texts where I recall seeing % used with a space before (excluding ones in languages such as French) were from the BIPM/CIPM/ISO themselves. I think there are also other recommendations by them which no-one else follows, but I can't remember any specific one right now. A. di M. (talk) 20:16, 20 November 2010 (UTC)

There have been some comments about the BIPM brochure being a recommendation. I don't know how the BIPM views the brochure, but in the US SI as defined by the CGPM (which controls the BIPM) as interpreted by the Secretary of Commerce is the law. There is a Federal Register notice in which the Secretary adopts NIST Special Publication 330 as the official interpretation of SI for the US. That publication is virtually identical to the NIST brochure, except for a few spelling and capitalization changes ("meter", "liter", symbol for liter is capitalized, a few others). So BIPM does not employ censors, but each state employs weights and measures officials who can seize improperly labeled goods. They can't censor encyclopedias, but they can censor product labels.

I think it is quite unlikely that goods would be seized over minor misuse of symbols, but the official interpretation could be used as a basis for disciplining any weights and measures official who tries to enforce erroneous pet peeves. Jc3s5h (talk) 19:25, 20 November 2010 (UTC)

This is a misrepresentation. On the one hand, selling 950 grams of something as a kilogram is (properly) a matter of law; just as selling 15 oz as a pound is. But we are not merchants; we are writing English to communicate. Where is the Secretary of Commerce empowered to edit prose of private and volunteer editors? Septentrionalis PMAnderson 23:08, 20 November 2010 (UTC)
I don't claim the government can force Wikipedia editors to write in any particular style; I claim it is inaccurate to describe the BIPM brochure as merely a recommendation. Because a slightly edited version of it has been adopted by the Secretary, it has the force of law in the US, although editing encyclopedias is not governed by that particular law. Jc3s5h (talk) 23:33, 20 November 2010 (UTC)
The WWW is heavily dominated by American practices. America has “Silicon Valley.” The French have “wine-making country.” Now they’re all *pissy* over the ever-diminishing influence of their language. Telle est la vie. Greg L (talk) 02:01, 21 November 2010 (UTC)

Encyclopedia, not collage

I find this change unacceptable:

*When the source uses one set of units, generally put that one first; if editors cannot agree, put the source's units first. If they are not first, this should be stated in the citation.

This would force us to change the order of units from one statement to the next, depending which source supports it. The purpose of an encyclopedia is to assemble sources into a cohesive whole, not cut and paste passages together. Jc3s5h (talk) 19:08, 20 November 2010 (UTC)

If the statements for which you have to cite different sources using different units are close enough that the switch of units is a problems, odds are you're trying to WP:SYNTH. A. di M. (talk) 19:32, 20 November 2010 (UTC)
This is a question of style, not numerical precision. It is confusing to read one sentence with kilometers first and the next sentence with miles first. As the prior version of the guideline indicated, it is worth the extra trouble if the number cannot be verified unless it is stated in particular units (perhaps the only source that can be found for one sentence uses chains and the rest of the article uses feet). It is also worth the trouble if the quantity is defined in particular units (an American football field is 360 feet long, including end zones). But numbers that are limited precision measurements should all give the same unit first in one article. Jc3s5h (talk) 20:34, 20 November 2010 (UTC)

It would be nice, Jc3s5h, if you would provide edit differences in your post. We’re all volunteers here with limited time and shouldn’t have to hunt for the edit difference underlying your objections. You are apparently referring to this edit diff by PMA, yes?

The old version was this:

When a mesurement can only be verified in one set of units, generally put that one first; likewise if the quantity is defined in a given set of units and is therefore exact in them. The exclusion zone is ten nautical miles (about 11 statute miles, 18.5 km) in radius

PMA’s new is this:

When the source uses one set of units, generally put that one first; if editors cannot agree, put the source's units first. If they are not first, this should be stated in the citation. Likewise, if the quantity is defined in a given set of units and is therefore exact in them, put it first. The exclusion zone is ten nautical miles (about 11 statute miles, 18.5 km) in radius

I’m not seeing a big problem here. Your statement here of what an encyclopedia supposedly is (The purpose of an encyclopedia is to assemble sources into a cohesive whole) seems to be a paraphrasing of another thing you wrote earlier above: The whole world allows, and most of the world requires, SI in most situations because the world has judged that people should be able to move from field to field without having to learn a new set of units. We should respect the world's judgment. That “respecting the world’s judgment”-bit sounds awfully noble but is clearly contrary to the consensus view here and the reason for that is such views simply flout Technical Writing 101. Wikipedia simply goes with the flow for any given discipline and is not to be hijacked by lovers of all-things-SI to promote earth’s adoption into the United Federation of Planets at the expense of either accuracy or clarity. The bullet point in question pertains to maintaining accuracy and scientific rigor with measures.

I am so reluctant to get mired in this argument. The basic principle as evidenced in the bullet point’s {xt} example seems perfectly clear. I’m curious; what precipitated PMA’s edit. Was PMA’s move a response to an edit conflict? And if so, was it one in which you were active, Jc3s5h? Either way, I wish he would weigh in here and you two resolve your differences. Greg L (talk) 19:46, 20 November 2010 (UTC)

Wikipedia sets style article-by-article. We don't have a strict style for the whole encyclopedia, nor do we use the style of each source for the sentence it supports.
I don't know what prompted PMA's edit; it does not seem to be part of a back-and-forth series of edits. Jc3s5h (talk) 20:21, 20 November 2010 (UTC)
The scope and applicability of MOSNUM is spelled out in the lede of MOSNUM, which speaks of how editors should “use common sense.” Clearly, that little tid-bit of advise is open to interpretation. (I’m just saying…) Greg L (talk) 21:18, 20 November 2010 (UTC)
I edited because the recent changes led me to look at the entire section - and I noticed a few holes and some repetition. One reason I used generally is that it doesn't force anything; but if we are going to represent a source correctly, we should in general represent that it tells us. If all we know is that an object weighed "100 kilograms", it would be overprecision to convert that ourselves (to, presumably, 220 lb) before stating it; the statement is likely to be approximate in the original primary source. The same reasoning applies if what we know is that the object weighed "300 lb"; we should not lightly convert to "135 kg".
On the other hand, if something is defined as 10 nmi, saying it is 18 km is not the case; saying it is "about 18 km" is a loss of data, and compels anyone who wants to figure out what it really is to reverse-engineer - a matter of guesswork.
There may well be other countervailing considerations, including consistency within an article; but the guideline says that, and advises editors to balance them. Septentrionalis PMAnderson 22:52, 20 November 2010 (UTC)
The guideline should encourage consistency except when consistency would create confusion or a loss of information. I don't believe there is any expectation on the part of readers that source units will be first. I believe this for two reasons:
  1. Most good quality publications use units in a consistent manner and make it easy for readers to compare numbers from various sources by stating them in consistent units.
  2. Most publications do not routinely give conversions for quantities, so readers have not even thought about the issue of choosing the order.
Since inconsistent order makes the articles look sloppy and creates confusion, and the order does not actually communicate anything to most readers, I intend to restore the previous wording. Jc3s5h (talk) 23:55, 20 November 2010 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────

  • OK. Here’s a challenge for you two. What if Scientific American took some abstruse technical jargon about ultra-high-energy gamma ray photons and wrote this: To understand just how energetic this is, if all the energy in an ordinary green photon from the Sun was converted to kinetic energy, it would be sufficient to make a 25 milligram rice grain jump upwards against Earth’s gravity by half the diameter of a helium-4 nucleus. If all the energy of an 18 TeV gamma ray photon was converted to kinetic energy, it would be sufficient to make that same rice grain jump upwards by 0.5 inch.

    My question to both of you is this: If you were tasked with adding that bit to our Gamma ray article, how would your proposed wording deal with this? Greg L (talk) 01:18, 21 November 2010 (UTC)

    "18 trillion electron volts, enough to raise a 25-milligram (0.39 gr) rice grain by 0.5 inches (1.3 cm) against Earth's gravity"? I don't think many readers would even notice that the mass is given in metric first and the height in imperial first, let alone be bothered by that. A. di M. (talk) 16:58, 21 November 2010 (UTC)
Assuming your intent is that it would be cited to the original RS, this seems to be a reasonably accurate way of taking somewhat sketchy information (a measure using a U.S. Customary unit expressed at low precision) and adding a conversion to make the statement more accessible to that portion of our readership who think in metric terms.

On a separate note: I’m not even sure the conversion to grains would be necessary. At least in the U.S., about the only people accustomed to grains are handloaders who make their own ammunition (both bullets and powder are measured in grains). Given that this would be in a scientific article, the fact that it is a rice grain conveys the general magnitude of the beast ten-thousand times better than a parenthetical to grains.

How about you guys, Jc3s5h and PMA? Would you address the issue differently than A. di M.? Greg L (talk) 17:16, 21 November 2010 (UTC)

I would probably paraphrase, but to make it easier for those discussing this, I'll just modify the units for this example. I will suppose the article goes on to discuss the energy of photons at length, so electron volts will be listed first. I would cite the work but not point out that 0.5 inch is the source value because this is obviously a low-precision example. I might decide otherwise when more precision is involved. I do not convert 25 milligrams because US readers are most likely to encounter such small masses while reading drug labels, which are in metric. Likewise I do not convert the energy to US customary because they are rarely used for such small energies. I will further suppose the article uses UK spelling and I will do likewise, even though Scientific American is a US source.
To understand just how energetic this is, if all the energy in an ordinary green photon from the Sun was converted to kinetic energy, it would be sufficient to make a 25 milligram rice grain jump upwards against Earth’s gravity by half the diameter of a helium-4 nucleus. If all the energy of an 18 TeV<ref>The teraelectronvolt is an energy unit equal to about 160 nanojoules.</ref> gamma ray photon was converted to kinetic energy, it would be sufficient to make that same rice grain jump upwards by 13 millimetres (0.5 in).
Jc3s5h (talk) 18:31, 21 November 2010 (UTC)

I think PMAnderson is correctly identifying the hazard we need to avoid when quoting sources. Jc3s5h’s above method of paraphrasing a cited source falls victim to not heeding this lesson. I note that A. di M., who hails from Italy where the metric system is used, recognized the pitfalls of trying to paraphrase a source in a manner that would reverse the order of the primary unit of measure to one not originally used by a reliable source. This occurs because conversions are understood to be generally close and mustn’t be used in a manner that introduces false precision.

In the above scenario, the mass of the rice grain can be taken to be defining (it is 25 milligrams). So too is the energy of 18 TeV as it is the subject of interest (“what is 18 TeV like?). Even if the two values have uncertainties of one part each (one part in 25 and one part in 18), since the math in converting to velocity (height) uses the two terms in division, the precision statistically would be one part in 14.6. Accordingly, it would have been appropriate for the original RS to express the value to the nearest millimeter had they chosen to do so. Unless one can cite all the way back to the original scientific paper or do the math themselves (math doesn’t have to be cited), the value of interest (18 TeV) converted to kinetic energy would cause a defining mass of 25 mg to jump upwards against one standard gravity by 11.763 millimeters, or 0.46311 inch.

In this hypothetical, a scientist who wrote a technical paper (or a consulting technical writer for Scientific American) calculated the value and a copy editor for Scientific American wrote “0.5 inch”, which is true; 0.46311 inch rounds to 0.5 inch. As mere wikipedians, we should properly quote the RS and do our best with the available information. As A. di M. wrote, it would be …jump 0.5 inch (13 mm). Accurately quoting like this instantly conveys that the measure is expressed in relatively low precision (somewhere between 0.451 and 0.549 inch) and the parenthetical is provided to make the magnitude of the measure more accessible for those readers who think in metric terms.

Assuming one hasn’t crunched the math (or can’t because mathematical pieces of the puzzle are missing), the approach A. di M. used above is the only one that does not introduce implied precision and accuracy where none exists. It is incorrect practices for mere wikipedians to write …13 mm (0.5 inch) because doing so would introduce false precision and would suggest that it is the conversion that is approximate. That is not the case. If Scientific American had elected to convey the measure with greater precision, they would have written “12 mm” or “0.46 inch” and we would have had more to work with. But they didn’t in this hypothetical. Greg L (talk) 19:22, 21 November 2010 (UTC)

Congratulations are due to Greg L for providing such a subtle, well-stated, and mathematically correct example. I still think my version is acceptable because (a) in a publication where conversions are provided for most measurements, most readers will not realize that "0.5 inch (1.3 cm)" means that 0.5 inch is the source value and 13 cm is the converted value, and (b) this isn't the sort of passage readers would draw numbers from for further calculations.
If I were writing a publication with good control over both the writers and readers, and the publication were dedicated to comparing and analyzing sources, I might try to introduce the convention that the source value goes first. But in the case of Wikipedia, I think this would amount to a secret code known only to careful readers of this style manual. Jc3s5h (talk) 20:23, 21 November 2010 (UTC)
Thanks: An SI-Nazi (I mean that in a good way) who is a darn good sport. Greg L (talk) 21:54, 21 November 2010 (UTC)

Time to gauge consensus

We’ve flogged this more than enough now; it shouldn’t have to drag on any further. And, as evidenced by the above RfC, it is unrealistic to try to get any more issues addressed in the green-div. I propose a straw poll. How say the community on the following green-div:

  • 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.
  • Write unit symbols properly. The symbol for kilowatt is kW, not Kw for example.
  • Some disciplines often use non-SI units or write SI units differently from BIPM-prescribed format. When the reliable sources in a field normally use SI units, articles should do so; when they do not, articles should follow reliable sources. For instance, it is cc in an article on Honda motorcycles engines and not cm3; the term "micron" (rather than micrometre) is also still in widespread use in certain disciplines. Such non-standard units or unit names 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 be best 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); likewise, most articles should use Celsius or Fahrenheit, not the SI kelvin, for ambient temperatures on Earth.


  • Support The general principals adhere to and reflect real-world practices and best enhance Wikipedia’s efforts to educate its readership on topics and properly prepare them for their continuing studies on the subject. Greg L (talk) 18:53, 17 November 2010 (UTC)
  • Oppose the motorcycle example. While articles in online Consumer Reports, and the manufacturers site for Harley-Davidson and Honda indicate that a substantial amount of rider-oriented literature uses "cc" rather than "cm3", all the literature I saw omits the space between the number and the unit. It would take too many words to explain that in the example, so I would like to see an example that can be explained more concisely. Jc3s5h (talk) 18:59, 17 November 2010 (UTC)
  • Weak Support - If a gross majority of reliable sources surrounding a particular topic/subject use a certain type of unit, that unit ought to be reflected on WP. Seems like it would be the best thing for readers. One thing I'd note though is that rule A) "When the reliable sources in a field normally use SI units, articles should do so; when they do not, articles should follow reliable sources." and rule B) "Use familiar units rather than obscure units" might conflict. What happens if a gross majority of reliable sources use obscure units when dealing with a particular topic? Anyway, I haven't read deeply into this dispute, so I'm only offering qualified support. I don't really get Jc3s5h's "all the literature I saw omits the space between the number and the unit" comment. NickCT (talk) 19:25, 17 November 2010 (UTC)
Well, in one article Consumer Reports writes about a Harley "Powered by a 1200 cc, two-cylinder engine..." (subscription might be required) but in another article they write "...the Honda Rebel and Kawasaki Ninja, have more powerful 250cc engines". Notice that one sentence uses a space between the number and the unit, the other does not. None of the manufacturer sites I looked at use spaces. So maybe we should find a less complicated example. Jc3s5h (talk) 19:38, 17 November 2010 (UTC)
Ummmm..... Is your only argument with the unit "cc"? It seems like what Greg L is proposing deals more with units in general. I don't understand how what Greg L is saying conflicts with your cc exmaple. In fact, in my reading of the language above it seems like "250cc" would be the correct way to display it, as the majority of reliable sources display it that way.
What's the conflict here? NickCT (talk) 19:56, 17 November 2010 (UTC)
I have two issues: readers unfamiliar with motorcycles or small lengths in certain field will just have to take the MOSNUM editor's word for it that these examples are valid; no source is provided to show these examples are true. Providing such a source could also serve as an example of how editors should handle any disagreements that might arise. Furthermore, we have not delved into whether nonstandard unit punctuation may be adopted in Wikipedia as well as nonstandard abbreviations. If a certain field usually omits the space between a number and a word, should Wikipedia follow suit? If a certain field indicates certain units are multiplied by putting a hyphen between them, (kW-hr), should we follow suit? By using motorcycles as the example, we open up that whole discussion. Jc3s5h (talk) 20:05, 17 November 2010 (UTC)
Ok. I think I get where you're coming from. Here are counterpoints/questions regarding your points,
"readers unfamiliar .... examples are valid" - Ok. So do you have an alternative to suggest?
"no source is provided to show these examples are true" - It seems like it would bulky to include a bunch of references to the fact that "cc" is the common unit. References aren't typical for policy pages.
"we have not delved .... nonstandard abbreviations" - Yeah... This seems like a valid point. Something should be mentioned about punctuation/formating of units. Perhaps change "Write unit symbols properly. The symbol for kilowatt is kW, not Kw for example." -> "Write unit symbols properly, in a manner that reflects how they are written in the majority of reliable sources. The symbol for kilowatt is kW, not Kw for example."
You know, I haven't read deeply into this debate, so forgive a rather uninformed opinion, but I imagine a simplier policy might just be -
Units in articles for any given topic should be used in a way that reflects thier usage in the majority of reliable sources for that topic
This to me seems to be simplier way of saying "Use what RSs use". Let editors figure out how to apply the rule after that. NickCT (talk) 20:40, 17 November 2010 (UTC)
Another problem with nonstandard punctuation is whether it impedes showing a clear majority exist. Let us say a large sampling of rider-oriented motorcycle literature were sampled, and we found 30% used "250cc", 20% used "250 cc", 20% used "250 c.c.", 19% used "250c.c.", 10% used "mL", and 1% used 250 cm3. Do we decide there is no clear majority, so use "mL"? Do we decide there is a majority on the letters, but not the punctuation, so treat it as if it were an SI unit ("250 cc")?
As for references being bulky, they need not be, if we can find a recent reliable source that explicitly says a certain nonstandard abbreviation is generally used in a certain field, despite it being a non-SI unit. This is a higher standard than merely finding a reliable source that uses a non-standard abbreviation, with no mention of what other sources do. Jc3s5h (talk) 21:34, 17 November 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Now we’re arguing about spaces between the numeric value and the unit symbol “cc”??? Just how is it that we so easily allow ourselves to fall into playing another editor’s argument-game? We are not here to solve whether or not Harley Davidson or Honda or Wikipedia are all washed up for leaving a space out from between the “750” and the “cc” or any other such silly details. It suffices to say that List of Honda motorcycles uses “cc” and the practice was clearly not the product of some CGS-nazi who snuck around when everyone else was asleep and changed hundreds of articles overnight. If one follows the links from that list, one comes for instance to CB750A Hondamatic, which has a template box that reads 736.6 cc (44.95 cu in) and doesn’t even link the “cc”. Go fight these battles about the missing space (or excess space depending upon your SI religion) or the lack of the link on the Honda motorcycle pages. If the community there discusses the issue and decides by general consensus to ignore the principal outlined here that “cc” should be linked on the first occurrence or in the template box, that’s their prerogative. The only principal we’re acknowledging and addressing here is that non-SI or specialty units like “cc” are fine for those disciplines that consistently use them. We are not worrying about the atomic-level details of what goes on in the motorcycle world; that’s left for the wikipedians specializing in those Honda articles. Greg L (talk) 03:18, 18 November 2010 (UTC)

  • Neutral. While I agree with all it says, it's more verbose than what's currently in MOSNUM and I think that's unnecessary. (And if such an experienced editor as Greg L has difficulties locating the bullet about symbols for plural units, that means that MOSNUM is already waaaay too bloated.) The point about ambient temperatures 1) is so completely obvious it'd better be left out, per WP:BEANS, and 2) it definitely doesn't belong in a section titled "Scientific and technical units", does it. A. di M. (talk) 22:50, 17 November 2010 (UTC)
    • I don't insist on it; but it is an example of using familiar units, and if we are going to say anything on Celsius, it seems to be as much as consensus will support. If we can get by saying nothing on Celsius, that's fine by me. Septentrionalis PMAnderson 15:30, 18 November 2010 (UTC)
  • Comment. Here is a reliable source that uses "cc" (without a space) to describe motorcycle engines. -- SWTPC6800 (talk) 04:53, 18 November 2010 (UTC)
"New Raider Hits the Boulevard". American Motorcyclist (American Motorcyclist Assoc) 61 (11): p. 18. November 2007. ISSN 0277-9358. "The air-cooled, fuel-injected, four-valve-per-cylinder 1,854cc motor …"
Also in the name of the group 10cc. McLerristarr | Mclay1 05:18, 18 November 2010 (UTC)
Probably a effect of using 1854 cc as a compound adjective. 9mm rifle also seems fairly idiomatic; the closure replaces the hyphen in 9-millimeter ammuntion. Septentrionalis PMAnderson 15:30, 18 November 2010 (UTC)
And "5000m relay" and "5000 m relay" are about equally popular. Art LaPella (talk) 21:01, 18 November 2010 (UTC)
  • Support. It's good to reflect real-world practices.  GFHandel.   18:23, 18 November 2010 (UTC)
  • “Real-world practices”. Yes; I agree. Like “megabyte” instead of “mebibyte” (no matter how good of an idea it was). This new wording is making only two minor changes: stop writing mHz when it is properly MHz, and it’s Celsius (not centigrade). And it also stops speaking of “automotive” in relation to the example used for “cc” and instead uses Honda motorcycles. That’s all that’s being changed. For years now, the basic principle has been to show a strong preference for SI. It is unrealistic to expect a wholesale change to the basic governing principle for making exceptions to the rule, which is to show wikipedian specialists in their respective fields how to resolve the issues over what units to use on their own; that being: FCL, for Follow Current Literature. Wikipedia should go with the practices observed by the clear majority of RSs for any particular subject. I don’t want to get bogged down in holy wars over nautical miles any more than anyone else here at MOSNUM; leave it up to the specialists in marine matters to argue point of fact. The existing wording (with only the above minor tweaks) outlines a realistic, workable framework that gives proper guidance to others so they can solve their own problems. Greg L (talk) 19:33, 18 November 2010 (UTC)
  • Support. Per Greg L and GFHandel. The proposed ideas are just good common sense. SteveB67 (talk) 03:52, 19 November 2010 (UTC)
  • Support. At present, our units are all over the place, whether in form or in substance. Time to put this house in order. --Ohconfucius ¡digame! 06:48, 19 November 2010 (UTC)
  • Support It is not our role to promote a particular form of displaying units, and absolute consistency over the wide range of topics covered on Wikipedia is unachievable and would conflict with the usefulness of individual articles. Consistency is good when it aids comprehension, but it has no other purpose. Johnuniq (talk) 01:31, 20 November 2010 (UTC)
  • Support I support anything that recognizes that units are endeavor-specific and doesn't try to cram SI units where they don't belong. Gigs (talk) 00:34, 23 November 2010 (UTC)

Talk pages by size

Please see the new page Wikipedia:Database reports/Talk pages by size (to be updated weekly).

Perhaps this will be a motivation for greater efficiency in the use of kilobytes.
Wavelength (talk) 21:09, 20 November 2010 (UTC)

  • Perhaps a realization of how warring religious factions fighting each other in Iraq affects plant life in the region would make the combatants play nice. That would be *extra special*, wouldn’t it? But thank you for galloping in on your tall white steed—nostrils flaring in the morning mist—to point out to the mere minions here doing the heavy lifting of debating guidelines on MOSNUM that they shouldn’t be so wordy. Your words look exceedingly wise, but methinks you might dismount from your tall steed as you blocketh the sun for those down in the trenches. Wikipedia is a hobby engaged in by volunteers. Editors are endeavoring to find a way to streamline and shorten WP:MOSNUM and the accompanying debate on the talk pages (like WT:MOSNUM) about how to go about the task of improving the work product sometimes results in protracted debate; talk pages like WT:MOSNUM tend to get filled with thought. Not everyone can communicate with the laser-like surgical precision you apparently exhibit, Wavelength. But we will all try to make you proud and do a better job of living up to your expectations. Greg L (talk) 21:14, 20 November 2010 (UTC)
    • Or to be shorter: all this really records is whether auto-archiving is set and what its parameters are.Septentrionalis PMAnderson 22:54, 20 November 2010 (UTC)
      • I had a shorter version (two words) but I would have been taken to ANI and had my tongue cut out for using the sort of plain-speak with which we are all familiar in real life. And then I might have been taken to The Inquisition (ArbCom) where my most coveted body parts might be amputated. Greg L (talk) 00:53, 21 November 2010 (UTC)
      @PMAnderson: No, it doesn't. It's a "[l]ist of talk pages by their total size, including subpages" (emphasis added). A 16 MB page would crash many browsers indeed. A. di M. (talk) —Preceding undated comment added 01:44, 21 November 2010 (UTC).
      It's been spammed all over the place. All it really reflects is what project pages were established a long time ago and have extensive archives. What does "greater efficiency in the use of kilobytes" even mean? Franamax (talk) 02:04, 21 November 2010 (UTC)
  • More specificity: Two words and only seven letters. Greg L (talk) 02:05, 21 November 2010 (UTC)
Remember, the Inquisition is ever vigilant, and I tend to sympathize with it. You meant "I differ", right? As for what the list reflects, my user talk would have to be about 8 times longer to be on that list, and it started in 2005. Sure we could eliminate some unnecessary conversation, but I'm not sure how the list helps, or even that this very section is one worth keeping ... Art LaPella (talk) 05:03, 21 November 2010 (UTC)
You could have saved several kilobytes :-) without using the F word by using a two-letter word, a four-letter word and a question mark. (OTOH, there have been many discussions on this page – including pretty bulky ones – which lead to bloody nowhere (is bloody wrong with the Inquisition?) and I do think that it would be more productive for editors to improve articles – or to do something off Wikipedia – than to take part in them.) A. di M. (talk) 22:02, 21 November 2010 (UTC)
Greg L, I did not present myself as an overlord addressing minions, and I did not exclude myself from possibly being motivated to achieve greater efficiency in the use of kilobytes. I simply offered a suggestion, as any editor can do.
Wavelength (talk) 00:44, 23 November 2010 (UTC)
Very well; thanks. In light of that sentiment, I regret having taken you to task as I did and apologize. Peace, bro. Greg L (talk) 00:56, 23 November 2010 (UTC)

A "600-metre (2,000 ft) hill with a 650-metre (2,100 ft) hill" ?

Under "How to present the units" in Units of measurement, we have this example:

  • Avoid inconsistent usage. Write a 600-metre (2,000 ft) hill with a 650-metre (2,100 ft) hill, not a 2,000-foot (610 m) hill with a 650-metre (2,100 ft) hill.

What does "a 600-metre hill with a 650-metre hill" mean? I suspect that this is just a long standing typo, and that it should be something like "a 600-metre hill with a 650-metre peak/circumference/plateau/...", but not being a geomorphologist, I don't know. Could someone either fix the example, or explain what it means? Mitch Ames (talk) 06:12, 21 November 2010 (UTC)

Unless I'm overlooking something: Sights visible from the town include a 600-metre (2,000 ft) hill with a 650-metre (2,100 ft) hill. Two hills. Art LaPella (talk) 06:34, 21 November 2010 (UTC)
It doesn't make sense in several respects. Better two different measurements: Sights visible from the town include a 610-metre (2,000 ft) hill more than 20 miles (32 km) to the south". Tony (talk) 06:54, 21 November 2010 (UTC)
It is better that they are the same measurements as it emphasises the point. Maybe changing "with" to "next to" or "alongside" would make it read better. wjematherbigissue 07:36, 21 November 2010 (UTC)
But it's more likely to occur when it's not quite as obvious; i.e., different units close by, or same units at some distance. The same units adjacent to each other in the example are too obvious, and won't make readers alert to the more likely situation. But it's no big deal. Tony (talk) 08:36, 21 November 2010 (UTC)
The trouble is that mixing imperial and metric for different measures (exactly like in your example) would be perfectly acceptable in, for example, a UK based article. wjematherbigissue 13:03, 21 November 2010 (UTC)
Should one be mixing symbols with text that is written out in full. Surely one should either write "They could 650&nbsl;m (2,100 ft) hill" or "A 650 metre (3,100 foot)" hill, but not mix the two. Martinvl (talk) 14:28, 21 November 2010 (UTC)

The guideline speaks to not mixing systems used for primary (non-parenthetical) units of measure. The sentence fragment used for the {{xt}} example can’t be expected to convey a complete coherent thought precisely because it is a fragment. But, I could suggest the below to address Mitch’s observation:

Write They could see the peak of a 600-metre (2,000 ft) hill from a nearby 650-metre (2,100 ft) hill, not …a 2,000-foot (610 m) hill from a nearby 650-metre (2,100 ft) hill.

I’ll put it in now since doing so ought to be uncontroversial. Greg L (talk) 16:53, 21 November 2010 (UTC)

That does indeed address my issue. Thanks. Mitch Ames (talk) 12:19, 22 November 2010 (UTC)

AD/CE as disambiguator

An editor added:

(CE or AD may be needed to avoid ambiguity when the date is during the first few centuries of the Common Era. For example, "Sosigenes the Peripatetic was a philosopher living at the end of the 2nd century AD.")

Why is that necessary, or even helpful? — Arthur Rubin (talk) 16:15, 15 November 2010 (UTC)

As I found the text, it gave an example of where no disambiguation was necessary (Norman Conquest of 1066), but no example of where any well-edited reference work (including Wikipedia) should provide CE/AD to avoid ambiguity. In principle, since ambiguity is possible, I thought it would be good to give an example of where it exists. Since the Peripatetic school was going strong in the 2nd c. BCE and in the 2nd c. CE, it is quite ambiguous to say that Sosigenes is a Peripatetic writer of the 2nd century, so those who write about antiquity would not omit it. If this isn't persuasive, I need a fully spelled out argument about the non-necessity and non-helpfulness of such commonly provided disambiguators, because I can't figure it out. It is certainly not the case that "2nd century" in English refers by default to the 2nd century CE. It is ambiguous, and the disambiguator is normally left off only where context leaves no room for doubt. For readers looking up new subjects, we have to assume a little doubt. (I intended to provide a good example but not add any guideline. If I were to attempt a guideline, I'd say that disambiguation is common in good scholarly reference writing for the first through fifth centuries.) Wareh (talk) 16:20, 15 November 2010 (UTC)
I am of the opinion that AD/CE should be required as a disambiguator only for individual years up to 999, not centuries. Ex. — 999 AD/CE, 1000, 1st century, 2nd century. — CIS (talk | stalk) 16:38, 15 November 2010 (UTC)
Let me just interject that I didn't want to speak of any requirements. All I believe is that when subject editors who work in antiquity have added CE/AD to a low-numbered century to prevent ambiguity, no one should have the impression that this guideline advises against such disambiguation. There are many contexts, I'm sure, where years 1-999 require no disambiguation (especially in the upper part of that range); I'm here to insist that there are many contexts in which centuries should have disambiguation upon the judgment of informed and neutral editors. Wareh (talk) 16:55, 15 November 2010 (UTC)
Well, actually, it is the case that the "2nd century" in English refers by default to the 2nd century AD, or sometimes the 2nd century of life. It never refers to the 2nd century BC. I could see the 1st century or 2nd century being ambiguous with the 1st or 2nd century of some other entity, but not to the 2nd century BC. — Arthur Rubin (talk) 17:09, 15 November 2010 (UTC)
"It never refers to the 2nd century BC." This is falsifable, and in fact false. In discussing classical antiquity "2nd century" etc. are ambiguous and very frequently refer to the 2nd century BC. E.g. these books[4][5] have titles in English, yet with no "BC." Once we accept that "2nd century" etc. do often enough refer to BC, surely we'll see the point of disambiguation. I am a Classics scholar myself and sometimes experience the annoying situation that my own ignorance requires me to do further research to clarify an ambiguous reference the author considered unambiguous; our readers should not generally be put into that position. Finally, if the "never refers" were true (though I've already falsified it), we could simply delete "unless the date would be ambiguous without it," because ambiguity would be impossible! We might wish for that level of tidiness, but we can't impose it unilaterally on the English language. Wareh (talk) 17:35, 15 November 2010 (UTC)
I agree with Wareh completely. An article on Augustan Rome, for instance, will straddle BC/BCE and AD/CE. Sometimes 1st century will be the former, and sometimes the latter, and you can't assume the outside reader not initiated into WP ways knows the secret handshake: you have to specify the era for both. The purpose of stylistic consistency is clarity and accuracy; when the application of a "rule" creates ambiguity, it's an obvious case for an exception. To make an analogy, I've worked in newsrooms for daily and weekly papers that followed the stylebook of the Associated Press as holy scripture; however, all newsrooms have an internal stylebook of exceptions in order to accommodate a localized frame of knowledge. In our case, it's usually self-evident which era is meant, except when you're in an article, or disambiguating situation, that falls within the three or four centuries before and the three or four centuries after, and both eras are in play within the topic; then it becomes usefully informative to specify. The opposite is also true: if you're writing an article with a lot of dates in one paragraph, say a biography, and it's established that this person is living in the 1st century BC/BCE, it's perfectly clear if you don't use the era for every single numerical year: "Flaccus was a quaestor in 97 BC. He was a praetor in 87 BC, and after that a governor in Asia Minor throughout the 80s. His last known assignment was in 78, and he died in in 72." In this case, nothing is added by continuing to use the era, since no one would assume he died in 72 AD/CE, and the only thing diminished by omitting it is cluttered, naive prose. Cynwolfe (talk) 18:12, 15 November 2010 (UTC)
I see your point, although scholarly English can have completely different disambiguation requirements than ordinary English. Just consider the term field: in ordinary English, it's almost always field (agriculture) or field of study; while in mathematics, it's always field (mathematics). — Arthur Rubin (talk) 18:44, 15 November 2010 (UTC)
For some values of ordinary, the playing field meaning is far more common than the agriculture one, and for some values of mathematics, it's not "always" as it can also refer to tensor fields or special cases thereof. But one of the weirdest things is that, although field has umpteen completely different meanings, pretty much all occurrences of it (except in proper names) can be satisfactory translated in Italian as campo. —A. di M. (talk) 19:06, 15 November 2010 (UTC)
But if someone reading the Flaccus article were to search Wikipedia for 80s, 78 or 72, they would be redirected to the AD/CE year(s) in every instance. We cannot allow for such an ambiguity. I stand by my position that AD/CE should only be used with the first 999 years of the era, and that "1st century AD/CE" should only be used when there is a direct contrast with a BC century. — CIS (talk | stalk) 18:52, 15 November 2010 (UTC)
Thank you for stating so clearly the position that Wikipedia must conform to (a personal interpretation of) the Manual of Style, even when it means actively misinforming the reader. That is the sort of nonsense WP:IAR was written to prohibit; we are here to write an encyclopedia, not to impose an arbitrary set of rules.
If date links from the Flaccus article were necessary - and it is the present position of MOS that they usually aren't - it is the responsibility of the editors to link correctly at both ends, neither distorting the text linked from nor conveying the reader to the wrong article. Fortunately, [[72 BC|72]] is not hard to type. We should above all be accurate, verifiable, sourced, and clear. Septentrionalis PMAnderson 19:20, 15 November 2010 (UTC)
For a reader to whom it does not occur to hover on the link (or who can't do that, e.g. because they are using a mobile phone or reading a hard copy), "72" is no clearer than "72", though. A. di M. (talk) 19:33, 15 November 2010 (UTC)
Quite so; as Cynwolfe said, good writing will only use 72 if it is already clear from context (here the earlier offices in 97 and 92 BC) which one is plausible; similarly, one should only use 73 by itself in discussing Masada if it is clear in context that it is AD. Septentrionalis PMAnderson 19:49, 15 November 2010 (UTC)
In any case, Septentrionalis makes a good point, which is that e.g. this edit (which brought this issue to my attention), at once confuses the reader about whether the reference is to BC or AD, and takes away one possible means of resolving the confusion (the wikilink). I agree with delinking dates, but I don't think we can expect readers to type "2nd century" in the search box, just to determine that Wikipedia as opposed to published books on the same subjects has a special system whereby the reader is supposed to know that "2nd century" always means "2nd century AD." (It's not just scholarly writing by the way: it's in any writing that has occasion to refer to these centuries--it just happens that scholars talk more about the 2nd century CE and BCE than most others!) In short, a reader of Sosigenes the Peripatetic after this change would have to (1) know the date of Alexander of Aphrodisias, or (2) consult the categories carefully, in order to know when Sosigenes is thought to have lived. I don't think Wikipedia can impose "all dates are CE unless otherwise noted" on our readers who come to us from Google etc., since (A) that is not the normal English practice elsewhere, and (B) it would, reasonably, require a banner on every article about the early centuries CE to remind them (such a banner would be no less called-for than notices that you may be missing IPA font characters, etc.). Wareh (talk) 19:36, 15 November 2010 (UTC)

On the substance of the matter: What MOS now says is that AD and CE should not be used "unless the date would be ambiguous without it". But fourth-century Greek writer is ambiguous; the first page of Google Book hits for the phrase includes Xenophon and Ephorus (of the 300s BC, although the text quoted doesn't say so) and Athenaeus and Cyril of Jerusalem (of the 300s AD). Why should our readers have to guess which is meant for Aeneas Tacticus or Longus or Horapollon or Theron? The solution may be to include one of these as a bad example (probably Theron, because there is a genuine controversy whether he is third century AD or BC, which adds a layer of richness). Septentrionalis PMAnderson 19:42, 15 November 2010 (UTC)

I agree with Wareh, Cynwolfe and Septentrionalis. Paul August 19:49, 15 November 2010 (UTC)

The important thing here is that, in Greek literary history, "third century" is not by default AD; until quite recently, it was by default BC, because the Greeks under the Roman Empire were held to be degenerate and uninteresting. I don't know the literature on China well enough, but it may do the same thing; Confucius is more studied that the literature of the Sui. Septentrionalis PMAnderson 20:17, 15 November 2010 (UTC)

Very well, I withdraw my objection to the change in the MOS, except I'm not sure that example is appropriate. I'm still not convinced that it's for the best, but it makes sense. — Arthur Rubin (talk) 05:45, 16 November 2010 (UTC)
Is the inappropriateness that Sosigenes is obscure? I agree, but am not sure who would be clearer. Most of the examples I can think of to replace him will still be obscure to the average reader: Lucian? Claudius Ptolemy? Athenaeus? Procopius? Plotinus? Please suggest one. Septentrionalis PMAnderson 16:04, 16 November 2010 (UTC)
I suggest the example of Augustus, whose rule spanned the eras. Also, there is the example of Halley's Comet, which is visible from Earth every 75 or 76 years.
Wavelength (talk) 16:21, 28 November 2010 (UTC)
  • Other than one minor change to the content, this discussion seems to have trailed off to nowhere land. It would be helpful to have more precise statements in this MOS regarding use of AD/CE as disambiguator. Hmains (talk) 01:24, 27 November 2010 (UTC)

Positive rewrite suggestions

Currently: Do not use CE or AD unless the date would be ambiguous without it (e.g. "The Norman Conquest took place in 1066" not 1066 CE or AD 1066).

Suggested: CE or AD should only be used to prevent ambiguity, which generally arises only with the first several centuries CE/AD: e.g. [acceptable example from antiquity]. CE/AD is not used with more recent dates: e.g. "The Norman Conquest took place in 1066," not 1066 CE or AD 1066.

I'm optimistic, following Arthur Rubin's agreement, that we are pretty much ready for this discussion, though I don't want to be the person to add back the text I wrote before. The bottom line is that the guidelines should give an example of an appropriate CE/AD, so that editors can understand the real possibility of ambiguity, and remove superfluous CE/AD's while leaving disambiguating ones in place. Wareh (talk) 15:15, 16 November 2010 (UTC)

  • As written, the proposal above says 'dates', to which some editors would say 'centuries' are not included. If centuries are to be included in this rule then this should be so stated with a century example. Also, the MOS used to say that if a BC/BCE date was used in an article, and the next date was a AD/CE date, then the latter date should always be marked as AD/CE. I see that many articles still follow this rule though the MOS no longer has it. I think something like this rule should be re-added. There is also the situation of an article written with numerous dates going back and forth between BC/BCE dates and AD/CE dates. Should every date/century in such articles be marked with BE/BCE and AD/CE to prevent confusion to the reader? Hmains (talk) 06:40, 18 November 2010 (UTC)
    What, do you mean that Rome#20th and 21st centuries should use AD before each year on the grounds that Rome#Monarchy, Republic, Empire contains BC dates? A. di M. (talk) 08:16, 18 November 2010 (UTC)
Not at all. Only the Rome#Monarchy, Republic, Empire sections contains enough BC dates to consider putting the AD notation on all its AD dates. In other sections, I would reduce/eliminate the existing AD notation as no ambiguity should exist there for the reader. Hmains (talk) 03:59, 19 November 2010 (UTC)

Other matters of substance: AH years

To further confuse things, we have many Arab/Islam articles written with AH indicating Hijri year. These articles have AD/CE dates also included for the AH dates--some/many/all depending on the article. To avoid confusion, should all AD/CE dates be labeled as such in any article that contains any AH dates? Similarly, should Arab/Islam articles without AH dates also have all their AD/CE dates labeled as such so that readers will not think that AH dates are being assumed in the writing? This MOS does not cover this. Hmains (talk) 01:09, 27 November 2010 (UTC)

Ditto what I said below. McLerristarr | Mclay1 16:35, 28 November 2010 (UTC)

And more matters of substance: V.S. and B.S. dates

And we have Bikram Samwat dates in many South Asia articles (India, Nepal, etc). These are variously indicated by V.S. or B.S. and vary from AD/CE dates by years/months/days, not just years. These articles have AD/CE dates also included for V.S./B.S. dates--some/many/all depending on the article. To avoid confusion, should all AD/CE dates be labeled as such in any article that contains any V.S./B.S. dates? Similarly, should South Asia articles without V.S.,B.S. dates also have all their AD/CE dates labeled as such so that readers will not think that V.S./B.S. dates are being assumed in the writing? This MOS does not cover this. Hmains (talk) 01:19, 27 November 2010 (UTC)

In my opinion, non-international dating should not be used at all, unless it is important for the context, in which case it should be in brackets after a normal date (normal for most English speakers). If we put normal dates first, AD/CE should follow the same guidelines as normal. McLerristarr | Mclay1 16:34, 28 November 2010 (UTC)

Percent or per cent

I notice that "per cent" was removed from the sentence "Percent or per cent are commonly used to indicate percentages in the body of an article". This was reverted but then removed again. Yet MOS:PERCENT still includes the "per cent" variant. "Per cent" appears in the OED, which notes that the "percent" form is chiefly American English. As the manual of style shouldn't be prescribing which variant of English to use, and because "per cent" is in wide use, I suggest that we reinstate it here. Cordless Larry (talk) 16:10, 22 November 2010 (UTC)

Oxford Dictionaries Online says per cent is chiefly British and percent is US. It says both are acceptable. They should definitely both be acceptable on Wikipedia. Perhaps we should add a note about which to use per WP:ENGVAR. McLerristarr | Mclay1 16:31, 22 November 2010 (UTC)
WP:MODLANG also suggests that "per cent" is archaic, so I've pointed this out here as well. Cordless Larry (talk) 16:35, 22 November 2010 (UTC)
I've removed it. I thought I'd done that a while ago... McLerristarr | Mclay1 16:57, 22 November 2010 (UTC)
In the British National Corpus, excluding the "spoken" section for obvious reasons, "per cent" occurs 9 times and "percent" 508 times. In the Corpus of Contemporary American English the figures are 2521 and 166799. (There must be something wrong because I can think of no plausible explanation for the three orders of magnitude difference in the totals.) A. di M. (talk) 19:38, 22 November 2010 (UTC)
The explanation is that the spaced version is clearly obsolete, just like to-morrow, R.A.D.A.R. and forsooth. — SMcCandlish Talk⇒ ʕ(Õلō Contribs. 19:48, 22 November 2010 (UTC)
It is very obviously not obsolete, as demonstrated by the Google News search I linked to above. This clearly shows that UK newspapers, amongst other sources, continue to use "per cent". This is the variant listed in the style guides of The Guardian, The Times, The Telegraph and the BBC. Cordless Larry (talk) 21:22, 22 November 2010 (UTC)
That's not what I meant. Since the COCA is only 4.1 times as big as the BNC, it appears thatÞ – disregarding spelling – Americans are talking about percentages in average 80 times – i.e (2521 + 166797)/(9 + 508)/4.1 – as often as Britons. That can't be right. A. di M. (talk) 00:38, 23 November 2010 (UTC)
Where are you accessing the BNC, A. di M.? I've never used it before but I can find 2,900 results for "percent" versus 38,069 results for "per cent". Cordless Larry (talk) 09:07, 23 November 2010 (UTC)
I was using the BYU interface. It must be broken, or I'm missing something: I'll ask its maintainers. (Your search also includes the "spoken" section where the spelling must be an arbitrary choice of the transcriber, as AFAICT there's no difference in pronunciation between per cent and percent; but the spoken part makes up only 10% – no, I won't spell out this one :-) – so it shouldn't radically change the results.) A. di M. (talk) 13:06, 23 November 2010 (UTC)
The maintainer of the interface I used answered: '"per cent" is a BNC multi-word expression (http://www.natcorp.ox.ac.uk/docs/multiwd.htm), and these don't work well in my interface (it's a function of how they tagged the corpus back in the 1990s).' A. di M. (talk) 19:24, 23 November 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Fine Cordless Larry. You have established that some English sources use “per cent”. In light of A. di M.’s evidence, above, I think it is entirely proper to expect you to provide convincing evidence that *most* British-English sources use that spelling before it becomes yet another ENGVAR-thing to make a few editors happy. Just because some subscribers on an island might read 55 per cent of readers voted “LOVE ‘EM” for the jugs of Page Three girl #2 comes up short. I can’t see why any practice that isn’t universally observed by people on a group of islands should be adopted as a *good practice* here until then.

I note too that in the French language, it is “cinquante pour cent” for “fifty percent.” Nevertheless, for English-language publications, the French BIPM recommends “percent” and doesn’t mention “per cent” in their manual on SI writing style, 5.3.7 Stating values of dimensionless quantities, or quantities of dimension one. Note that this spelling is from the people who don’t go with the ‘American flow’ on the spelling of “meter”. Greg L (talk) 00:15, 23 November 2010 (UTC)

I don't think it matters how many people use "per cent". Since reliable sources use it, we should not be preventing its usage. It does no harm. McLerristarr | Mclay1 00:21, 23 November 2010 (UTC)
ENGVAR doesn’t look towards “some RSs” as a litmus test of what spelling practices are approved for use here. If we did, we would be allowing editors to use all sorts of obsolete British spellings because some RSs in England still use them. If it is to be sanctified for use on Wikipedia as an ENGVAR-based issue, a spelling variant should be widely used. Now the test is to establish the true facts in that regard. If A. di M. is correct, it isn’t looking good for “per cent.”

As for your I don't think it matters how many people use "per cent"-comment, have you considered the scope of such a statement? There are a number of places where English is now a peoples’ first language. Just because some people in the Seychelles (see List of countries where English is an official language’) might have a weird spelling for some word (they have their own RSs on the Seychelles don’t-cha know) is no reason for spelling chaos to confuse the majority of our readership. Remember, Wikipedia is about doing what is best for our readership; it is not about making every single volunteer editor who is still in 8th grade happy as long as he has sufficient intelligence to register on en.Wikipedia. Greg L (talk) 00:30, 23 November 2010 (UTC)

Firstly, I don't consider the three most respected UK newspapers and the national broadcaster as just "some RSs". I don't think it's really necessary to prove that the majority of British sources use "per cent", just that a significant number do. However, Google suggests that "per cent" is in wide use (not just in the UK). A news search for percent returns 80,700 results, versus 46,200 for "per cent". Cordless Larry (talk) 07:54, 23 November 2010 (UTC)
But searching "per cent" also finds hits for "percent" and vice versa, and Google hit numbers are notoriously bogus when larger than a few thousands. A. di M. (talk) 08:04, 23 November 2010 (UTC)
Looking through the first few pages of results, there doesn't seem to be crossover between "per cent" and "percent" so I think the search is reliable in that sense. I defer to you about hit numbers though. I still don't really think that the numbers matter than much though, given that the OED states that "percent" is chiefly the US variant and that the UK's major news sources use "per cent". Cordless Larry (talk) 08:49, 23 November 2010 (UTC)
Weird: the first time I looked at the result there was at least one "percent" in the first page of results for "per cent" and vice versa. So either Google or I (or both) are even more erratic than I thought. A. di M. (talk) 13:13, 23 November 2010 (UTC)

The correct BNC data (thanks Cordless Larry) settles it, IMO. Unless we have some evidence that it changed since 1993, British usage is "per cent" in more than 90% of the cases (though my spell checker doesn't mark "percent" as incorrect). A. di M. (talk) 13:17, 23 November 2010 (UTC)

  • More facts: Notwithstanding the BIPM’s declaration (¶ 5.3.7) of the correct spelling for international consumption of English-language material is “percent”, here are “Advanced” Google searches: Find “percent” and exclude "per cent" = 487 million hits. And this one: find "per cent" and exclude “percent” = 176 million hits. Given that the BBC uses “per cent”, it is clear that it is a legitimate British spelling variant.

    Grokking the mix: I would concede that just because the BIPM says it is “percent”, doesn’t mean they are correct on all matters; they are still being ridiculously obstinate that the unit symbol for the term have a space between it and the value (75 %). However, the BIPM stays out of spelling wars with regard to words used in general prose and only weighs in on spelling that is germane to communicating measures to an international readership consistently and unambiguously. So issues pertaining to the rule of SI and its writing style should certainly not be dismissed easily.

    The test: It seems that only decision lies in balancing the BIPM’s recommended practice for international consumption of English-language material against allowing editors from all over the English-speaking world to contribute to Wikipedia without untoward discouragement. So what MOSNUM guideline best serves the interests of our global readership?

    Suggestion: In this case, I would suggest that we follow the BIPM’s advise. Even in the case of MOSNUM’s ignoring the BIPM with regard to the space before the percent symbol, we are clearly being consistent about which practice to use across the project. But I don’t feel strongly one way or another because I don’t know how discouraged our British contributors would be over this issue. I would think they could take the BIPM’s recommended practice on the chin in this instance, but I might be wrong. Greg L (talk) 16:49, 23 November 2010 (UTC)

"I don't know how discouraged our British contributors would be over this issue". I obviously can't speak for anyone else, but as a British contributor I would never write "percent" on Wikipedia or away from it (and my work has involved using and writing about statistics) without being forced to do so by a style guide. Even then I would do so reluctantly because I have always been taught that the correct version is "per cent". Cordless Larry (talk) 17:05, 23 November 2010 (UTC)
The dictionaries, style guides, and corpus counts cited so far show that "per cent" is by far the dominant British spelling, rather than a variant. Also, the two spellings are similar enough that readers are quite unlikely to be confused by seeing a spelling other than their own, so it should be treated like any other word with different British and American spellings, like colour/color, litre/liter, travelling/traveling, and so on. (Maybe there could be some contrived context where per cent would be ambiguous because it could also plausibly mean ‘per hundredth of euro/dollar/semitone/whatever’, but such a situation wouldn't be much more common than American meter being able to plausibly mean both ‘metre’ and ‘measuring instrument’.) A. di M. (talk) 19:06, 23 November 2010 (UTC)
  • In any case, we should encourage using the % sign whenever possible (i.e. except in quotes or at the start of sentences) as it is more succinct and avoids this entire issue. --Cybercobra (talk) 20:44, 23 November 2010 (UTC)
    • Well, that would involve a big change in policy, since the manual of style currently suggests that "percent" or "per cent" are preferable in prose, other than in technical articles. Cordless Larry (talk) 20:47, 23 November 2010 (UTC)

In light of all of the above, may I suggest that we restore the "per cent" option to the text, thus making it consistent with MOS:PERCENT? I think it has been demonstrated that "per cent" is the common spelling in British English, so including it as an option avoids there being a conflict between the manual of style and WP:ENGVAR. Cordless Larry (talk) 08:32, 24 November 2010 (UTC)

OK, done. Cordless Larry (talk) 10:47, 27 November 2010 (UTC)
I tweaked the wording - the way Cordless Larry had it gave the impression that "per cent" is obligatory in British English. We must also not lose sight of the fact that US and UK English are not the only varieties that exist. Roger (talk) 17:34, 27 November 2010 (UTC)
I chose the wording to match that at MOS:PERCENT, which was recently amended to note the American and British English versions (presumably as a result of this discussion), but I agree that perhaps the wording needs to be tweaked. Cordless Larry (talk) 11:31, 28 November 2010 (UTC)
My change was reverted by A. di M. but without addressing my main concern that the way it now stands implies that the "per cent" form is compulsory in British English, which is clearly not the case. I have a secondary concern in that the current wording ignores the existence of other varieties of English - the language is not neatly bisected into "Yanks" and "Limeys". Roger (talk) 12:23, 28 November 2010 (UTC)

Trimming

There was some confusion about the SI standard and how it relates to percent and spelling. I've now added two bullets relating to non-SI units and indicated that mosnum specifies deviations. The improved clarity allowed the removal of duplicate/redundant sections. This goes towards the objectives of making mosnum more succinct and suited to debates on talk pages. As far as I can see, the changes don't change the intended meaning of mosnum. Lightmouse (talk) 15:34, 26 November 2010 (UTC)

I think this version gives too much weight to the BIPM brochure. For example it requires to symbolise the astronomical unit as "ua", but I can't recall any English-language source using that symbol; it requires to call 1 fm a femtometre even if nuclear physicists (which are the people who use that unit most often) call it a fermi, and so on. --A. di M. (talk) 17:55, 26 November 2010 (UTC)
On further thought, the second bullet in "Scientific and technical units" does address that, though I'd retitle it to something broader e.g. "Specialized units" so as to include the nautical mile (who the hell uses the symble "M" for it when writing in English?) etc. A. di M. (talk) 18:02, 26 November 2010 (UTC)

I agree, the section titles are a mishmash. You've also revealed some gaps that were previously there but obscured by the unclear text. It might be better to have a section like "Specific units" which could contain a lot of the detail that is currently scattered. It could list micron, cc, astronomical units, fermi, and any other units where a decision has been recorded. Then another section that describes the process for resolving disputes about units e.g. "When you want to do something else". This would contain the current text about how to measure which units is most popular, most understandable, most cross-cultural, etc. Lightmouse (talk) 20:26, 26 November 2010 (UTC)

It must be remembered that a significant proportion (20%?) of English version W]ikipedia readers do not use English as their home language. The proportion is probably even higher for technical articles. This is, I beleive, good enough reason to stick with the BIPM definitions. Martinvl (talk) 21:39, 26 November 2010 (UTC)
So what? Non-SI names for SI units such as "fermi" or "micron" are also widely used in other languages, and anyone able to understand technical English enough to read the articles will likely know them. (OTOH, I do think that typically English units such as the fortnight or the stone should only be used if there's a good reason for using them, and even then they should have a parenthetical conversion.) A. di M. (talk) 02:18, 27 November 2010 (UTC)