Wikipedia talk:Manual of Style/Dates and numbers/Archive 123

What do others think? Michael Glass (talk) 10:47, 21 May 2009 (UTC)

You are encouraging others to read through the MOS, but have you read it yourself? The "Which units to use" sub-section covers all this quite nicely. As such there is no requirement for the principles to be specific, yet you continue to try and push SI into these introductory statements, while ignoring the fact that SI, and SI related, units are inappropriate in many instances, not just US and UK related articles. This is becoming tedious. _wjemather^bigissue 11:43, 21 May 2009 (UTC)

Seriously, it's like the 4th time your trying to rephrase and re-introduce the same discussion. The

WP Physics

} 12:55, 21 May 2009 (UTC)

Should we really write a 43-centimetre (17 in) CRT television set despite the fact that those things are measured in centimetres in Australia and South Africa but in inches everywhere else? Should we write 0.23-millimetre (0.009 in)-gauge guitar strings despite the fact that no-one measures them in millimetres? C'mon. Saying "international units" is fine. The inch is the internationally used unit for CRTs and it should be used to measure CRTs in non-regional articles. Also, the wording suggests that we should "generally" write stuff like 0.69 gigaseconds (22 years); matter of fact, I can very hardly imagine cases where anyone would ever want to do that. --A. di M. (formerly Army1987) — Deeds, not words. 19:12, 21 May 2009 (UTC)

Saying that you couldn't use minutes, hours, days and years under SI is a load of bovine manure. [8]. To argue this way is nonsensical. However, if people on MOSNUM are determined to stand against the world-wide system, they are welcome to it. I have tried my best to work out a wording that would deal with any rational objection to my words. This is my last attempt.

* Wikipedia is International; in general, apart from US and certain UK related topics, put SI and other units accepted for use with SI first, supplemented, where appropriate, with other national or internationally used units.

The words that are bolded are based on the official policy. Michael Glass (talk) 09:20, 22 May 2009 (UTC)

You are still completely missing the point, and it shows bad faith on your part to suggest that those opposing your attempts at changing the guidelines are opposed to global standards. You must be aware by now that the use of SI (and other units) is well covered in the "

Which units to use" section. I am thankful that you will not be pushing this any further. _wjemather^bigissue

09:51, 22 May 2009 (UTC)

That proposal is identical to the one you made yesterday except that you replaced "SI related units" with "other units accepted for use with SI" (which, as you admitted, were intended to mean the same). So I can't see how it addresses any of the problems which were raised about it. Writing 0.23-millimetre (0.009 in)-gauge guitar strings when guitar strings are measured in inches to within a thousandths of inch everywhere makes no sense regardless of whether the article is about the US, the UK, some other place, or about no place in particular. The same applies to a whole lotta other things, too: if you understand French, just see fr:Pouce (unité)#De nos jours for a list of things which aren't commonly measured in anything else than inches. The same applies to kilocalories, light-years, parsecs and other such units which aren't approved for use with the SI but are the only units most people all over the world would use to measure stuff such as food energy values, interstellar distances, yadda yadda yadda. (As for time units, I did not mention hours and days—which I know are accepted for use with the SI, but weeks and years. Misquoting people won't help supporting your position at all.) Can you just point out exactly what is wrong with the current wording, apart from region-specific topics, use internationally accepted units for the topic at hand (namely, measure CRTs in inches first, unless if you are specifically dealing with Australian or South African CRTs, in which case you would use centimetres first)? --A. di M. (formerly Army1987) — Deeds, not words. 11:33, 22 May 2009 (UTC)

The guideline is international units first, not SI and related units. This 10th (or whichever iteration this is, and I am glad this is finally the last one) proposal is no different than all the ones that came before it. There is no "official policy" that Wikipedia follows. We do not have the luxury of being able to follow one.

WP Physics

} 13:09, 22 May 2009 (UTC)

All the exceptions you refer to are covered by the words "in general". I proposed changing the words because they are too vague. "International units" to me suggests SI units in a way that does not come out and say so directly. In any country that uses metric measures, international units would mean SI. That is the problem with the expression '"international units". It means different things to different people.

Where units other than SI are used to measure things such as guitar strings, or hands are used to measure the height of horses, it strikes me as anomalous, like using Roman numerals. These exceptions are covered by the words "in general".

The expression "region specific topics" is so vague and woolly that it doesn't make any sense to me at all. That is why I wanted to sweep away this gobbledegook and replace it with something that makes sense in plain English. But, and this has been made abundantly clear, some editors want the policy to be vague, so there's no point in pursuing it further.Michael Glass (talk) 13:40, 22 May 2009 (UTC)

I think it is better to use a wording which makes clear which principle we should be following, rather than putting the cart before the horses and just state a rule of thumb ("SI and other units accepted for use with SI") which has zillions of exceptions hoping that the "in general" somehow makes clear not only that they exist, but even what they are. It's almost like realizing that most mathematicians are male, suggesting that "in general, mathematicians should be referring to by masculine pronouns" and claiming that the words "in general" cover cases such as Emmy Noether. (The main difference is that, in my fictitious example, there would be very little—if any—risk that some over-zealous editor could take the rule too seriously and refer to Noether as "he".) "Internationally accepted units for the topic at hand" is as unambiguous as it could possibly be: the units which are most commonly used around the world for the given class of measurements: kilometres per hour for road speeds, kilocalories for food energy values, inches for cathode ray tubes, and degrees Celsius for atmospheric temperatures. "Region-specific topics" are topics where different units are used locally rather than the internationally most common ones, for example road speeds in the US and in the UK, for which miles per hours are used instead of kilometres per hour, or air temperatures in the US in degrees Fahrenheit, or Australian CRTs in centimetres. Now I agree that might not be the most explicit possible wording for that, but if you can find a clearer wording which doesn't change the meaning, you're welcome to propose it. --A. di M. (formerly Army1987) — Deeds, not words. 14:22, 22 May 2009 (UTC)

BTW, I am Italian and I am a university student of physics, so I'm definitely not a person who could want to oppose metrication for the sake of opposing metrication. Anyway, I have gotten sick of this silly discussion, too. Farewell. --A. di M. (formerly Army1987) — Deeds, not words. 14:28, 22 May 2009 (UTC)

Because it fits so well here: I am German, have lived in the UK for barely 2 years and am leaving the country in a few months. I believe the level of resistance to metrication here is silly, and I am suffering from this totally unnecessary problem in that I constantly have to translate miles into kilometres when driving. (The worst thing is that while British cars have dual-unit speedometers, European cars don't show miles per hour. So I have to translate 50 mph into 80 km/h. Then after a while I may ask myself whether it was 80 km/h or 80 mph.)

I am interested in obsolete historical units as a curiosity. It's absurd that the US and to a lesser extent the UK and Canada continue using them, but it's a fact that we have to deal with. --Hans Adler (talk) 19:11, 22 May 2009 (UTC)

If the wording of the guideline contained the phrase "Internationally accepted units for the topic at hand" it would be much clearer. However, it just reads "international units". Why not put that forward as a suggestion? Michael Glass (talk) 16:22, 22 May 2009 (UTC)

Because it already written that way.

WP Physics

} 18:20, 22 May 2009 (UTC)

Yes, it's just been added, though I think the passage needs editing. Please see below: Michael Glass (talk) 04:08, 23 May 2009 (UTC)

I fail to see how the section below is related in any way to the editing of the passage in question. Nor was the recent addition about the passage of the below section. It's a very weird thing to misrepresent.

WP Physics

} 04:46, 23 May 2009 (UTC)

New wording on weights and measures

The recently added wording reads as follows:

• In articles which are not region-specific, prefer internationally accepted units; usually, they are the units of the
  non-SI units accepted for use with SI—for example, 25 °C (77 °F) rather than 77 °F (25 °C)—but there are many exceptions
  for particular classes of measurements.

• In articles which are not region-specific, prefer internationally accepted units. Usually these are the units of the
  non-SI units accepted for use with SI, for example, 25 °C (77 °F) rather than 77 °F (25 °C). However, there are a number of exceptions
  for particular classes of measurements.

Michael Glass (talk) 04:08, 23 May 2009 (UTC)

I've attempted a compromise, because having two consecutive sentences starting with "However" sounds weird to me; also I feel "various" hasn't the same connotations as "many" but isn't as long as "a number of". (BTW, I am glad we've moved on to discuss wording rather than meaning.) --A. di M. (formerly Army1987) — Deeds, not words. 10:08, 23 May 2009 (UTC)

Wording of the policy on weights and measures

I note that Headbomb has restored the earlier wording of the introduction. As the words replaced had some merit, I have tried to bring the two versions together to combine the best of both wordings. Michael Glass (talk) 12:55, 14 May 2009 (UTC)

Frankly, that is no better. If anything it is worse, so I have reverted again. _wjemather^bigissue 13:16, 14 May 2009 (UTC)

Let us look, clause by clause, at the wording that Wjmather rejected:

A:Follow these principle in using units of measurement:

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

A is simple, straightforward and short. Why reject this simple wording in preference for a more complicated sentence? Is anything of importance omitted by using B instead of A? I don't think so. All that is lost is ponderous language.

A: Readers need to understand the measurements, so avoid ambiguity and aim to write clearly.

B: Avoid ambiguity: Aim to write so you cannot be misunderstood.

Yes, A is longer but B tells us to aim for the unattainable. No-one can write so they cannot be misunderstood. Why not substitute the more modest aim, to write clearly? But let us say that you were not happy with A or B. Then why not try to recast the sentence? That would require work, of course, and I challenge Wjmather and others to work with me on this wording.

A: Different readers are familiar with different measurements, so both metric and US/Imperial measurements should be supplied.

B: Familiarity: The less readers have to look up definitions, the easier it is to be understood.

Yes, A is longer, but it points out that different people are familiar with different measurements and that is why both metric and US/Imperial measures should be supplied. B is shorter but I doubt that it is better, because it offers no solution to the problem noted.

A: Wikipedia is international; apart from region-specific topics, use international units.

B: International scope: Wikipedia is not country-specific; apart from region-specific topics, use international units.

Once again A is simpler, more straightforward and shorter. Why reject it in favour of the longer and more ponderous B?

Reverts are easy and quick, of course. However, perhaps if Wjmather and others could work with me we could come up with something that is better than either A or B. What about it? Michael Glass (talk) 10:44, 15 May 2009 (UTC)

To be totally frank: So soon after your wearisome attempts at rewriting this guideline so that it conforms to your misreading of it, I don't think you are going to find much cooperation for minor stylistic tweaks. And "both metric and US/Imperial measurements should be supplied" instead of "familiarity ..." is more than a tweak, it's a change of meaning. 1) It can be read as saying that where applicable all three measurements should be applied, as in "1 pint (0.57 litres, 1.2 US pints) of cider". 2) It doesn't address the often heard argument that a unit is linked, so readers can simply look up its equivalent in another system. 3) It doesn't provide for historical articles where the most natural units to use may be obsolete ones that are neither current US nor Imperial units. For example a historical article on minting in England would in one period naturally measure the gold taken to the mint in pre-imperial troy pounds and the finished coins in Tower pounds. It would give metric and possibly Imperial equivalents in parentheses. 4) "Imperial" is not the system currently used in the UK. E.g. in the Imperial system gold was measured in troy pounds, not in avoirdupois pounds. If we write "Imperial", sooner or later someone will come along, point to this guideline, and insist on something like "The robbers stole 267 kg (716 lb troy)" where we should really convert this as 588 lb avdp. I object to all 4 changes of meaning. --Hans Adler (talk) 11:23, 15 May 2009 (UTC)

In a nutshell, what Hans wrote. This is getting tendentious.

WP Physics

} 12:04, 15 May 2009 (UTC)

Michael, yes is was easier simply to revert your entire change. However, three of your changes made no improvement to the current wording, so nothing has been lost there. Obviously the main objection is to change 3 (bullet 2). I see very little wrong with the current wording. It needs to be ambiguous to take account of instances when none of metric, imperial or US customary measures are appropriate, as Hans says in historical contexts, but also for subjects such as oil (generally measured in barrels). Also, as I have said before, conversions are dealt with in another section, so do not need to be covered here. _wjemather^bigissue 12:21, 15 May 2009 (UTC)

Hans, it seems you have a ready argument for rejecting anything I suggest. If my edit changes the meaning, you reject it; if it doesn't, you dismiss it as a 'minor stylistic tweak' and then reject it anyway, because it changes the meaning. Let us leave aside the changes that you have argued against. I assert that what you call minor stylistic tweaks are an improvement, because they make the language is simpler, clearer and shorter.

Wjemather, it seems you admit to reverting at least one of my changed wordings, not because you disagreed with it, but because it was easier. Thanks a lot.

I will comment on the argument about ancient measures and looking up meanings later. Michael Glass (talk) 14:42, 15 May 2009 (UTC)

Hans's criticisms Only one criticism made sense to me.It could be read as saying that all three measurements should be applied. As this is a possiblity I'll withdraw that proposal. However, the points he makes are a very good argument for using SI and related measures, at least as supplementary units, everywhere.

I cannot see how the proposed wording would affect the use of other measures, whether ancient or modern, whether they are troy pounds or barrels of oil or whatever. However, I reject the excuses that personal animosity is a good reason to refuse to cooperate or that Hans finds it wearisome to argue his point of view. Personal animosity is not a logical reason for rejecting proposed wordings.

Wjmather's points Wjmather has admitted that he took the easy way out in reverting my edit in full. Therefore when he writes 'Your changes made no improvement to the current wording,' the statement is self-serving. In at least one case the charge of 'no improvement' is demonstrably false.

A: Wikipedia is international; apart from region-specific topics, use international units.

B: International scope: Wikipedia is not country-specific; apart from region-specific topics, use international units.

A is obviously shorter, clearer and easier to follow than B. Wjmather's argument simply falls away. He rejected this wording, not because it was worse, but because a total revert is easier than a partial revert.

I believe that this proposed revision is

• neutral
• shorter
• clearer

Why should we reject it simply because Wjmather took the easy way out? Michael Glass (talk) 01:01, 16 May 2009 (UTC)

And the previous version is also neutral, concise/short and clear. The removal of three words is utterly trivial and pointless, and breaks the style of the three bullets.

WP Physics

} 01:40, 16 May 2009 (UTC)

No more trivial and pointless than your last edit [9]. You are prepared to make what you call minor tweaks to the wording to improve the style; so am I. If you are concerned about breaking style, consider how you could improve bullet 2, which also breaks the style of the other two. Michael Glass (talk) 02:12, 16 May 2009 (UTC)

WP Physics

} 05:36, 16 May 2009 (UTC)

In fact, I tend to agree with that edit though I feel that "'plus one' and 'minus one'" could be considered. As for quoting the policy on not making personal attacks, I suggest you look at your own behaviour first. For one who is rude and doesn't even care, this wins the prize for gall. Now, about your latest edit. I reverted it because it implies that problems over wording are one-sided. Problems with wording are more often disputes between editors. The policy shouldn't imply that these are one-sided problems. Michael Glass (talk) 08:30, 16 May 2009 (UTC)

Nowhere did I attack you. What I said is I can't be helped to give a damn that you take offense at me calling your version of things "BIPM-pushing POV" because you thought this made you sound like a drug-pusher.

WP Physics

} 15:05, 16 May 2009 (UTC)

That's not the way it came across to me, but thanks for your explanation. Michael Glass (talk) 23:33, 16 May 2009 (UTC)

Bureau International des Poids et Mesures International Bureau of Weights and Measures!

"UK-related topics may have either SI or imperial units as the primary units." Does not work in my opinion because it my use both as the primary units it depends on the things being discussed (the shot of whiskey and the half pint chaser, cc and horse power, litres of petrol and walking five hundred miles).

Surly it is better just to recommend using that which is used in modern reliable sources on the subject with some sort of suggestion of what to do if the sources are divided on usage, as it is supported by the Wikipedia Verifiability policy, and it will take care of things like whether the altitude of a flying aircraft should be in feet or metres, without having to have complicated rule set to arbitrarily impose standards on articles. It results in the units used being the least surprising to the readers, and it would certainly makes writing articles about UK weights a measures much simpler to follow.

Forcing one system on areas of specialization reminds me of the mess, we get into, if we try to impose a dating system on British history articles, which does not follow the conventions used in reliable sources -- should the execution of Charles I be done using the contemporary Julian calendar, the revised calendar, or the Gregorian calendar-- Keep It Simple Stupid, keep to the

talk

) 11:17, 21 May 2009 (UTC)

100% agree. --A. di M. (formerly Army1987) — Deeds, not words. 19:23, 21 May 2009 (UTC)

Why are there two "Unnecessary vagueness" sections?

I don't know if anyone else has noticed, but there are currently two sections of the manual page entitled "Unnecessary vagueness", one slightly more complexly formatted than the other. There are all kinds of good reasons to avoid identical names in the same article (which is why I hate multiple polls with just "Support", "Oppose" and "Neutral": when I see the edit summary I ask Support what?) Can we at least have consensus about which to delete? —— Shakescene (talk) 20:53, 21 May 2009 (UTC)

(Apparently not, according to the recent edit history.) —— Shakescene (talk) 05:17, 23 May 2009 (UTC)

I deleted the second of them by mistake (I had noticed Michael Glass had re-added the first one after it was removed, but I hadn't noticed he had reverted himself); I don't really care about which one is retained. (If I had to choose, I'd keep the second one, because that's a general principle and it's weird to mention it in a section about fine details.) --A. di M. (formerly Army1987) — Deeds, not words. 09:59, 23 May 2009 (UTC)

User:SimonTrew's Understanding of the discussion so far

Hello everyone,

I've been following this discussion for some time but not put anything into it. I don't want here to do anything but set down some deliberately vague thoughts that I have taken from this discussion.

As a UK editor and based in the UK but having lived elsewhere I thought I would just add a general rambling discussion, partly as a little comic relief but also maybe to steer us back to the point, which is that we invent units of measure to suit the things we are measuring.

I've been designing units of measure subsystems in computer systems for many years and have heard many of these arguments before, and shan't repeat them. It seems to me Michael Glass has a reasonable argument with the inconsistency between different articles of similar standing (e.g. counties or islands) in the UK. However, that is a general problem in WP: MOS only really says articles should be consistent, not that topics etc should. That is undesirable, but is not specific to numbers: spelling suffers the same problem, and for the same reasons (different authors, subsequent editors, sources, and so forth). Personally I think there should be a broader recommendation that groups of related articles (topics or projects) should have consistency but then that totally hoists the need to discuss it out of MOSNUM (which would be a good thing).

It is incontrovertibly the case that in the UK road distances must be given in miles-- not just on signs but also in written law etc. The same also applies to clearance heights. A swedish lorry driver was cleared of careless driving a couple of years ago because the bridge he hit was not signed in Imperial measure. The fact his cab would have had its height in metres etc was irrelevant. (CN I know.)

Ordnance Survey maps are metric, by the way.

I am thirty-seven. I have no idea how tall I am except in feet and inches. I use SI every day of my working life, and Imperial in my life outside work. I like the dual system, it conveys appropriate imprecision. To say something is 200 yards away gives a rough idea, to say it is 200 metres away sounds very exact.

As always, it seems to me to think of the intended audience (who is likely to be reading this article?) and go from there. SI, even with appropriate rounding, can imply a level of precision that is unwarranted. That is why in all cultures at all times people invent units of measure suitable for the task, be it hands, carats, parsecs, calories (kilocalories), or whatever. The french livre and the spanish libre are different (400g and 500g if I remember it the right way round) but both just mean "a pound" because that's a handy size for measuring butter and fats, some other foodstuffs, stuff like that.

The paper sizes thing is interesting actually, since few who don't really draw would know the measurements, let alone say that an A0 sheeet of paper has an area of a square metre, and that the other sizes are half of that, and the aspect ratio is the square root of 2.

I'm a little surprised at Hans because he seemed to have the most sensible grasp of the UK system of most of the regular contributors to this argument then he says he finds himself converting it (not quoting you verbatim Hans)! That's kinda the point, really. If you give a British person a distance in kilometres mentally they will convert it into miles. I think, at least for people say 30 or under, they are less likely to do that for smaller measures (a metre or below), but still a lot of work is done in inches and feet etc. This is not purely out of spite. If you have a room built to imperial measure, a metric width wallpaper WILL NOT FIT the same as an imperial width wallpaper. I just viewed a house today and the old tiles are imperial and the new ones metric. THEY DO NOT LINE UP. 25mm IS NOT AN INCH. So, some things are kept in Imperial simply because they have to be. Screw threads are another example, as is plumbling: if you have Imperial plumbing and try to fit metric plumbing to it, beware, a compression joint WILL fit (just about), but seal it well or you will have it all leaking out of the joint.

As a small point 100l/km is always used by the manufacturers in the small print (and I think legally must be) but in practice the given figure is always in MPG. One difficulty here with converting (in one's head) is that it is a reciprocal conversion, one can't simply multiply by some rule of thumb (e.g. C to F one can multiply by 2 and add 30, that's about right. It's not spot on, I know the exact conversion, but it does for day-to-day conversation and is very easy to do in one's head).

Residential property guides in the UK virtually always give room sizes in feet × feet and the footprint of the property in (fractions of) acres. I am not sure nowadays about commercial property, that may be slowly moving over to metric, it does seem to be.

Although I use SI all the time at work, from the subatomic level (where again they give way to other units such as Van de Vaals derived units) to astromonical size units, and they're very handy for that kind of thing, my horses still always lose by a short head, my pencil is 2B not measured in Rohms hardness, and so on.

Carpets give an interesting example in that prices are generally given in square yards and metres both. There was a concerted effort in the mid 70s to move it over to metric. However, one large trader realised that a square yard was perceived as pretty much the same as a square metre but the price would seem a lot cheaper-- part of the metrication process was a slogan was "a metre is about three foot three, not much bigger than a yard, you see?" and so of course then all the other traders also followed suit and everyone was soon back on square yards.

I remember when they switched from 1/6 gill to 25ml measures for shorts a lot of pub chains had ads for "new larger measures" even though the difference is only about 5%, and 1/4 gill to 35ml measure is actually a shorter measure (1/4 gill is 35.5ml). It is that kind of thing that makes people very wary of switching to metric. The same happens, of course, with currency changeover (e.g. for the Euro)-- I can't seem to find a good article discussing the worry about that I think the German wikipedia has a good one "Teuro"? (Fat euro or something).

Draught beer and cider MUST be served by a third of a pint, half a pint, or multiples thereof. Milk (and I thought Orange Juice but have never been able to find this) may be served in pints to the doorstep in returnable bottles-- though really to me that is rather irrelevant since one could just as easily deliver 568ml.

Just some rambling thoughts to show that in the UK the use of Imperial measure is not dead, it is not restricted to "those over 75" (I am half that age), it is not anti-European or anything, it is people just using units that make sense in their daily lives. SimonTrew (talk) 18:32, 29 May 2009 (UTC)

Simon, you have given a marvellous picture of the use of weights and measures in Britain. It explains why similar articles are inconsistent: different editors have different preferences, and so Dartmoor is metric first in some places and Imperial first in others. Elsewhere in Wikipedia it's supposed to be US customary measures first in US articles, and metric first elsewhere. But not in Great Britain. Articles like the one about Big Ben appear to be thoroughly metric while that on the Tower Bridge is Imperial first. You would think that this inconsistency would be indefensible, but its defenders will support it to the last. Michael Glass (talk) 08:54, 30 May 2009 (UTC)

Dunno about "different editors", but a sentence starting with "Jones' engineer, Sir John Wolfe Barry, devised the idea of" should use the units Barry used, whichever they are. Sentences which are not that specific, e.g. "The highest point is High Willhays", should use the units which are in most common use in UK nowadays; according to the Times style guide cited above, "Use feet when dealing with aircraft height but metres when dealing with the heights of hills and mountains", so that sentence should use metres. What about throwing that "primary" units nonsense and say: Articles about UK-related topics should generally use metric units first, except for measurements which are still commonly made in imperial units in the UK. These include road distances, road speeds, personal heights and weights, heights of aircraft (but not heights of hills and mountains), volumes of beer, cider, and sometimes milk (but not other liquids), and miles per gallon for fuel consumption (but not gallons for fuel volumes). As I said earlier, there's no point in insisting that the diameter and length of an optical fibre between two UK towns be given either in micrometres and kilometres or in

thous and miles, when everybody would use micrometres and miles. "A foolish consistency is the hobgoblin of little minds". --A. di M. (formerly Army1987) — Deeds, not words

. 11:03, 30 May 2009 (UTC)

There's no inherent reason why the fact that you measure the distance between towns in miles should mean that you should also have to measure the average temperature in Fahrenheit or that you should have to measure its elevation in feet. Better to apply the system actually in use in the UK. Pfainuk talk 10:44, 30 May 2009 (UTC)

Exactly. And in practice since people don't calculate in their heads any more it doesn't make a difference whether you subdivide a mile into 1,760 yards or 1,609 metres. So it's only natural people in the UK are moving towards doing the latter in spite of their affection to their traditional system. --Hans Adler (talk) 10:57, 30 May 2009 (UTC)

(ec)Michael, I am not sure you are getting it. These two articles are using feet or metres first, respectively, for building heights and building lengths. This is one of the few areas where metric and pre-metric are about equally in use. (Mostly because "a metre is a yard".) Another is square miles vs. square kilometres, and hectares vs. acres. You are not going to find a similar example with miles vs. kilometres for road distances; not even on the

Channel islands articles as I just checked, even though due to their separate history and close proximity to France they have very special issues as witnessed by Lihou, which uses vergées

first, then acres, then hectares.

Inter-article consistency is desirable and currently a problem. But this is a wider problem than just units. E.g. compare the dual mathematical concepts of "equaliser" and "coequalizer", and look at what happened when somebody tried to clean this up: Talk:Equaliser (mathematics)#Requested move. --Hans Adler (talk) 10:57, 30 May 2009 (UTC)

I can only come back and give my own perspective. When I am on the continent, or in Canada, and have to use kilometres, I kinda mentally translate that into miles, even if I am driving a rental car that only has an odometer measuring kilometres. (BTW Hans I believe there is a legal requirement for cars that are stabled here to measure speed in miles per hour, you should get a new odometer fitted, or perhaps they have dropped that now.) I agree with Hans (and said so) that yes the inter-article consistency is a problem and that it is wider and so should be hoist out of this discussion (I don't know quite to where). I imagine it has been discussed elsewhere many a time, I presume there are some guidelines for projects or something? I disagree "people don't calculate in their heads any more", I do all the time, I am always comparing prices etc by doing some mental arithmetic (you sometimes find for example that buying two smaller bottles of something is cheaper than one larger bottle, kinda a reverse psychology on the part of the retailer) and sometimes that involves metric/imperial conversions too.

I am not sure how Hans divines the mind of a British person and works out that he translates "1 mile" into "1609 metres", I would doubt one in ten British people even know how many yards there are in a mile let alone metres. It's just "a mile" or "a quarter of a mile" or whatever. That's kinda the point. The measure, or the unit of measure (not really sure which here) is appropriate to the concept being measured one doesn't think in terms of any kind of number really just oh it's about THAT far. I don't think, then, it really varies whether one's mental model also includes metric (SI) or imperial or US customary.

I am guilty of this myself when I said "british people mentally convert kilometres into miles". How should I know?

With housing it is especially important to keep Imperial in daily life simply because something like 90% of housing stock is built to Imperial measure. Most plumbing nowadays is metric but if you have Imperial measure plumbing it simply WILL NOT FIT. Same with carpets, wallpaper, etc etc etc. Last time you bought 6 rolls but now you need 7 because not only is it marginally narrower but also the drop on the wallpaper has changed because the pattern is also on a metric-based repeat. See, these things aren't easy.

You may wish to have a look at my work at Coca-Cola formula where I have tried to back-translate ridiculous metric measure to US Customary. I am not sure I have it all right, but it is patently obvious the recipe was originally given in US Customary and some idiot converted it to metric (not necessarily the Wikipedia editor it may have been in source) ending in stupid amounts because of overprecision and all kinds of other faults. I am sure that that article needs improvement, but I hope it may serve as an example when you see the struggle there when someone carelessly translates into metric. That is nothing against metric; it's against its careless translation. SimonTrew (talk) 19:53, 30 May 2009 (UTC)

I think you missed my point that while there was once a time when people knew the algorithm for the complicated conversion between miles and yards, they now use a pocket calculator or computer at those rare occasions when they have to convert. Hence for all practical purposes, a system with metres for short lengths and miles for longer distances is no worse than the original one with yards and miles. In fact it's superior because the metre is an international unit. But I neglected the constraints that you describe, which come from Imperially measured goods. Of course there are other conversions that we have to do more often. --Hans Adler (talk) 21:35, 30 May 2009 (UTC)

For the pre-metric quantities (including drams) in the purported Coca-Cola formulas, look (for example) at this earlier version of the article (pulled from its edit history): http://en.wikipedia.org/w/index.php?title=Coca-Cola_formula&oldid=9415661 (It took me a year to understand that you can pull the actual version in one motion by just clicking its date in the edit history, in this case 08:38 14 January 2005.) Note that the third, 1968, formula, whose original text is quoted here seems to have been stated originally largely in metric quantities (as a chemist, apothecary or pharmacist in 1968 might very well do.) —— Shakescene (talk) 20:22, 30 May 2009 (UTC)

Interesting discussion. I would like to respond to some points.

• One point appears to be that I have identified the problem but that my solution is impractical. If I propose a rule to force metrication, this would not work, and if it's only a recommendation, that wouldn't work either, because it would be ineffective. Actually, what I proposed was to allow both systems, but state a preference for SI, and then leave time and patience to sort it out. This might sound ineffective, but it apparently worked with temperatures, where Celsius gradually supplanted Fahrenheit in British usage.
• A second objection is that the British are wedded to the mile, so Wikipedia has to follow this preference in UK-based articles. I admit that this fact alone would stump any chance of applying consistency in measures on many articles.
• A third objection is that it we should fix on current British usage, and have Imperial roads winding over metric mountains.

Actually, this last provision is actually an oversimplification of what really happens in British articles. Some, like

By the way, there are 1760 yards, 5280 feet and 63,360 inches in a mile. Michael Glass (talk) 08:42, 31 May 2009 (UTC)

I wouldn't object to changing the first so that it gives distances in miles and the second so that it gives heights in metres. As for consistency, I don't think there is anything intrinsically wrong in giving road distances in miles and heights in metres: from a theoretical POV, it isn't much worse than measuring lenghts in metres and durations in seconds. (In Italy, the same identical length would be given as 2260 metres if vertical and as 2.26 kilometres if horizontal. Nobody ever complains that this is inconsistent, probably because you don't usually have to directly compare horizontal and vertical lengths of that size. Now maybe a factor of conversion of 1000 is much nicer than one of 1609, but so what?) --A. di M. (formerly Army1987) — Deeds, not words. 13:07, 31 May 2009 (UTC)

I think the most common conversion is probably 1600 metres to the mile. The extra nine metres are generally insignificant given that miles are not generally used for precise measurement. I would also note that the fact that, say, 6,600 feet (2,012 m) and 1.25 miles (2.012 km) are actually the same length is far less obvious in the non-decimal imperial measures than in metric. So it's understandable that people used to a non-decimal system of measurement might not see the need, and indeed, find it strange, to insist on rigorous adherence to a single system.

It should be mentioned that many imperial measures are used only in specific circumstances. The fathom of six feet is traditionally only used for water depth. The hand of four inches is only used to measure horses. The furlong of 220 feet is only used to measure horse racing tracks. That several different units might be used for the same dimension, depending on context, is quite normal - and these units are generally not mixed. That metric units get added to the mix as well is not particularly surprising.

I would propose that we go ahead with the sort of change we've been discussing. I would support your wording above. I'd suggest it be firmed up slightly (to avoid the inevitable argument as to what exactly constitutes common usage) - though that said, this is supposed to be a guideline, and I suppose some ambiguity isn't necessarily a bad idea. Pfainuk talk 21:38, 1 June 2009 (UTC)

Both the

So, you're now arguing that we should not use have consistency between articles, a position that you described on Saturday as "indefensible"? Curious. Pfainuk talk 17:13, 3 June 2009 (UTC)

Perhaps I should rephrase this: if a distance is measured primarily in miles (inches, feet, yards) why, in your opinion, is it necessary that mass, when measured elsewhere in the article, has to be measured primarily in pounds (ounces, stones, long tons)? And why, in your opinion is it necessary that - because distance is primarily in miles - temperature has to be measured primarily in Fahrenheit? Particularly bearing in mind that this defies local custom (and so inter-unit conversion is not a strong factor).

Second question, how does insisting on this make our articles better, when it means that (for example) the height of Mount Usborne is measured primarily in metres on Mount Usborne and on East Falkland, but in feet on Falkland Islands?

The fact is, the same measurement is frequently carried out in different sets of units on closely related articles. You appear to be saying we have to have this because otherwise different measurements - measurements of different and unrelated quantities - might be given in different sets of units on a single article. Yes, I am arguing for a 10-mile (16 km) road across a 1,200-metre (3,900 ft) mountain. But you appear to be arguing that that same mountain must be measured as 1,200 metres (3,900 ft) on one article and 3,900 feet (1,200 m) on another closely related article, depending on the whims of the editors. I think we should be measuring the same thing the same way on closely related articles, and I think that applying the Times style guide - or something similar - is the most consistent practical way of doing this. Pfainuk talk 18:15, 3 June 2009 (UTC)

In response to Pfainuk, I have the following to state:

• My comments above were directed at messing about with
  Shetland Islands
  to impose a forced inconsistency on them.
• Voluntary inconsistency may be problematical; enforced inconsistency is far worse.
• In my opinion, the best option is to allow both Imperial and metric measures to come first in articles but to state a preference for metric measures. This could help to sort out inconsistencies between the measures that Pfainuk noted in articles about the Falkland Islands.
• I repeat that though this is my opinion I would not put this onto MOSUM at this time. Thank you, however, for making it necessary for me to restate my position on this matter. Michael Glass (talk) 06:36, 4 June 2009 (UTC)

You are not stating clearly what you mean by (voluntary/involuntary) "inconsistency", which prompts me to say the following: Just because it is possibly transitional and apparently so young that nobody has named it yet, is no reason to discriminate against the system of weights and measurements currently in use in the UK. It is a hybrid of the metric and Imperial systems, with inches, feet, metres, miles for lengths/heights, acres, hectares, square kilometres and square miles for area, Celsius for temperatures, pounds avoirdupois and kilogrammes for weights, and litres and pints for volumes. We should use it for the UK, although whenever we have to choose between a metric and an Imperial unit of similar size and about equal popularity we should prefer the metric unit because it will be understood more widely and in some cases we avoid misunderstandings by American readers who may not know that the UK pint is larger than 1/2 litre while theirs is smaller.

I am not sure that Hampstead Heath is internally consistent:

• 790 acres (320 ha)
• 440 feet or 134 metres high
• 790 acres (320 ha)
• 321 feet (98 m) high
• 0.5 km² (50 ha; 120 acres)

The choice of the square kilometre here may be due to the round number.

Shetland is internally consistent in using metric first, but that's an odd choice:

• 280 km (170 mi)
• 1,466 km² (566 sq mi)
• 967 km² (374 sq mi)
• [...] gave future kings of Norway the right to redeem the islands for a fixed sum of 210 kilograms (460 lb) of gold or 2,310 kilograms (5,100 lb) of silver
• 64 kilometres
• 5 °C (41 °F) etc.
• 1 mm (0.039 in) of rain etc.
• 40 km (25 miles)
• 11 km

It's amazing to me that some of the metric lengths are not even converted; my original research using Google has shown that miles seem to be predominant on these islands. The strange use of kilogrammes for a historical treaty comes from the original (Norwegian) source; the number for gold makes sense insofar as it is equal to 350 historical Danish pounds. Wherever possible such historical data should use the original units because there is normally some uncertainty in the conversion; I can't fix this problem because I can't find a better source. --Hans Adler (talk) 11:34, 4 June 2009 (UTC)

What I meant by involuntary inconsistency is when the use of units in the UK in 2009 is frozen in aspic and foisted on all, whether they want it or not. I think it is better to leave it to the good sense of editors to make their own arrangements in this transitional period. However, I would add that it is helpful to suggest the use of the metric system. I hope that this makes my meaning clearer. Michael Glass (talk) 13:05, 4 June 2009 (UTC)

It wouldn't be helpful to recommend the metric system. As I say, either the MOS would be considered a reason to change to all-metric - in which case it would mean metric units would be primary everywhere on UK-related articles - or it would be optional and would not help sort out the inconsistencies in Falklands articles and others.

You talk about "leaving it to the good sense of editors", but that's not what we do at the moment - we require that all units for a given article be given in either the imperial system or the metric system. Which, in your interpretation, means that editors on an article like London have to choose which measurements they want to give in units that are barely used by the locals and which will only be used as conversions by sources.

If they choose to use miles for distance (like all their sources), then the current guidelines tell them they have to measure rainfall in inches and average temperature in Fahrenheit. If they choose to use the commonly-used units for climate (again, like all their sources), they have to convert all the distances into kilometres and then convert them back into miles using the template.

Personally, I see the as a solution in need of a problem - in my view the benefits of using the system of units currently applied in the UK (generally speaking - and the whole point of the MOS is that it allows for common sense exceptions) far outweigh the benefits of imposing either the imperial system or the metric system on a given UK-related article.

So, again (since I don't believe you answered this before), why should measuring distances in miles mean that temperature has to be measured in Fahrenheit as the MOS currently proscribes? On what basis should such an insistence be made? A scientific basis? A cultural basis?

And how, in your view, do the benefits of imposing a single system on a given article outweigh the flaws of such a system - such as the fact that the units preferred in articles are not likely to be the same as those preferred by sources, editors and readers, such as the fact that this system will inevitably mean that exactly the same measurement is given in different systems on closely-related articles (something that will not be changed by your proposal), such as the fact that - in practice - such a system creates more work in converting articles to meet the MOS, when they have originally been written using units preferred by British editors? Pfainuk talk 17:55, 4 June 2009 (UTC)

First of all, I do not believe in imposing on editors either the Imperial or the metric system for UK articles, but rather leave it to the good sense of the editors to choose the units to put first.

There are three things that I would like to recommended about UK articles.

• The first is to accept the fact that there will be inconsistency in usage with UK articles, simply because British usage is inconsistent and also because different editors have different preferences.
• The second is to use both metric and Imperial measures in British articles as a general rule. This is to provide for the maximum number of readers and help to smooth out any difficulties that might arise from using one system or another.
• The third is to recommend that while editors may have either metric or Imperial measures in first place, metric measures should be suggested.

All three together would provide the guidance that is needed with the flexibility that is required. Michael Glass (talk) 14:18, 7 June 2009 (UTC)

The current system does impose metric or imperial on British articles in that it insists that a single article use one system as primary in all cases. In this way, it achieves the worst of both worlds. It leads to the inter-article inconsistencies that we both agree are a problem, while at the same time imposes one or other system on each article. In practice, this system is routinely ignored - but if the MOS is going to be routinely ignored, why have an MOS?

How about this. We replace the wording we have now with:

On UK-related topics, primary units may be either imperial or metric, depending on the context in which the units are used (which may be determined based on sources). Where there is doubt as to which unit is most appropriate in a given context, the metric unit is to be preferred. It is strongly recommended that similar measurements should use the same primary units, but a given article may use primary units from different systems in different contexts. Conversions between units are considered particularly important on UK-related articles.

This no longer gives a list of units to be in imperial, and contains your recommendation that metric be preferred in case of doubt, while allowing imperial units to be used where appropriate. It no longer requires that articles take their primary units from a single system. Indeed, pretty much the only requirements it makes are that similar measurements should use similar units (so that we're not talking about going 5 miles (8.0 km) down a 10-kilometre (6.2 mi) road or something), and that conversions should generally given.

I would imagine that, faced with the same issue of which units are most appropriate, editors dealing with closely-related articles are likely to make the same decisions as to which units should be primary, at least partly resolving the inter-article consistency issue. Pfainuk talk 15:37, 7 June 2009 (UTC)

I think there is some merit in your suggested wording. I'll think this over. Michael Glass (talk) 03:40, 8 June 2009 (UTC)

First of all I am impressed by the work you have put into the suggestion above. It has real merit, but I think that the use of units in an article should be largely determined by the use of units in the majority of the sources quoted. Here is a suggested revision of what you proposed:

In UK-related topics, the primary units may be metric or imperial, but it is important to provide conversions to the other units. The choice of units should largely depend on the units used in the sources referred to. If in doubt, put the metric unit first. Similar measurements should use the same primary units (this is strongly recommended), but a given article may use primary units from different systems in different contexts.

I have tried to preserve all the good features of your suggested wording, but to put them in a more concise form (72 words instead of 87) but it may still be too long to fit comfortably into the guideline. I look forward to your feedback on this suggestion. Michael Glass (talk) 14:05, 9 June 2009 (UTC)

Excellent. I agree with this version. --Hans Adler (talk) 14:36, 9 June 2009 (UTC)

I would support your version.

Looking at the preview screen, we do say that we should make similar measurements using the same primary units in the bullet point below the one we're discussing, so I have no objection if you choose to take that as read and delete the first part of that sentence (as far as the "but"). I think the rest of the sentence is an important clarification, so I think we need it. On my screen this would mean that the comment would be two lines long, which is not exactly short, but is manageable. I support it in either case.

My thoughts on the sources question - I would expect that most sources would use the most appropriate units in the context concerned anyway, so I think the question as to whether it is non-source-based context or source-based context we're dealing with is generally going to be moot. And sourced-based context does make it easier for non-Brits to understand the MOS properly. Thanks, Pfainuk talk 20:49, 9 June 2009 (UTC)

Other than clarifying that the current wording does not mean use all metric or all imperial, the proposed changes add nothing but unnecessary duplication of what already exists in the guidelines. Simply replacing "either... or" with "both... and" would achieve the same.

UK-related topics may have both SI and imperial units as the primary units.

The rest of the bullets in the guidelines deal with the remainder, especially with regards sources and using SI first – I quote: "put the source value first and the converted value second. If the choice of units is arbitrary, use SI units as the main unit". This already seems quite clear to me. _wjemather^bigissue 23:22, 9 June 2009 (UTC)

I noticed that - but would suggest that (since it's a change from the previous method), we ought to note at least that a given article may use primary units from different systems in different contexts. Pfainuk talk 17:26, 11 June 2009 (UTC)

I feel the best clarification would be:

UK-related topics may have both SI and imperial units as the primary units, largely depending on the units used in the sources of the information.

This would help to encourage the use of units based on the sources rather than the whims of individual editors. Michael Glass (talk) 00:41, 12 June 2009 (UTC)

You seem to have an ability for finding sources that quote in metric to satisfy your whims, and using them as justification for changing articles. Just because one out of a hundred sources use metric, does not mean that is what should be used. _wjemather^bigissue 01:03, 12 June 2009 (UTC)

To clarify, and this time fully quote the existing guidelines: "If editors cannot agree on the sequence of units, put the source value first and the converted value second. If the choice of units is arbitrary, use SI units as the main unit, with converted units in parentheses." Sources should not determine the order, unless the order cannot be agreed. A source may be cited with the order reversed in the text if appropriate. It is entirely possible to find sources using whichever units you may desire, and this should not be exploited to push an agenda of imposing metric (or anything else) on all articles. _wjemather^bigissue 09:05, 12 June 2009 (UTC)

What is your problem with finding or verifying sources for information?Michael Glass (talk) 01:36, 13 June 2009 (UTC)

I don't know where you get that idea from. Clearly, the only issue I do have is that you are still continuing this needless discussion in the hope that one day everyone will give up and let you metricate everything. _wjemather^bigissue 08:12, 13 June 2009 (UTC)

Do you mean you don't have a problem with finding sources of information and verifying information? Good. End of discussion. Michael Glass (talk) 11:43, 13 June 2009 (UTC)

It's a tad irritating that we seem to be agreed on a broad principle, but cannot change because of this one detail.

My view is that, on a general principle, we should not be writing rules based on what one editor might do or not do. In many cases in the United Kingdom it is pretty obvious - to a Brit - which unit is most appropriate. But I can quite see that to people who may not have been to the UK, this is less obvious.

There are clear benefits to relying on sourcing. For one thing, it means that we avoid original research. For another, it allows non-Brits to be clear on units on the same basis as Brits. That said, I see no reason not to allow editors sufficient flexibility to base unit conversions on whichever unit is considered most appropriate by consensus, while meeting policies such as

WP:NOR

.

How about, UK-related topics may have both SI and imperial units as the primary units, and a given article may use primary units from different systems in different contexts, based on the principles outlined for all articles below.

As Wjemather notes, the principles include using sources where there is disagreement. Pfainuk talk 17:12, 17 June 2009 (UTC)

Channel Islands: Terminology

Please note, by the way, when giving examples of inconsistency in

While of course you are correct, the average user coming for guidance to the Manual of Style won't take the time to determine what sense of UK is involved (e.g. while the Channel Islands and Isle of Man are considered distinctly by such bodies as the European Union and the European Court, for most practical purposes—outside banking, finance, taxation and customs—their international representation is understood by everyone as included in the UK's.) And much of what holds true about customary usage and familiarly-understood units for Great Britain will also hold true for both Northern Ireland and the Republic of Ireland (which joined the then-European Communities at the same time as the UK), so the reader needs to understand when we're talking about general geography and when international and intra-national boundaries do coincide with differences in usage. —— Shakescene (talk) 21:18, 30 May 2009 (UTC)

Quite agree. I make the point simply that if we are going to get picky about "UK articles", especially about articles *about* the UK not those written by UK editors, we better make sure what we are talking about. It's no big deal (I am English not British) but let's make sure we don't get the argument cluttered with nonsense like this, simply by making sure we are on the same terms. SimonTrew (talk) 00:16, 31 May 2009 (UTC)

Not all international bodies recognise them as being included in the UK's. For example, they have different number plate registration and different international plates (GBM, GBJ, GBG etc). It has been a long-standing grumble in Northern Ireland that although they have their own driver licensing and plate issuing agency their international plate is GB, even though they are not even part of Great Britain. SimonTrew (talk) 00:23, 31 May 2009 (UTC)

It must be even more galling for Northern Irish (of all beliefs and loyalties) when they form part of a UK athletic team whose international abbreviation is GB or GBR. While the athletes from the dissolving USSR were still competing as a single team, at least its abbreviation was changed from SU (Soviet Union) to EUN (Équipe Unifiée). —— Shakescene (talk) 00:34, 31 May 2009 (UTC)

What matters is which units of measurement are used in Jersey etc., not any other geo-political stuff. --A. di M. (formerly Army1987) — Deeds, not words. 13:14, 31 May 2009 (UTC)

a d m i disagree. what matters is what people elswhere can understand. so put in miles and metres both. the oddity with the channel islands pointed out up there is jersey is one way and guernsey another, it's the inconsistency BETWEEN articles which doesn't break any wp style rules but is anomalous (or between derbyshire and lincolnshire or whatever). I would like to campaign for a guideline that related articles i mean things you could group into a project or whatever they need their own style guidelins. That's kinda deiberatley vague but you need to say hey channel islands articles use imperial first (or metric first, whatever you decide) so that they are consistent between themselcves. The inconsistency between similar articles it certainly gets my goat so i imagine it does the average wp reader too. That's what we';re alll trying to improve isn't it? SimonTrew (talk) 10:38, 6 June 2009 (UTC)

I meant "what matters in deciding whether to write 32 kilometres (20 mi) or 20 miles (32 km)". If kilometres are normally used in these islands, use the former. If miles are, use the latter. I've never advocated using any unit of measurement without a conversion (except for things which are measured with the same unit all over the world, such as ages of people in years, electric tensions in volts, etc.). --A. di M. (formerly Army1987) — Deeds, not words. 20:50, 11 June 2009 (UTC)

Aircraft measures

Aircraft height is always in feet, but usually just given as a level e.g. climb to level 250 is 25,000 feet. With modern altimeters, autopilots and so forth, a commercial airliner will be almost bang on to 25,000 feet with an astonishing degree of accuracy (as, say, reported by GPS systems). Of course "height above sea level" depends on the definition of sea level but that is another story. Similarly, distances are in nautical miles which are defined in SI (whatever) as 1852m (the knot, nm per hour, being a derived unit). SimonTrew (talk) 00:12, 31 May 2009 (UTC)

By the way what I may not have may made clear is this is international, worldwide as far as I know. Of course internally the autopilit software uses SI (far easier) but levels are always given in feet. It's one of those things that just "sticks" i.e. once you've got it, it's very hard to change. Similarly wind speed is given in knots and stuff. Yet without doubt all new aircraft (and I think all new cars now even america ones) are built in metric. SimonTrew (talk) 03:47, 19 June 2009 (UTC)

Cars actually suffer from the same hybridisation in the UK. I just replaced a tyre which is defined as "185 R15", i.e. 185 mm wide on a 15 inch rim! --RexxS (talk) 13:15, 19 June 2009 (UTC)

Aircraft navigation suffers from the early domination of the aircraft industry by the Americans and British, with altitudes measured in feet and speeds measured in knots because Anglo-American nautical methods were applied to aircraft navigation. Things haven't changed because there are safety issues involved and human lives are at stake. (see

Gimli glider for the results of a metric conversion error in measuring fuel levels.) Communications are always in English for the same reason. If the French or Germans had dominated the early aircraft industry instead of the Americans and Brits, the results might have been different.RockyMtnGuy (talk

) 16:57, 23 June 2009 (UTC)

Technically, communications weren't always in English; the strict standardization was put in place after the Tenerife airport disaster. --Golbez (talk) 17:24, 23 June 2009 (UTC)

This request for comment is for several related questions. This RfC is about the appearance of the numbers, and whether they can be successfully cut-and-pasted into other applications. The templates used to format them is not the topic of this discussion; template coding can follow whatever conclusions the RfC comes to.

1. Is 4,046.8564224 an acceptable number format? (Notice the comma to the left of the decimal point and the gap to the right used to group digits into groups of three digits.)
2. Is 4046.8564224 an acceptable number format? (Notice that gaps are used on both sides of the decimal point to group digits.
3. If gaps are acceptable both to the right and left of the decimal point, and an article needs gaps to the right in some numbers in the article, should all the numbers in the article, including numbers with three or fewer digits right of the decimal, be formatted with gaps?

The Manual of Style (dates and numbers) (MOSNUM) for quite a while advocated comma separators to the left of the decimal, and either no separators, or gap separators, to the right of the decimal. However, none of the examples in the MOSNUM or in the documentation for the {{Val}} illustrated that this could occur in the same number, thus creating a style clash within the same number. About a week ago this was changed to allow 4046.8564224 or 4,046.8564224 but not 4,046.856 422 4. Today the change was reverted and the reverting editor suggested wider input be sought.

(In case you are curious where the number 4046.8564224 came from, it is the exact number of square meters in an international acre.) --Jc3s5h (talk) 02:08, 13 June 2009 (UTC)

• The short answer is that 4046.8564224 may be suitable for scientific articles and may (but don’t have to be) used there. For general-interest articles, use either 4046.8564224 or 4,046.8564224. The reason for using {{val}} is to prevent editors from leaving a single dangling digit, like you used above when you wrote 4046.8564224.

  It is not the job of Wikipedia to promote change in the way the world works by adopting good ideas by standard bodies just because they are good ideas and come from standard bodies. Wikipedia generally simply follows the way the world really works. Truly good ideas from standards bodies usually rapidly gain traction in the real world. When that happens, Wikipedia follows suit. In other cases, such as the IEC’s proposal that the world adopt the new binary prefixes like “256 mebibyte”, the proposal falls on its face. And again, notwithstanding the compelling virtues of the IEC’s proposal, Wikipedia ignores the proposal so as to not confuse our readership. Note also that the rule of SI is to have a space between the number and the percent symbol, such as “50 %” rather than the “50%” we are all accustomed to. Like the rest of the world, we ignore that one too.

  As regards the delimiting of large numbers and/or high-precision numbers in all our general-interest articles, Wikipedia specifies the U.S. method of delimiting numbers using commas in the significand because that causes the least confusion with its readership. U.S. readers are far and wide entirely unfamiliar with other delimiting conventions whereas their European counterparts are typically comfortable with several different methods and are not in the least confused by the U.S. convention. Scientific articles are a notable exception, since the language of science is scientific papers which follow the SI method so that such papers can be read by a world-wide audience. Even then, delimiting with narrow spaces is allowed in our scientific articles—not required. Greg L (talk) 02:19, 13 June 2009 (UTC)

Minor correction and comment: I thought the {{

• No that is not correct that {{Val}} does not profess to follow any international standard. It follows ISO convention, observed by the SI and the NIST to the right of the decimal marker. That progression is as follows:

  1.1
  1.12
  1.123
  1.1234
  1.12345
  1.123456
  1.1234567
  1.12345678
  1.123456789
  1.1234567890
  etc.

  I noticed also that you have had difficulty using {{val}} and have been getting error flags as shown here. Simply omit the commas and it will work fine. Your unfamiliarity with the template may underlie why you are uncomfortable with it.

  {{Val}} produces scientific notation values like this:

  Conductance quantum G₀ = 7.7480917004(53)×10⁻⁵ S. Compare that to the value here at the NIST. They are identical because {{val}} follows the convention used by science for scientific notation.

  {{Val}} is smart enough to handle the entire progression of delimiting, up to 14 significant digits, it won’t do a line-end word-wrap anywhere in the equivalency, it uses narrow spaces on both sides of the × symbol, and it even knows to use a true minus sign in the exponent (10⁻⁵) instead of the hyphen (10^-5) you would get if you simply type it off the keyboard.

  Editors are encouraged to use {{val}} because it delimits long, high-precision strings of digits to the right of the decimal marker and provides consistent, well-formatted output.

  Your objection seems to be over how it combines the practice of commas to the left, and narrow spaces to the right in numbers that aren’t scientific notation. Indeed, {{val}} follows MOSNUM guidelines for numbers containing four or more digits to the left and delimits them with commas. Your proposed solution (simply adopt the European practice of thin spaces both left and right) is contrary to MOSNUM and is very confusing to U.S. readers.

  The simple solution, is rather than try to change the way the world works, is for you to simply not use {{val}} to delimit on the right of the decimal marker for those relatively rare occasions where there are simultaneously four or more digits to the left. You are perfectly free to write 4,046.8564224 rather than the 4046.8564224 {{val}} generates for you. That should make you happy, and it would make me happy. I would encourage you, however, to avail yourself of the {{val}} template. Why? Because, since you are expressing five or more digits to the right of the decimal point, I assume you intend readers to actually read and understand those digits. Delimiting to the right helps to parse those digits just as delimiting with commas to the left helps to parse the digits on that side. Greg L (talk)

I agree that general-interest articles that do not contain any numbers with more than four digits to the right of any decimal should be the US format, which is widely understood. When there are more than four digits to the right of the decimal, we could just not use any digit grouping to the right of the decimal, but that is hard to read. We use commas to the left and gaps to the right; I really don't know if general-interest readers will find that less disconcerting than the BIPM style or not. The disadvantage is that rather than choosing between a style that has wide acceptance but is hard to read, or the BIPM style, which has limited acceptance, instead we go off and create our own style, which so far as I know, is not used by any other publication.

I think that readers are accustomed to seeing numbers displayed in many formats (phone numbers, social security numbers, and various gimmicks used in advertising) so even readers who have not been exposed to the BIPM format before will only be confused for a few moments. --Jc3s5h (talk) 03:11, 13 June 2009 (UTC)

In reply to Greg L's comments above, I see an example in the previously mentioned International System of Units {SI) brochure on page 126 that confirms Greg's belief that it is OK to eliminate the rightmost gap to avoid isolating the last digit after the decimal, even if there are more than four digits to the right of the decimal: 9.109 3826 (16) × 10⁻³¹ kg. Still, the vals template does not follow the BIPM/NIST/ISO convention for all cases because for some numbers, it violates the convention: "Neither dots nor commas are inserted in the spaces between groups of three." (ibid. p. 133)

However, I can't accept the idea of not using the val template when a number has both 5 or more digits to the left, and also 5 or more digits to the right of the decimal. It is very common for one template to invoke another template, so the editor who is actually typing the number may not be aware that the val template is being called. It would also be very confusing for an editor who is not especially familiar with the val template, and is just updating a number that was originally put in by someone else. It would be much better to alter the behavior of the val template so it produces acceptable results for any numbers, no special cases. (The problem of high precision numbers in scientific notation is understandable, since it is an issue for representing such numbers in most computer applications.)

As for using thin spaces to the left of the decimal being confusing to the general interest reader, I'm not convinced that is any more confusing than using two different digit grouping marks in the same article. Remember, the context here is an article where some of the numbers have five or more digits to the right of the decimal, so the reader will have to figure out what the thin spaces mean. Given that the reader figures this out, how hard is it to realize the same convention is being used to the left of the decimal? --Jc3s5h (talk) 04:43, 13 June 2009 (UTC)

• As a general comment regarding this RfC and other style issues, note that Wikipedia often uses bits and pieces of different standards, for different reasons. The reasons for choosing a style should go beyond simply finding sources that support a certain format, and should justify that style on the basis of practical merits. I saw it recently argued that Wikipedia ought to only use citation formats that are already accepted as standards by other bodies—but that argument didn't clearly articulate why this was preferable on Wikipedia, nor did it explain why Wikipedia contributors should not exercise their prerogative to create formatting methods that are more suitable in a principally online medium than (for example) publishing industry standards. On the other hand, many standards are the product of a valuable design process, and shouldn't simply be ignored because they weren't invented here, or aren't popular, or aren't pretty. We should take popular usage under advisement, but ultimately build the MOS based on what we can establish to be good, rather than what we feel is most popular. Familiarity is one major objective, but absolutely not the sole one, nor necessarily the most important one.

  It's not a matter of changing the world, but rather of serving the needs of the encylopedia most effectively. We ought to do what we can to improve Wikipedia, even if there are prevailing conditions in the wider world that make Wikipedia's style decisions inappropriate elsewhere.

  So on the topic of the RfC at hand (starting with item #2), the answers vary depending on the degree to which we're willing to be prescriptive, and conversely, the degree to which we're bound by local convention. I'd suggest that in an encyclopedia with an international audience, we'd be best served by using a number format that is understood in the least ambiguous terms. It just so happens that this is the exact problem that the BIPM convention is designed to address. It's a good system, not simply because respectable authorities recommend it for contexts where interlanguage or international communication might be necessary, but because it's strong on its own merits. (You can't confuse 1,234 and 1.234, most importantly.) It's weakest feature is the fact that American audiences are very likely to be unfamiliar with it. On balance, however, most Wikipedia users are likely to correctly understand numbers formatted in this fashion, even if it takes a moment the first time they see it. As a bonus, the use of CSS allows this number to be copied without digit grouping symbols (i.e. copy 4046.8564224 and the clipboard contains 4046.8564224), makes the number non-breaking, and employs spacing in an unobtrusive and readable fashion. Item #2 is acceptable and preferable (and therefore should always be permitted as an option).

  In terms of item #1, I don't particularly like the mixing of delimiting characters, but that's an aesthetic issue with relatively little traction. From a readability perspective, it is superior to not grouping digits on the RHS, because the thousands places are clearly displayed. Given that we will presumably maintain a version of the U.S. customary digit grouping method, I wouldn't be opposed to #1. (I've seen a variant on this using commas on both sides, but I'm informed that in modern U.S. practice this is rare. I wouldn't be opposed to that either.)

  For #3, if the output with or without {{gaps}} would be identical, then there's no need to require the template. If the number needs digit grouping (according to whatever style is in effect), then the output ought to be consistent within the article. I definitely like the concept of {{val}}, because of its ability to format difficult numbers. Its design is such that for some cases, however, {{gaps}} has to be used, because of technical limitations. We shouldn't mandate converting an article exclusively to one or the other, when both output similarly-formatted results. TheFeds 06:52, 13 June 2009 (UTC)

• Commas to the left and spaces to the right looks weird, to me. Why does example 1 have a three- then a four-digit group to the right? Tony (talk) 07:35, 13 June 2009 (UTC)

• TheFeds, both {{val}} and {{gaps}} provide the appearance of gaps without actually including thin space characters, so both facilitate cut-and-paste. Both can accomodate a unit of measure with a non-breaking space between the number and the unit. I imagine that if a pasted number has no commas, more computer applications will accept it (but an important one, Excel, will accept the commas).

• Tony, in the case where a number has digits to the right of the decimal, and the right-most digit would otherwise be by itself, the ISO/BIPM/NIST style allows (but does not require) that the thin space be eliminated. (In other words, if dividing the number of digits to the right of the decimal by three gives a remainder of 1, the rightmost thin space may be dropped.) --Jc3s5h (talk) 13:47, 13 June 2009 (UTC)

• Thanks, Jc. User:JimP created a combined thin and non-breaking space for just such an application, I think in 2007. It was dropped when we realised silly Internet Explorer displayed it wrongly. Is this function now available? Tony (talk) 14:35, 13 June 2009 (UTC)

• Both templates prevent breaks in numbers and the associated units. If an actual space character were found in Unicode that is both non-breaking and thin, some browsers would not display the character correctly and many applications (such as Excel) would not parse the number correctly if it were cut-and-pasted. I think including actual space characters of any kind should wait until most numeric-oriented computer software and most computer languages can deal with numbers in that format. I expect that to happen around the same time °, μ, and Ω appear on computer keyboards. --Jc3s5h (talk) 15:20, 13 June 2009 (UTC)

• Get a Mac. Here: °, µ, and Ω. All from the keyboard. Always the same in all applications because it is a system-wide resource. Greg L (talk) 15:42, 13 June 2009 (UTC) BTW, going off the keyboard of the Mac also gives you the special Unicode “micro” symbol µ that most people would want rather than the Greek letter μ. Greg L (talk) 15:45, 13 June 2009 (UTC)

• The Unicode narrow no-break space (U+202F) is available, and seems to render without error on IE7. (For more examples of Unicode spaces, see here.) The regular thin space breaks, unless surrounded by non-breaking markup. Unfortunately, you're right that most software wouldn't know what to do with spaces other than the regular one. TheFeds 17:18, 13 June 2009 (UTC)

• Jc3s5h, I think we're talking past each other a little bit: I agree with you that the CSS margin spacing is an advantage of {{gaps}} and {{val}} (and said so above, using the clipboard example). TheFeds 17:18, 13 June 2009 (UTC)

• I suspect that most have little idea of how much effort went into deciding how much of a CSS margin space went between the numbers. Different browsers and operating systems produced different results. So a bunch of screen shots were taken by editors who participated in the project and they were e-mailed back and forth so others could understand what still others were seeing. Some browsers could resolve to 0.05 em; others couldn’t and only resolved to 0.1 em. Some browsers would round “0.25 em” up and others wouldn’t. One editor looked at sandboxes using his iPhone. And all that was just to settle the width of the gap. The discussions for {val}'s functionality and protocol was just as lengthy. The template is locked down for a reason too: we need stability in our articles. Greg L (talk) 17:26, 13 June 2009 (UTC)

Let me know if you guys ever figure out what number format works best for Wikipedia. I'll be afk for a month, so please take your time if you need me to update {{val}} after you're done. May I suggest you make the final decision as explicit about number formatting as possible to prevent having to go through yet another one of these discussions again in the future?     — SkyLined (talk) 22:19, 13 June 2009 (UTC)

• So that we don’t go over old ground here, note that {{
  a bit lower, also here on Archive 94 on WT:MOSNUM, and…

  …it was well received here on WT:MOS Archive 97 where its functionality tweaked to a compromise that received broad acceptance.

  Also, a main testing sandbox is here on a subpage in my userspace, which also has links to other related sandboxes. Greg L (talk

  ) 02:21, 14 June 2009 (UTC)

Guys, I've read over this discussion. This is at the cutting-edge of how to strike an appropriate balance between non-expert and expert readers. It may be that the relevance of the discussion overwhelmingly involves expert/semi-expert readers: is that the case? Although it's not close to my expertise, I can't quite see why the {val} template needs to be fixed, and why the guidelines for using {gaps}, and how editors should hand-delimit numbers, need to be fiddled with. I've seen the relevant guidelines evolve over the years, albeit at a distance, and have had the impression that their specificity WRT formatting numbers probably minimises confusion and maximises consistency between articles for our readers. In particular, I'm concerned to minimise the potential for confusion among semi-expert and non-expert American readers, if indeed we do need to factor non-experts in (I'd appreciate your thoughts). Tony (talk) 16:36, 16 June 2009 (UTC)

I don't want to needlessly confuse non-expert readers. However, I don't know which they would find more confusing, gaps instead of spaces to the left of the decimal, or two different methods to group digits being used in the same article, and maybe even in the same number. After all, they are used to two concepts in what they usually read: commas as the grouping character, and only one grouping character. All the proposals on the table violate one or the other concept. {I'm using "character" loosely, to include gaps that look like thin spaces but really are not.) --Jc3s5h (talk) 17:07, 16 June 2009 (UTC)

As I get older and my vision gets worse, I become more aware of problems of accessibility. I have to say that 4,046.8564224 would be my preferred format from a readability point-of-view. Even when commas and decimal points get blurry, I know that the last one is the decimal point and can tell immediately that the number is four-thousand and something; I can also see at a glance, because of the spacing to the right of the decimal point that it has precision to the seventh decimal place. It may seem obtuse to prefer commas to group digits to the left and thin spaces to group digits to the right, but it actually makes the number more recognisable for me. --RexxS (talk) 00:43, 17 June 2009 (UTC)

(unindent) RexxS, I can see why accessibility is an argument against using commas both to the left and right of the decimal, but I don't see why it would be an argument against using thin spaces to both the right and left of the decimal. --Jc3s5h (talk) 02:04, 17 June 2009 (UTC)

It's only a small preference for commas over thin space on the left, so I'm not opposed to either, it's just that I can miss the decimal point momentarily in something like 4046.856422 and it looks like 4 billion and something at first glance or when I don't have my reading specs on. Having the commas on the left doesn't guarantee I don't misread the number, but they help me to identify the significand a little easier. Hope that makes sense. --RexxS (talk) 02:30, 17 June 2009 (UTC)

I was kind of invited to this RFC. Some notes I took during the reading of this discussion:

• IEC "256 Mebibyte" seems good. But in some entrenched areas it's just hard to motivate. And "mebi" sounds weird ;)
• Maybe it's feasible to make numbers with per user configurable view-format?
• SI units are actually very useful. Because the old system with BTU, feet, mile, inch, ounce, etc.. have several definitions depending on often unclear context. Besides needing a lot of "magic constants" (with varying precision as well) to make simple physics formula to work.
• Also beware that 1.2000 * 10⁴ specifies number of value digits. But 12000 really won't. Which is the reason why the first number is used in many cases (101 in college physics).
• Copy and paste to other tools for calculation usually restricts everything to 12.3456 format. Most wikipedia numbers must be converted at all times.

In the Mile article the number "1,609.344" combines "," and "." and makes this a really confusing number. In school "," was taught as the comma delimiter. Besides the number can't be copy & pasted directly into any other calculation tool. And on a larger level the article just shows that a measurement such as a mile or mph etc.. have an unclear definition in contrast to km/h (1000 meters per 3600 seconds). At least it's not straightforward. Electron9 (talk) 01:53, 25 June 2009 (UTC)

This RfC really isn't about units of measure, just the number. As for copy and paste, numbers with commas (as digit grouping characters) can be pasted into Excel, but some other applications will not accept them. (The behavior of Excel might depend on the language and location settings; my operating system is set for US English.)

In what country did you attend school? It is really essential that the period be used as the decimal marker in Wikipedia. Indeed, if this cannot be enforced by placing it in the Manual of Style, we should propose it as a policy. Every Wikipedia article I have ever seen followed this convention. The only alternative would be to include a statement in every single article that contains numbers as to what caracter represents the decimal point. --Jc3s5h (talk) 02:09, 25 June 2009 (UTC)

I find I'm able to copy & paste both 1,609.344 and 4046.856422, not only into Excel, but also into the Windows calculator. What "tools for calculation" don't accept them? For "numbers with per user configurable view-format", please see the

date autoformatting saga. --RexxS (talk

) 02:27, 25 June 2009 (UTC)

On June 3, User:TheFeds changed the digit grouping section. I endorse the spirit of the change, but I mercilessly edited it to conform to my understanding of many printed style guides (I suspect my edit reflects what TheFeds had in mind). I wish to call editors attention to the fact that the {{Val}} template does not conform to the changed MOSNUM, in that it uses commas to the left and thin spaces to the right, even in the same number (e.g. 123456.78901). If this change is accepted, the Vals template can be changed to conform to the change. --Jc3s5h (talk) 11:53, 3 June 2009 (UTC)

Has anybody ever seen a number written as "0.123,456" anywhere? --A. di M. (formerly Army1987) — Deeds, not words. 14:40, 3 June 2009 (UTC)

Normally, I would refer to the BIPM/ISO 31-0/NIST convention which suggests spaces, on both sides—that's definitely the dominant convention in most modern (e.g. post-1980) texts with which I'm familiar. Other style guides (e.g. Chicago) just suggest grouping thousands (mathematically, 1000ⁿ) with commas, but don't specify whether n < 0 counts (as it does in the BIPM convention). It's not quite fair to cite an elementary-school introduction to these things, but the general guidance that we received involved using the BIPM convention preferentially, or using commas instead (on both sides) if desired. That seems to carry over to hand-written work with which I'm familiar. I've also seen commas used on the right of a decimal point in older books (like, early 20th century). Somewhat-recent American texts using imperial/customary units usually use the commas-on-the-left-only style, but when they convert to SI editions, they seem to adopt the BIPM convention. Tables of values often use digit grouping symbols on both sides—but they're just as likely to omit them altogether. So basically, the edit I made was referring to accepted style manuals, and when unclear, erring on the side of mutual consistency; it looks like you might prefer erring on the side of historical practice in the U.S.A. and (perhaps) Britain? Maybe we should be emphasizing the BIPM convention, because it is less ambiguous on this point? TheFeds 21:41, 3 June 2009 (UTC)

The

Chicago Manual

14th ed. (current edition is 15, but I don't have it) says on page 310 "commas should be used between groups of three digits, counting from the right....Exceptions to this rule are ...decimal fractions of less than one". Later on the page, under SCIENTIFIC COPY, it suggests the BIPM style.

The

Publication Manual of the American Psychological Association

5th ed. on page 130 gives the example "numbers to the right of a decimal point 4,900.0744". I wonder if the examples of using commas to the right of the decimal point mentioned by TheFeds were in English? This is, after all, an English encyclopedia.

As for emphasizing the BIPM convention, I think that is unrealistic, since most Wikipedia editors are not scientists, engineers, or the like. It is really only in science and engineering that the BIPM convention has made any headway in the US. It is generally rejected when dealing with money, due to the opportunity for forgery. Far more Americans write checks than read scientific journals. --Jc3s5h (talk) 22:23, 3 June 2009 (UTC)

I'm quoting from memory, but I'm quite sure that the Numbers section of the 15th edition of Chicago says something like 'In numbers with a thousands place, use commas to group digits in threes, counting from the decimal point'. It's possible that it also states the exception for decimal fractions less than one, but I don't remember it. It goes on to give the BIPM convention as an alternative. (If anyone happens to have access to that, they can check up; it's been a couple months since I last looked through the actual document.)

Either way, no commas on the right of the decimal isn't a huge deal, and I'd be amenable to either style. But if we're relying upon a particular style guide as our basis for the decision, we should be clear whether or not following one recommendation in a guide is an endorsement of the rest of the guide. (I suspect that it isn't necessarily.) TheFeds 03:42, 4 June 2009 (UTC)

I've deliberately avoided chigago manual of style since i am not american but my 1967 copy of the engineers number handbook (a little book full of figures and conversions) says that it should be written with commas in the integers and spaces in the decimals. i have never actually seen it written that way but that is what it says and that would come from ISO. I think WP:COMMON sometimes must prevail. SimonTrew (talk) 11:02, 6 June 2009 (UTC)

Can you provide bibliographic information for the engineers number handbook? Does it indicate that units should always be chosen so that most of the digits always fall to the left or rigit of the decimal point? If so, that is not feasible for templates. --Jc3s5h (talk) 16:20, 6 June 2009 (UTC)

Will do but I amjust in the process of moving so all my books are in boxes. And its only kinda a little handbook so it will take me a while to find as I unpack! Please bear with me. SimonTrew (talk) 02:55, 19 June 2009 (UTC)

Haven't forgotten just still unpacking books and it is such a slim volume is easily missed. It is maroon cover almost a pamphlet I would guess about 64pp. and the front is in that newfangled kinda neon font in white on maroon and I am not sure if it has an ISBN. However please could you remind me if I havenot done so by the end of the month, I am liable to forget these things. If it hasn't an ISBN (and would no doubt be out of print anyway) there's not a lot we can do as far as RS I guess, as it is probably mid 1960s, but I can at least as far as MOSNUM goes tell you what it says to give you and so give you an alternative way to toss up a decision where it differs. It has logs and trig tables and stuff but also all kinds of conversions etc. I am not sure I even have the title right, when I think of it, I think I do but I am doubting myself now. I have it in my visual memory but maybe I have not got number right it may be conversion or something like that. SimonTrew (talk) 18:24, 23 June 2009 (UTC)

Found it, it's actually called the Science Data Book. My memory must be faulty as it uses spaces both sides of the point. It is totally silent on the matter of formatting actually. Anyway for completeness it is published by Olicver & Boyd 1971,

ISBN 0 05 002487 6, Edited R M't Tennant. I *thought* it had a preamble with this stuff in it, but must be mistaken, just refers to various conventions on weights & measures etc. Generally it uses <3 to left of d.p. i.e. scientific/engineering notation, i.e. use 10^3, 10^6, 10^-9 etc but not 10^7 or 10^-2. (can't be bothered to type all those sups!). It does very occasionally veer from this but only when explicitly making a contrast etc. SimonTrew (talk

) 07:09, 29 June 2009 (UTC)

I just tracked down what the Chicago Manual of Style, online 15th ed. says on numbers:

9.59 Comma between digits
In most numerals of one thousand or more, commas are used between groups of three digits, counting from the right. (In scientific writing, commas are often omitted from four-digit numbers.
1,512
32,987
4,000,500
No commas are used in page numbers, addresses, and years (though years of five digits or more do include the comma.)
Punctuation conventions can be found on page 1535 of the tenth edition.
Our business office is at 11030 South Langley Avenue.
Human artifacts dating from between 35,000 BP and 5000 BP have been found there.

9.60 Space between digits
In the International System of Units, half-spaces rather than commas are used to mark off groups of three digits, both to the left and to the right of the decimal point. In numbers of only four digits either to the left or the right of the decimal point, no space is used (except in table columns with numbers having five or more digits). This system is far more common in Europe than in the United States. Chicago’s Astrophysical Journal, for example, uses commas, not spaces (see bibliog. 5). See also 9.22, 9.59.
3 426 869
0.000 007
2501.4865 (four-digit numbers require no space)
See B. N. Taylor, Guide for the Use of the International System of Units (SI) (bibliog. 2.4).

That's mostly along the lines of what I remembered, and doesn't really talk about RHS commas, except to suggest that the Chicago Astrophysical Journal uses them in preference to spaces.

It also looks like Chicago doesn't go into what to do about (for example) seven digits on the RHS when using thin spaces as the digit grouping symbol; it just says "four-digit numbers require no space"—which is like the BIPM's text, but not their examples. That suggests to me that they don't recommend grouping seven digits on the RHS in groups of 3 & 4, but rather would use 3, 3 & 1. TheFeds 17:55, 29 June 2009 (UTC)

For example, if an article currently reads there are 43,560 square feet in an acre and an exact SI conversion is added, should it revised to say there are 43,560 43560 square feet (exactly 4046.8564224 m²) in an international acre? --Jc3s5h (talk) 17:29, 9 June 2009 (UTC)

Being a Dane living in England, I have lived with two different ways of presenting the same information (Denmark uses a comma as the decimal separator and a point as the digit group delimiter, whereas England uses a point as the decimal separator and a comma as digit group delimiter) and I find the BIPM way the least confusing solution as 1,234.567 is ambiguous: Is it 1234.567, i.e. the English interpretation, or 1,234567, the Danish interpretation? By using spaces only as digit group delimiters, no number will be ambiguous, whether a point or a comma is used as decimal separator. —Preceding unsigned comment added by 79.78.201.76 (talk • contribs) 17 June 2009

• It is not the job of Wikipedia to promote change in the way the world works by adopting good ideas by standard bodies just because they are good ideas and come from standard bodies. Wikipedia generally simply follows the way the world really works. Truly good ideas from standards bodies usually rapidly gain traction in the real world. When that happens, Wikipedia follows suit. In other cases, such as the IEC’s proposal that the world adopt the new binary prefixes like “256 mebibyte”, the proposal falls on its face. And again, notwithstanding the compelling virtues of the IEC’s proposal, Wikipedia ignores the proposal so as to not confuse our readership.

  As regards the delimiting of large numbers and/or high-precision numbers in all our general-interest articles, Wikipedia specifies the U.S. method of delimiting numbers using commas in the significand because that causes the least confusion with its readership. U.S. readers are far and wide entirely unfamiliar with other delimiting conventions whereas their European counterparts are typically comfortable with several different methods and are not in the least confused by the U.S. convention. Scientific articles are a notable exception, since the language of science is scientific papers which follow the SI method so that such papers can be read by a world-wide audience. Even then, delimiting with narrow spaces is allowed in our scientific articles—not required.

  Please remember that the overriding principal on Wikipedia is to follow practices that cause the least confusion with its readership. Note also that the rule of SI is to have a space between the number and the percent symbol, such as “50 %” rather than the “50%” we are all accustomed to. So that our articles cause the least confusion and look most natural, we wisely ignore that suggestion (yes, even thought it comes from the BIPM and their SI).

  I note that the two of you (The Feds and you, Jc3s5h) have been having a discussion amongst yourselves, above, and have been quietly making changes to MOSNUM without buy-in and consensus of the rest of the Wikipedian community. I know you clearly believe what you are suggesting is a better way to do things. Please bear in mind that MOSNUM guidelines are often the product of tortuous and lengthy (often very lengthy) discussions and are done a certain way for a reason. When a situation arrises where a proposal, as you wrote above, has not been challenged for nearly a week, that does not equate to acquiescence and agreement with a proposal. Please widen the discussion so there is a clear consensus before revising MOSNUM. Greg L (talk) 02:09, 13 June 2009 (UTC)

  In

  WP:SOURCES ("In general, the most reliable sources are peer-reviewed journals and books published in university presses; ") we should use scientific journals as our best source, and from this it follows that we should try to addhere to standards like this (used in these journals all over the world, even in the US) Nsaa (talk

  ) 15:34, 18 June 2009 (UTC)

The period (stop) is well-entrenched as the decimal point in the anglosphere. Gee, those full spaces look awkward on my display—no wonder many people recommend thin spaces. Tony (talk) 16:07, 18 June 2009 (UTC)

But this isn't an 'Anglosphere' Wikipedia, our readers come from almost every country, and the use of the comma is common in South Africa and Australia it appears, so I'm not even sure it is well entrenched in the 'anglosphere'. Our article Decimal separator is a help on this.

Nevertheless, it is the 'English' Wikipedia, and it would seem odd to adopt a convention that is not in use among the (English) sources that we use. I can find no reference to Australia using a comma as decimal separator; in fact the article you pointed to, Decimal separator, lists it with all the other English-speaking countries in using the period. The same article only lists South Africa among English-speaking countries as using the comma (and that is qualified as "officially[citation needed], but dot point is commonly used in business"). I am pretty convinced that the use of the period is very well entrenched in the Anglosphere. --RexxS (talk) 18:09, 23 June 2009 (UTC)

Comparison of texts in "Large numbers"

This discussion came about as a result of two different versions of part of the "Large numbers" section. The first example below incorporates some clarifications I made, plus some changes Jc3s5h made. The second example is the prior version of that section. Now that we've been discussing it for a while, I'd like to see whether others feel that the first example is clearer overall, and should therefore be the basis for the eventual revision of the section. I know I'm of that opinion, principally because that version is clearer about what is permissible and in what situations, states individual possibilities in separate bullets, and covers cases that are ambiguous in the other version. Does anyone feel that the second option is clearer?

• In a number with many digits, digit grouping symbols (inserted at intervals from the decimal point) should be used to subdivide the number into easily readable groups. The acceptable digit grouping schemes are:
  • Commas every three digits (8,274,527) only to the left of the decimal point, with no grouping to the right of the decimal point. This is traditional usage in many English-language contexts.
  • Thin, non-breaking spaces every three digits (8274527 or 0.12345). This format is suggested in
    NIST style guides for scientific works, and is in common use in interlanguage contexts. The {{gaps}} template uses CSS
    to produce this output (using the syntax {{gaps|8|274|527}}). The   character may cause rendering problems in some browsers, and should be avoided when practical.
  • Other traditional digit grouping schemes, when relevant to the subject matter of the article (e.g. 82,74,527 in the Indian numbering system).
• When a number has exactly four digits on any side of the decimal separator, those digits may optionally be expressed as a group of four instead of three (e.g. both 9876 and 0.9876 are acceptable).

• In large numbers (i.e., in numbers greater than or equal to 10,000), commas are generally used to break the sequence every three places left of the decimal point, e.g. 8,274,527. In scientific and mathematical contexts, {{gaps}} may be used to insert thin spaces, e.g. {{gaps|8|274|527}} produces 8274527 (note: the thin space character and its HTML entity,  , do not render correctly on some browsers). Consistency within an article is desirable as always.

If we generally prefer the first example, we should implement that change to the MOS (it got reverted during the course of our discussion), because it's generally independent of the RfC going on below. (The result of the RfC can easily be added, once it is established—until then, we can stick to the status quo for the U.S. customary style guideline) Any objections or suggestions? TheFeds 07:04, 19 June 2009 (UTC)

The first version does not make it clear that for the thin space scheme, the thin space may be omitted to the left of the decimal if and only if there are exactly four digits to the left. But to the right of the decimal, the rightmost thin space may be omitted if the rightmost group would otherwise contain only one digit. That is, 7682533 is wrong but 0.768 2533 is OK.

Also, neither version endorses the current behavior of the {{Val}} template. --Jc3s5h (talk) 14:23, 19 June 2009 (UTC)

While I see what you're getting at, it doesn't seem to be something that NIST or the BIPM endorse (they say exactly-four-digit groups on either side are the lone exception). I'm not so sure we're well-served by having an exception to the exception (for 3n+4 digit groups on the RHS). Is it particularly advantageous to use that style on Wikipedia?

Per the BIPM:

Following the 9th CGPM (1948, Resolution 7) and the 22nd CGPM (2003, Resolution 10), for numbers with many digits the digits may be divided into groups of three by a thin space, in order to facilitate reading. Neither dots nor commas are inserted in the spaces between groups of three. However, when there are only four digits before or after the decimal marker, it is customary not to use a space to isolate a single digit.

Regarding {{val}}, I think the best way to handle it is to decide what the preferred formats ought to be (in the MOS talk arena), and then afterward, discuss changing {{val}} so that it meets one. Alternatively, if the RfC eventually determines that the format currently used by {{val}} is acceptable, we write it in as another option (distinct from the BIPM and local styles) TheFeds 22:30, 19 June 2009 (UTC)

TheFeds, the BIPM brochure which spells out the quote you gave contains examples where there are more than four digits to the right of the decimal, yet the rightmost group consists of 4 digits. See the number 9.109 3826 (16) × 10⁻³¹ kg on page 126. It seems that the BIPM didn't do a very good job of writing the rule that describes what they do in practice (or maybe the translation from French to English isn't very good).

As for which version at the top of this thread is better, I am not willing to choose between two versions, neither of which describes a fairly popular template ({{tl:Val}}). I think we have to wait to decide what we want to say before we can decide how to say it. --Jc3s5h (talk) 10:50, 20 June 2009 (UTC)

Quite right: I see those examples and wonder what the BIPM was meaning to say. It's not a translation error—the French means the same thing.

Cependant, lorsqu’il n’y a que quatre chiffres avant ou après le séparateur décimal, il est d’usage de ne pas isoler un chiffre par un espace.

I guess that lends more weight to the idea that we ought to be considering what's good for us, rather than following the convention religiously. TheFeds 17:08, 20 June 2009 (UTC)

I was browsing through the archives on this topic, and came up with several discussions that relate to our current questions:

• Archive 19#Spaces as number separators
• Archive 24#Units section, continued
• Archive 74#Commas in numbers
• Archive 87#Spacing after the decimal point
• Archive 94#Grouping of digits after the decimal point
• Archive 94#Grouping of digits after the decimal point (next attempt)
• Archive 98#{{delimitnum}} template
• Archive 122#WP:ENGVAR and Indian number format (lakh/crore)
• Archive 122#High-precision numbers

To summarize, in Archive 19 (May 2005), it was suggested that spaces, or better, thin spaces be used. ISO 31-0, BIPM and Chicago were cited. Editors were concerned about technical limitations in browsers (not supporting thin space characters). Objections to spaces included that text-to-speech was affected when used. No consensus, because of technical limitations, however it was agreed that either commas or non-breaking thin spaces were acceptable typographically.

In Archive 24, as part of a larger discussion on when to use SI vs. regional units, it was suggested that the &nbsp; character be used for digit grouping, due to technical limitations. This didn't really get anywhere, for unrelated reasons.

In Archive 74, a similar question to the above was posed: should digit grouping exist on the RHS? Consensus was that ISO 31-0 permits digit grouping with thin spaces on the RHS, and would be acceptable, but IE6 did not support them well. Some editors expressed preferences for commas, others for thin spaces (all relating their experience in English-speaking contexts). International preference was for thin spaces.

In Archive 87, it was suggested that 0.249 592 9 was ugly and unfamiliar. Another editor suggested that it should be permitted, given SI standards, and the goal of reducing ambiguity. There were a few

WP:IDONTLIKEIT

arguments, with editors advocating against certain regional preferences. Eventually agreed that SI formatting was less familiar, but more readable, especially when using comma-based lists of numbers. Technical limitations again come up, with regard to thin spaces. Ended in proposal to allow either no spaces after decimal point (U.S.), or spaces every 3 places (SI), or no spaces with editors' judgment in special cases. No consensus between those 3 options.

In Archive 94 (December 2007) the same issues were raised, but with a potential solution to the technical limitation against the use of thin spaces: CSS margins. This was apparently the origin of the proposal to group with commas on the left, and thin spaces on the right (current behaviour of {{val}}). Has advantage of being pasteable into computer programs, while still being readable. Viewed as good technically, and by proponents of allowing SI usage, but some editors still did not support using spaces (ever). Later devolved into personal bickering.

But later on the same page (Archive 94, February 2008), a more technically-refined version of the previous proposal was made. This addressed details of how wide CSS margins ought to be. This also explained a rationale for using commas on the left, and gaps on the right of the decimal—mainly that Americans were familiar with commas, but that the RHS needs delimiting for readability. It was considered rare that both commas and spaces would exist in a single value, but it was suggested that this was not a problem when it occurred. It was also suggested that WP:MOSNUM policy should reflect a particular formatting style ({{delimitnum}}). In the discussion, there was broad support for a formatting function, but some expressed concern that spaces and commas should not be mixed within a number.

In Archive 98 (March 2008), more details were given about the implementation of {{delimitnum}}. Here, {{val}} was introduced, as an alternative that didn't perform the spacing. (It would later use {{delimitnum}}'s spacing.)

In Archive 122 (April 2009), there was a discussion about using the Indian English terms lakh and crore (derived from Hindi, etc.). Some said that these were not appropriate in English usage, others argued that there were many English speakers in India, and that their usage should be respected. It was suggested that the use of these terms be linked. There was no consensus on how to handle digit grouping (which is different in this system).

Later in 122 (May 2009), concern was expressed that the text of the

WP:MOSNUM#High-precision numbers

was inadequate. This led to renewed discussion of why MOSNUM advocates commas on the left and spaces on the right, and that {{val}} implements this. Some felt that this compromise (to an unusual format) was better, others felt that it was contrary to any accepted external standards. No consensus was reached.

One major theme throughout this discussion was that technical limitations (mainly: IE6 did not handle certain Unicode spaces properly) prevented the use of spaces. This was addressed by implementing the CSS gap method, which is apparently well-supported on all major platforms (including modern versions of IE; IE6 is now dated software, and supporting it is not such a critical issue). Similarly, there was the need to paste numerical text from the browser into number-crunching software (e.g. Microsoft Excel). The CSS gaps allow this, whereas thin spaces were unreliable. (Also, on a related note, copying and pasting chunks of text can be affected by the spacing method. By using CSS, the spacing is encoded in the markup, rather than the text; this means that derivatives of Wikipedia will end up with unformatted numbers when quoting it. I don't think this is a big issue.)

I noticed that there are apparently two schools of thought about formatting numbers: SI is better or U.S. customary is better. (I've already expressed why I prefer SI.) However, given this divide in opinion, with reasonable justifications provided on both sides, it seems logical to suggest both on an equal footing, and leave the choice to the editors. The idea to use the hybrid format (commas on the left, and decimals on the right) in {{val}} and elsewhere apparently originated with Greg L, who advocates it as a compromise. Unfortunately, I (and apparently others, judging by the discussions summarized above), think that it has the disadvantage of satisfying neither camp. I'm personally of the opinion that if we continue to use that format, and people keep wondering about why it exists, and indeed whether it should exist, we invite the very long discussions for which this topic has become known.

In much the same vein as I wouldn't insist that editors commit to using the de facto "international" standard (the SI), and instead would allow the current U.S./historical British comma-based digit grouping scheme, editors also should have the choice to use (for example) Indian numbering, where the article context is appropriate. If at a future date, Wikipedia wants to standardize on the SI, I wouldn't be opposed to it. But for the moment, I think that the best solution is to avoid prescribing a particular hybrid format in the MOS, and instead allow editors to choose among the two prevailing conventions at will, and to choose an appropriate regional format when the article context would permit it. That's the thinking behind the wording that I used for the MOS.

So, I've made a couple minor adjustments to the text:

• In a number with many digits, digit grouping symbols (inserted at intervals from the decimal point) should be used to subdivide the number into easily readable groups. The acceptable digit grouping schemes are:
  • Commas every three digits to the left of the decimal point, and no grouping to the right of the decimal point (e.g. 8,274,527 or 0.12345). This is traditional usage in many English-language contexts.
  • Thin, non-breaking spaces every three digits (e.g. 8274527 or 0.12345). This format is suggested in
    NIST style guides for scientific and engineering works, and is in common use in interlanguage contexts. The {{gaps}} template uses CSS to produce this output (using the syntax {{gaps|8|274|527}}). Using HTML entities
    for this purpose (e.g.   or  ) may cause rendering problems in some browsers, and should be avoided when practical.
  • Other traditional digit grouping schemes, when relevant to the subject matter of the article (e.g. 82,74,527 in the Indian numbering system).
• When a number has exactly four digits on any side of the decimal separator, those digits may optionally be expressed as a group of four instead of three (e.g. both 9876 and 0.9876 are acceptable).

One consideration of this particular phrasing is that it gives equal weight to the alternatives—I think that will keep arguments over which format is most popular to a minimum. Also, the text is not overly prescriptive (this came up previously in the discussions); given that editors may ignore rules that don't improve the encylopedia, it isn't really necessary to insist that edits "must" be a certain way. Combined with the MOS caveat that exceptions sometimes apply, it's appropriate to say that certain things "should" be done. (So instead of saying "things are generally done", as if it's fait accompli, suggest that "things should be done".)

I'll defer to others' experience with regard to digit grouping using commas on the RHS. It's clear that no RHS grouping is acceptable according to some major style manuals and publications that use commas on the LHS.

I've retained the four-digit grouping language, because it seems that even the BIPM can't get this straight. The text is a faithful statement of what they prescribe. If someone follows the BIPM's examples (in contrast to the text of their standard) and groups the four rightmost digits of a number with seven digits on the RHS of the decimal, it would seem fair to call that an allowable exception under the MOS. If we want to, it would be straightforward to diverge from the BIPM's text and specifically allow other groupings (per Jc3s5h's suggestion).

As for {{val}}, {{delimitnum}}, {{scinote}}, etc., judging by the discussions I reviewed, I think the best way to handle them is simply to have them default to SI-style spacing (given that they are designed to handle scientific numbers), with a very simple parameter to allow other styles (U.S., Indian, etc.). The real issue is that essentially nothing endorses the way {{val}} works now—it's a compromise between readability and U.S. convention, and judging by the discussions, that formatting scheme (as distinct from the rest of the issues) never really achieved any sort of consensus—it was just along for the ride as part of the overall project to improve formatting of numbers. (The comments about formatting in that manner were overshadowed by other details, most notably CSS gap widths, but were generally slightly negative.)

Also, I've looked into the code of {{val}}, and it's straightforward to change; the necessary formatting already occurs on the RHS. SkyLined has suggested creating separate versions of the template for internationalization, but because of the difficulties inherent in supporting multiple sets of code—look at the discrepancies in the {{cite $thing}} templates, for example—we would want to do this in the backend of {{val}}, and make sure users get to see the template documentation as a family. (That discussion can continue at Template talk:Val; for MOSNUM's purposes, I think that's a fair resolution.) I don't know if it's necessary to require a particular template in MOSNUM (too many exceptions would apply), but it would be fine to recommend it (after the changes are made).

Are there any other aspects of this discussion that I haven't addressed? TheFeds 21:23, 29 June 2009 (UTC)

Excellent. I would have never spent all that time to dig through all those discussions. Thank you. I think the parentheses in the last sentence was supposed to say "e.g. both 9876 and 9,876 are acceptable", right? While we are at it, we might add "although four-digit serial numbers such as years and page numbers are written without spaces or commas". --A. di M. (formerly Army1987) — Deeds, not words. 23:01, 29 June 2009 (UTC)

I concur with the change. Please consider this case: an article had some integers with five digits on the LHS, and used commas to group digits. A new number is added: 0.12345. Should the revised MOSNUM be read to mean that the article should be revised to use (the appearance of) thin spaces in the integers with five digits on the LHS? —Preceding unsigned comment added by Jc3s5h (talk • contribs) 03:38, 30 June 2009 (UTC)

Jc3s5h, if I understand correctly, you're thinking about maintaining consistency, which is definitely an objective from the general

WP:MOS. We probably wouldn't need to repeat that instruction here. A. di M., the example of 9876 and 0.9876 was intended to show that we could group four digits on either side of the decimal point. But 9,876 works too (as does 9876). TheFeds

04:33, 1 July 2009 (UTC)