# Wikipedia talk:Manual of Style (dates and numbers)/Archive 32

 Archive 25 ← Archive 30 Archive 31 Archive 32 Archive 33 Archive 34 Archive 35

## Spell out "source unit", symbol for "derived value"

Current text of the manual:

• Spell out source units in text. Use digits and unit symbols for converted values and for measurements in tables. For example, "a pipe 100 millimetres (4 in) in diameter and 10 miles (16 km) long".

I think this came in with a batch of changes without much specific discussion of this rule in its own right, though the use of hyphens in an earlier version has been discussed on the talk page.

This is a bad rule, and one not often followed in actual Wikipedia practice. For example, there are zillions of articles with measurements in the form x km (y miles). In fact, it is pretty obvious to me without even checking who did it that there have been editors (likely including at least one bot) who have gone through Wikipedia systematically spellig out "mi" whenever that appeared as a symbol for miles.

In actual use, in conversions between degrees Fahrenheit and degrees Celsius, in most cases either both are spelled out or both are in symbols, and to me that seems vrey logical.

I disagree with both parts of it: that "original units" should always be spelled out in text, and that "converted values" should never be spelled out in text.

The first problem, of course, is determining what is a "source unit". Often, a Wikipedia editor will get information from a source which includes dual measurements. Other Wikipedia editors will often have no certain information of what the source information was; sometimes there will be obvious clues such as an overly precise conversion, of course--in which case, that overly precise conversion should be changed as well.

Furthermore, in devising a general rule, don't just look at the simple cases. After all, it is the exception which tries the rule. All our conversions aren't going to be between feet and meters, or between kilometres and miles. Stop and think about what we should do when the conversion is between gigaelectronvolts and nanojoules, or between angstroms and picometers. We also often have conversions between the U.S. EPA units of "picocuries per liter" and the SI units "becquerels per cubic meter", or between blood sugar levels in millimoles per liter (mmol/L) and in milligrams per 100 milliliters (mg/dL) (even though most people who use one or the other of those numbers couldn't tell you what the units are if asked).

Or, consider this one copied from an actual Wikipedia article: " 1 ft²·°F·h/Btu ≈ 0.1761 K·m²/W". Do we really think that this should be presented in Wikipedia this way?

"One square foot-degree Fahrenheit-hour per British thermal unit per hour = 0.1761 K·m²/W" or ambiguously as "One square foot-degree Fahrenheit per British thermal unit per hour" or clarifying that by using parentheses for grouping spelled out words ""One square foot-degree Fahrenheit per (British thermal unit per hour)"

In many cases, it would greatly help the readability of an article to spell out a unit of measurement on first occurrence, and use a symbol thereafter. This applies whether it is the "source" or the "converted value" (something not always possible to discern in any case).

Note also that there are many instances in which measurements expressed in dual units can both be considered "source units". In fact, sometimes both of those numbers can be individually rounded off from some more precisely known value (but whose precision is irrelevant to the article at hand), so that you would not end up with the same two numbers if you converted from one of the rounded-off versions to the other.

In many measurement intensive articles, symbols should be used in text as well as tables. Furthermore, some symbols such as °F and °C can pretty understandably and safely be used in any spelled out text as well as in tables. Of course, "tables" needs to be interpreted in a broad sense, including many lists in text format as well as the wikitable format.

BTW, what is supposed to happen when the "source unit" is omitted (or removed by a later editor), and only the "converted value" remains? That can, in fact, be proved to have happened in some cases, and it has in fact happened in a great many cases. Gene Nygaard 12:56, 28 October 2005 (UTC)

I forgot to mention one other consequence if this recently added rule were to be widely implemented: we'll have a lot more quibbling about metre vs. meter, milliliter vs. millilitre, even some kilogramme vs. kilogram issues. Gene Nygaard 13:32, 28 October 2005 (UTC)
The problems that you state seem valid. I generally recommend the use of digit+symbolic format for tables and for units in parentheses. I had thought the guideline addressed that but now I see that it could have unintended consequences. It adds much less value than the other 7 bullets.
You described problems but did not propose a solution. If you propose removing that guideline, I will support you. Bobblewik 13:37, 28 October 2005 (UTC)

I think this guideline is a slightly too-specific application of the principle that in running text it can be preferable to spell out figures and units, while in tables, and possibly in parentheses, it can be preferable to keep them compact. This is a matter of style which requires editors to apply their judgement, and so shouldn't be over-specified, but bears mentioning.

Regarding metre/meter spellings, we already deal with realize/realise and labour/labor, so I don't think it's a good reason to avoid spelling out units and numbers. Michael Z. 2005-10-28 15:09 Z

I think you're missing the point of the guideline.
Perhaps it's badly worded. The idea is, such abbreviations should not be used in prose, in much the same way you wouldn't write "he has 1 pen", but instead "he has one pen". Converted units, however, are provided for and only for convenience, hence it is acceptable and even preferable to abbrievate them so that they disturb the text itself as little as possible (this is not to say they should not be there; they are very necessary.)
I don't see this as an issue of source unit versus converted unit; just body text versus additions for the covenience and aid of users. It is generally considered (from what I gather) to be bad style to abbreviate in formal text (except acronyms).
In scientific articles, however it may be acceptable to abbreviate to symbols in context. This requires some judgement.
With regard to spelling, the British vs. American follow-the-first-guy rule applies; gram is generally spelt as such in British English (I think).
I wouldn't want to remove it altogether, though perhaps it could be reworded to better reflect its purpose. It must always be remembered that any guidelines here are for the average case and must be applied with a certain degree of elasticity (that cannot be emphasised enough, and it is perhaps underrated in this manual). Neonumbers 22:44, 28 October 2005 (UTC)
It's a bad guideline even for the "'average' case".
The "convenience" involved is the convenience of those who use that measurement, not the convenience of those who use the other. Those who ignore one set of units deserve to get basically the same information, as far as both precision of the measurement and flow of the text, as those ignore the other.
It isn't just "scientific articles". Look at the climate sections of many of the articles about cities or other administrative units, for example.
You don't even state the rules for national varieties of English correctly. The last resort rule is first significant contributor, not a stub. Gram may be often spelled as such even in the British Isles, but it isn't universally spelled that way, and "gramme" is an acceptable spelling under Wikipedia rules.
When we are dealing with these conversions, we are dealing with measured quantities, not counting numbers.
It is better to underspecify our rules than to overspecify them. You can talk all you want about applying them with a certain degree of elasticity, but whenever some zealot (who might be me, for instance) can cite the unqualified, overspecific Manual of Style rules, that makes such flexibility difficult. Better to indicate some examples (not a comprehensive list) of situations in which the rule doesn't apply, or not to make or mention any rule at all when there is no consensus for a guideline of that sort in the first place. Gene Nygaard 13:01, 29 October 2005 (UTC)
Much of the "judgement" Neonumbers talks about deals with factors other than some imagined and hard-to-determine dichotomy between "source unit" and "derived value".
• It is more important to spell out less familiar units (but in this case, a distinction can often be made between first appearance and later use in the article):
For example, it is more important to spell out joule than to spell out watt, and it is more important to spell out any prefixed unit where the powers of ten go beyond ±6.
This is true even if that unfamiliarity is of a geographical nature (stone as the unit of mass used for body weight, for example).
It is more important to spell out newton meter than to spell out foot-pound force (in part because some are familiar with the now-deprecated non-SI metric units such as meter kilograms).
• Familiarity with, and ambiguousness of, the symbol are also important factors:
It is more important to spell out gram (symbol g) than to spell out kilogram (symbol kg).
On this basis, it is more important to spell out coulombs (symbol C) or teslas (symbol T) than to spell out becquerels (symbol Bq).
Some have even used a claimed unfamiliarity to remove "mi" as a symbol for miles, but everybody is familiar with "km" for kilometers. But whether these two units are spelled out or not is almost never determined in actual practice on the basis of which is the "source unit" and which is the conversion or the "derived value".
• A distinction can often be made between measurement-intensive text and that with only sporadic measurements.
Of course, that measurement-intensive text often contains information which would lend itself to presentation in a table, or some other kind of list, but that may not have been done yet, and may never be done.
Any distinction based on some vague notion of a "scientific article" invites quibbling. Gene Nygaard 13:41, 29 October 2005 (UTC)
I must be allowed to defend myself against the charge of failing to state the rule correctly. I used the term "follow-the-first-guy rule" to denote the rule in the manual, not to imply that the rule is quite literally that — I am well aware of the fact that it is a major contributor, and I assumed that I could assume that everyone else would know that too, and hence the major contributor past would be implied. With respect to the part about "gram" being spelt that way, in New Zealand we spell it that way, which is pretty much British English, so I added "I think" at the end just in case.
I cannot (and, I note, has not) be denied that abbrevations and prose don't go together (in most cases, acronyms and i.e. e.g. q.v. are exceptions).
Those aside, I'll take a second to talk about familiarity. There is a fine line — and I almost certainly do not mean to imply that you or anyone have crossed it — between ensuring understanding and assuming stupidity. I understand your concerns on familiarity.
That said, this is where I talk about judgement, and a comprehensive list of every unit imaginable would be a bit too far — not because it is too restricting, but because it's, well, impossible to make a comprehensive list and allow for everything. So we can make guidelines on what the editor should consider and allow him to make a decision based on that. (Here, it is worth re-emphasising that not every editor is required to make this decision; one can make a decision arbitrarily and have it "corrected" by a more familiar one, that is no harm done at all.)
So, all this in mind, the only plea I make is that the source or primary or first or main or whatever-you-want-to-call-it unit is spelt out because it is in prose and you do not abbreviate in prose. Those in brackets are arguably not truly in the prose part and can be abbreviated, but if you must I will not stand too hard to mandate their abbreviation. Neonumbers 11:30, 7 November 2005 (UTC)

## Context

I've touched on the issue of context in several discussion now, and I note that there is no explanation of its importance on this manual. The Main Manual quotes from Chicago. With dates and numbers, judgement of context is also important; when I think about it, it's what not to do that is of concern more than what should be done.

There is never a one-size-fits-all rule. Different articles are about different things and will command different formats. This must be recognised, but without going too liberal with formatting. In all articles about the same topic, for instance, all formatting should be the same (except infoboxes and the like).

Is there any support for including in the introductory section (before the TOC) a sentence emphasising the importance of context and judgement? I'll have to think of the wording, and that can be done later, but it's worth thinking about first. Neonumbers 22:59, 28 October 2005 (UTC)

## Degree signs

I've been informed by a user that the current convention on Wikipedia is to insert a space between degree signs and numbers, thus: "10 °C". This contravenes the practice of every major publisher in the English language, so I suggest (a) Wikipedia stops using this convention (the space is wrong) or (b) finds some extremely good justification for this eccentric usage, and writes it up on the manual of style. --DannyWilde 00:04, 31 October 2005 (UTC)

The current MoS rule is a good one, well justified in modern practice, even if other conventions are also used. There is no reason to treat temperatures differently from any other unit of measure. DannyWilde has already been discussing this on my talk page, where I hae cited the following authorities for this rule. Gene Nygaard 01:39, 31 October 2005 (UTC)
File:Riedi degree symbols.jpg
Example of printing of degree signs in Oxford University Press book "Riedi: Thermal Physics".
The current MoS rule is incorrect, and unjustified in modern practice, since no one uses it. I have checked documents published by every single publisher I can think of, both American and British, and in not one case do I find a space between the number and the degrees of temperature. Note that the references below are from standards laboratories, not from typographers or publishers. A space is never inserted between a number and a degree of temperature sign in normal practice. The rule is incorrect, and you are in effect vandalising Wikipedia by applying this incorrect rule to articles. Further, the fact that you removed referenced and correct information from a Wikipedia article shows what shaky ground you are standing on. The above right image shows a typical example of a typeset degree of temperature sign. I invite any readers of this discussion to examine any examples, in a professionally typeset publication, except those of the following standards bodies, and see if it follows the "space" or "no space" rule. --DannyWilde 06:43, 2 November 2005 (UTC)
The U.S. Government Printing Office is one of your "typographers or publishers", Danny.
Your original research is woefully inadequate if you can claim that all "professionally typset publications" follow a no-space rule. Publications such as Encyclopædia Britannica, the World Almanac, and Isaac Asimov's New Guide to Science (1984) are some of the "professionally printed" material using the "900° C" format, with a space but in a different location; that doesn't mean that it is generally considered acceptable in modern usage either. Encyclopædia Britannica followed that space between the degree sign and letter rule in 1971 and 1979 editions, when it was owned by the University of Chicago, whose University of Chicago Press publishes the Chicago Manual of Style.
Others use a space, but omit the degree sign, or omit both in the case of The Times. That doesn't mean we need to accept either omission.
Examples with space in proper position according to the modern rules:
USDA, Agriculture Handbook Number 66, The Commercial Storage of Fruits, Vegetables, and Florist and Nursery Stocks, pdf file
"Water loss of about 1% causes 6 °C (11 °F) product cooling (Barger, 1963)."
There are also differences in most common usage on a chronological basis and probably some on a geographical basis. The fact that our standards-keepers haven't yet achieved total uniformity doesn't mean we should ignore their rules, which do in fact have significant usage in the real world. Gene Nygaard 14:38, 2 November 2005 (UTC)
Random House Webster's Unabridged Dictionary, 2nd ed. 1998. freezing point ... The freezing point o fwater is 32 °F, 0 °C.
Here is one clearly indicative of the direction of the trend. World Book encyclopedia, s.v. temperature:
1991 ed.: "-273.15° C"
1996 ed.: "-273.15 °C"
2000 ed.: still "-273.15 °C"
Gene Nygaard 15:45, 2 November 2005 (UTC)

National Physical Laboratory, the keeper of the standards for the U.K., [1]
"A space is left between the numerical value and unit symbol (25 kg but not: 25-kg or 25kg). If the spelled-out name of a unit is used, the normal rules of English are applied."
This does not mention the degree mark. --DannyWilde 13:10, 2 November 2005 (UTC)
It would if it were an exception. Note also that these NPL rules refer the user to NIST SP811 for more detailed rules; they are well aware of NIST's more detailed statement of the rule including specifically with degrees of temperature; it cannot be an oversight. Gene Nygaard 15:53, 2 November 2005 (UTC)

BIPM (the keepers of the international standards) SI brochure [2] (this unofficial--only the French is official--English version uses British English, with "litre" spellings and the like, and it uniformly and consistently uses spaces between numbers and units).

NIST Guide for the Use of the International System of Units (SI), 1995:[3]
7.2 Space between numerical value and unit symbol
In the expression for the value of a quantity, the unit symbol is placed after the numerical value and a space is left between the numerical value and the unit symbol.
The only exceptions to this rule are for the unit symbols for degree, minute, and second for plane angle: °, ′, and ″, respectively (see Table 6), in which case no space is left between the numerical value and the unit symbol.
Example: α = 30°22′8″
Note: α is a quantity symbol for plane angle.
This rule means that:
(a) The symbol °C for the degree Celsius is preceded by a space when one expresses the values of Celsius temperatures.
Example: t = 30.2 °C but not: t = 30.2°C or t = 30.2° C
This reference also says that a space should be inserted between a number and a percentage sign. This is certainly not the normal or standard practice. --DannyWilde 13:10, 2 November 2005 (UTC)

Society of Automotive Engineers Standard TSB003, Rules for SAE use of SI (Metric) Units, May 1999:pdf file
7.3.10 When writing a quantity, a space is left between the numerical value and a unit symbol. For example, write:
35 mm, not 35mm; write 20 °C, not 20°C.
Exception: No space is left between numerical values and symbols for degree, minute, and second of plane angle. Example: 45°. However in SAE Practice, the ° symbol is not used for plane angle. The word degree is spelled out.
This is an interesting reference, but looking at the PDF file, the space after the number looks absolutely freaky. Is there any example of a real document typeset using this convention? --DannyWilde 13:10, 2 November 2005 (UTC)

U.S. Government Printing Office, Style Manual, 29th ed. (2000), section 10.6:chapter 10 pdf file
10.6. Any symbol set close up to figures, such as the degree mark, dollar mark, or cent mark, is used before or after each figure in a group or series.
\$5 to \$8 price range
5′–7′ long, not 5–7′ long
3¢ to 5¢ (no spaces)
±2 to ±7; 2°±1°
but
§ 12 (thin space)
from 15 to 25 percent
45 to 65 °F not 45° to 65° F

The last example is interesting; does the document specifically state that 65°F is preferred? There is a convention that a repeated mark must be separated with a space. The Chicago Manual of Style also mentions this, although they don't use a space before the degree mark. --DannyWilde 13:10, 2 November 2005 (UTC)

Incidentally, I also found this discussion: [4]. --DannyWilde 13:10, 2 November 2005 (UTC)

I'm in favor of using the modern, international standard; i.e., 37 °C (37 degrees Celsius, 310 kelvins, 310 K). --Simian, 2005-11-03, 04:05 Z
Internal consistency is good, but I'd like to point out that the choice of format is a matter of style, so it's not possible to state that there's one right way. In contrast to the technical publications listed above, the Times [scroll to "temperatures"] and Guardian [scroll to "temperatures"] both use the compact format "10C" and the former prefers "minus 20C" in running text, while the Economist [5] uses °F or °C, but its style guide isn't clear on spaces. Michael Z. 2005-11-3 23:18 Z
Setting aside consideration of spaces for the moment, I would certainly hope that one thing everyone her can agree on is that neither a degree sign standing alone, nor a letter standing alone (note that F is the symbol for farads, and C is the symbol for coulombs as well as the chemical element carbon), are acceptable for use for temperatures in Wikipedia. The symbol for degrees Celsius is "°C", both of them, unseparable. By extension, that applies to any other degrees of temperature as well. Gene Nygaard 01:55, 4 November 2005 (UTC)

## Number ranges

The text says:

• Sometimes numbers and dates are expressed in ranges, such as "4–7" for the numbers 4 through 7. Use an en dash for these when possible. It is often preferable to write this out (for example, "4 to 7" or "four through seven") to avoid confusion with "four minus seven". See Wikipedia:Manual of Style (dashes) for more information.

I think it is easy to miss the point about negative numbers. It could be made more explicit. Here is one possible revision:

• Dates ranges can be expressed as 1982 to 1988 or 1982 - 1988.
• Number ranges can be unambiguously expressed using to e.g. 4 to 7. A dash or hyphen can also be used if negative values or subtraction is not valid. [... possibly provide examples e.g. temperatures, tolerances etc ...]

Other suggestions welcome. I don't know whether this section belongs here or in the section on dashes. Perhaps these two pages should be brought together. What do others think? Bobblewik 16:23, 8 November 2005 (UTC)

Since an en-dash is signoficantly differnt from a hyphen or a minus sign/subtraction symbol, it can be used even if negatives or subtraction are possible, but it might in those casaes be better avoided. Also, "through" is slightly less ambigious than "to" IMO, as some people use "to" for "up to, but not including", whereas "through" i think always means "up to and including" Howw about:
• Sometimes numbers and dates are expressed in ranges, such as "4–7" for the numbers 4 through 7. When writing ranges in this form, use an en dash. However, it is often preferable to write this out (for example, "4 to 7" or "four through seven") to avoid confusion with "four minus seven", or with a list including negative numbers, such as "four, negative seven". See Wikipedia:Manual of Style (dashes) for more information.
Note that your version above lost the specific suggestion of an en-dash, as opposed to a hyphen. DES (talk) 16:37, 8 November 2005 (UTC)
Thanks for the response. I thought through was merely a synonym for to. What you say about inclusiveness is interesting. I would like it to be true. Can we confirm this?
The term through is alien to me because it is a variant of english that I don't use. I would have no personal problem with standardising on it, particularly if it has a less ambiguous meaning. But:
• I suspect it would be hard to get other non-US editors to use it because of the english variant.
• All editors seem to prefer the brevity of a dash. The use of to uses more space but it seems to be accepted sometimes. I wonder if many would accept the extra space used by through.
I did not remove the en-dash on principle. It was partly lazy drafting of the text and partly that I don't get it. I rarely notice the one or two pixel difference. If I do notice, it just seems as meaningful as the difference between km² and km2 ('''km& sup2;''' and '''km<sup>2</sup>'''). I don't know why people bother, but it does no harm. If people think it adds value, keep it in. Bobblewik 12:19, 9 November 2005 (UTC)
In many fonts the hyphen is much shorter than the en dash. The en dash in this case is a more widely supported substitute for the figure dash (U+2012). The figure dash, hyphen-minus (U+002D), and minus sign (U+2212) are supposed to occupy the same width as one numeric decimal place in vertical alignment, eg, in a table or complex equation. Michael Z. 2005-11-9 20:31 Z
• Note that the exixting Mos version endorses the use of either "to" or "through" it does not say that "through" is prefered. I have expereinced people findign "to" ambigious, but I can't say how widespread that view of things is. I find the en-dash noticably larger than the hyphen or minus sign (whn the special character for the latter is used) in most standard web fonts that i see. If you want to make the advide to use a word and not a dash stronger than my version, that might not be a bad idea, but we should still indicate that a dash is an acceptable usage except where it is likely to cause confusion, IMO. DES (talk) 23:01, 9 November 2005 (UTC)
It would make sense that an en dash could be used if and only if there is no possible confusion with the minus sign given the context. (I personally prefer it to be spelt out "to" or "through" to avoid symbol abbreviations but never mind that that's just me.) Don't know about the to versus through thing, I'd say it's arbitrary. Hyphens must be discouraged from number ranges, and editors allowed to edit articles accordingly.
In terms of linking this and the mos on dashes, as long as the two pages are consistent (last time I checked which was more than a month ago, they weren't) it should be okay, but (there's always a but) there's every chance someone will change one without changing the other. If someone wants to merge the two pages, or remove the duplicate on one, I'll be behind them, but I won't push for it too hard. Neonumbers 11:10, 17 November 2005 (UTC)
– − Okeh, fellahs. These are en dash and minus sign (I had to C&P, since my keyboard only has _- and ㅡ (Korean letter closest to em dash). Do you really mean to say you can see a difference? Here in edit mode I can see no difference; in preview one is a little lower but not any shorter or longer that I can see. Give the readers a break!
Yes, I can see a difference. I've been correcting dob / dod to meet style guidelines, and have been using the correct symbol in the process. Next time you post a message look below the "Save page", "Show preview" and "Show changes" buttons, and you will see links to several dozen symbols. The correct hyphen to use is available there. Just click on it to insert it into your text: "Ə ə – — …". It should be the first dash in the subset I just quoted (at least that's the one I've been using). --Dan East 13:57, 19 November 2005 (UTC)
That's exactly whence I got my two (where I mentioned C&P). Okeh, if You say You see a difference, but either it is too small for my eyes or computers don't register it (I am using public computers in Korea with various common browsers). Kdammers 00:06, 21 November 2005 (UTC)
Not seeing a difference, or a minimal difference, in edit mode is a reason not to convert &ndash; and &minus; and &mdash; to the equivalent characters on the edit page. Just watch for bots making that change. The difference, for most people, is clearer when the pages are viewed normally. Gene Nygaard 15:44, 26 November 2005 (UTC)
Like Bobblewik the use of through to indicate a range is alien to me. Unlike Bobblewik I'd not be comfortable with standardising on it. In no way does it seem less ambiguous to me. I never knew that it meant anything different to to. I'm not alone here. I'd generally use to for ranges inclusive and between exclusive of the starting and ending points. Jimp 27Nov05
Between is usually not exclusive of the endpoints either. And there is no normal distinction between to and through in this context; anything which needs to make it clear will generally say in some other way that it is exclusive of the endpoints. Gene Nygaard 15:44, 26 November 2005 (UTC)

I think that a hyphen or en dash should be acceptable in limited purposes, such as birth and death dates after a persons name, e.g. William Thomson (1824–1907), where I find it preferable to any spelled out word. I think it should be strongly discouraged in situations such as temperature readings where the numbers can be and often are negative, clearly subject to change by any editor. In other situations, I don't like it but could let it stand if it was already there. Gene Nygaard 15:44, 26 November 2005 (UTC)

I also don't think a hyphen should be used in "between...and" or "from...to" constructions. Use "between 35 and 45" not "between 35–45". Use "from 100 m to 200 m" not "from 100-200 m". Gene Nygaard 15:54, 26 November 2005 (UTC)

I guess you're right, Gene. Between can mean including the end point. Yes, the best way to make things clear is to explain them in words. As I say I'm not familiar enough with through used this way to comment on whether it is different from to but I had expected it was the same. Otherwise I second all that you write, Gene. Jimp 27Nov05