# Wikipedia talk:Manual of Style (dates and numbers)/Archive 25

 Archive 20 ← Archive 23 Archive 24 Archive 25 Archive 26 Archive 27 → Archive 30

## temperatures, spaces before or after degree symbol

Is there a policy about spaces before or after the degree symbol in temperatures? For example, 32°F (no space), 32° C, (space after degree symbol) 27 °C (space before degree symbol)? The main pages shows the second one for degrees, minutes, and seconds, so is that preferred for temerature too? Bubba73 02:45, July 25, 2005 (UTC)

ISO 31-0 clearly says "27 °C". Markus Kuhn 07:09, 26 July 2005 (UTC)
This was discussed somewhere up thread (was a consensus reached?). Anyway, there are no spaces in non-scientific American style, e.g., 60°F. British style and scientific style use a space or half-space between the value and the degree symbol, e.g., 20 °C. I would recommend no spaces for an American-based article and spaces for a technical or British-based article. —Wayward 08:16, July 25, 2005 (UTC)
Yes there is a policy, following what Wayward calls British or scientific style. Christoph Päper 08:55, 25 July 2005 (UTC)
It is, of course, NIST (the United States national standards laboratory) which expresses Wayward's "British style" rule more clearly than any other style guide. [1]
There are other related issues, too. Don't use the degree sign standing alone for temperature: not "33° to 35 °C" but "33 °C to 35 °C". Use either both words spelled out, or the entire symbol consisting of the degree sign and identifying letter or letters. Don't write "22° Celsius" or "22 degrees C". Gene Nygaard 13:26, 25 July 2005 (UTC)
"33 to 35 °C"? Just curious. Rl 14:24, 25 July 2005 (UTC)
Thank you everyone. I made some additions to Heat index and Dew point and I think there are the three different styles I mendtioned in those articles. Bubba73 15:19, July 25, 2005 (UTC)
PS - I'm going to consider Heat index and Dew point technical articles, unless someone thinks otherwise. Bubba73 16:08, July 25, 2005 (UTC)
PPS - I made that change. In two cases it made a line break between the degree symbol and C/F on my browser with my fonts, though. Bubba73 16:18, July 25, 2005 (UTC)
Use a non-break space, e.g., 70&nbsp;&deg;F = 70 °F. —Wayward 01:59, July 26, 2005 (UTC)
That does not prevent breaking after the degree sign, before the C/F as stated above.−Woodstone 02:50, July 26, 2005 (UTC)
Which browser do you use? Mine certainly do not insert line breaks for anything other than a space character. Markus Kuhn 07:09, 26 July 2005 (UTC)
It actually happens for me with IE6 in the mentioned article dew point (full screen 1024x1280, medium size font). I also tested separately. So inserting the "&nbsp;" actually increases the chances of an inappropriate line break. IE6 breaks only after the °-sign, not before! −Woodstone 07:20, July 26, 2005 (UTC)
I had noticed this some time ago, and hadn't even stopped to think that it might be browser-specific. BTW, the IE method does get a more acceptable break in latitude and longitude and other angles. Gene Nygaard 13:10, 26 July 2005 (UTC)
Which is probably why it does it that way. The solution is to use the Unicode characters for "degree Celsius" and "degree Fahrenheit", i.e. &#x2103; (℃) and &#x2109; (℉). Too bad for the Réaumur and others... Urhixidur 20:20, 2005 July 26 (UTC)

## Policy on "3rd", etc

Is there a policy on 3rd versus 3rd versus third? (If there is a policy, I couldn't find it.) I'm wondering about it in general and in connection with militry units, such as the 101st Airborne. To me, if you're going to use the numerals, the superscript looks much better. On the other hand, the Search needs to work on non-subscripted versions. Also, modern word processors such as Word Perfect and Microsoft Word have automatically converted these to superscript form for years. This must be considered the preferred form. Bubba73 16:52, July 24, 2005 (UTC)

I don't think we have a policy or need one. Maurreen 17:04, 24 July 2005 (UTC)
I'm unaware of any policy. Having said that, superscripts break line spacing (in some fonts, skins, browsers, etc.). That can be resolved by putting them in a <small> font (3rd), but that's a blatant violation of the KISS principle. Not something I'd start a holy war over, but that's one of the Word autofunctions I always turn off. Hajor 17:18, 24 July 2005 (UTC)
Symbolic forms of numbers are better for an international resource. For example, if your French is not so good, would you prefer:
• 3ème
• 3ème
• troisième
Of the two symbolic forms, I prefer the one without superscript to keep it simple and legible. Bobblewik 17:47, 24 July 2005 (UTC)
I hate those superscripts; I turn mine off in WordPerfect, too. But I'm not convinced we need a policy against them, and I wouldn't like to see a policy preferring them. How about "2nd" and "3rd" and "23rd" vs. "2d" and "3d" and "23d"? Gene Nygaard 19:49, 24 July 2005 (UTC)
Since Bubba73 mentioned military units, that was the prescribed form in U.S. Army Regulation 15 (AR-15), the Army style manual, when I was in the Army over 30 years ago, and I imagine it still is. But we have 2nd Infantry Division (United States), with a mixture of 2nd and 2d in the text of that Wikipedia article. The link to the opening page of the official website of this division also uses 2nd. But for the U.S. 3rd Infantry Division, its website is a haphazard mixture of spellings. See for example, this 3d ID page with "3D Infantry Division" in the logo at the top, two conflicting header lines for " 3d Military Police Battalion (Provisional)" and "3rd Military Police Battalion (Provisional)", the navigation bar at the left including "2d Brigade" but "3rd Brigade", "3d MP BN(P)" in text, "3D MILITARY POLICE BN" on the banner on their crest, and a phone number for "CDR, 293d MP CO". Gene Nygaard 20:07, 24 July 2005 (UTC)

"2d" is not only astonishingly non-standard (in normal English, at any rate), but ambiguous for anyone over a certain age in this country. The important point, surely, is that – aside from technical or naming conventions, as in the division names – we're writing English, and should use English words like "second" and "third". Some articles are already drifting towards a sort of pseudo-technical look that's both ugly and harder to read (see the slew of teeny-pop articles that attempt to ape music journalism ("her 3rd single debuted at #2, and she had 4 #6s, and 2 singles that hit the charts at #3", and so on). --Mel Etitis (Μελ Ετητης) 22:11, 24 July 2005 (UTC)

My feeling about that is that the the #-whatever is OK but the other numbers should definitely be spelled out. Bubba73 22:47, July 24, 2005 (UTC)
My understanding is that uses such as "3rd" are standard to designate military units, "third" is standard in other cases, and "3d" and "2d" are seldom used in civilian publications anymore. Maurreen 07:45, 25 July 2005 (UTC)

In most cases I reckon they'd be spelt out, second, third, etc., but if they must be used in numbers for whatever reason, I don't like superscript. It is meant to be convention to use superscript, but for us it ruins the line spacing (which, btw, was my reason against negative exponents, but that's irrelevant). If ordinals without superscript are already in common usage, then I say we may as well add it to the manual — that's what it's for: guideline that represents best and common practice. (No comments on military.) Neonumbers 11:27, 25 July 2005 (UTC)

## Fractions?

Should the guide mention any preference between using [itex], <sup>/<sub>, plain ASCII, or Unicode when writing out fractions? Examples:
 [itex] $\frac{123}{456}$ / 123/456 ASCII 123/456 Unicode ¹²³⁄₄₅₆
Unicode has 0x215F for the "1/" part (e.g. ⅟₄₅₆) but the example above had to mix superscripts (0x2070-0x2079), the fraction slash (0x2044), and the subscripts (0x2080-0x2089). There are also a slew of Unicode "standard fractions":
 half ½ 0x00BD thirds ⅓ ⅔ 0x2153-0x2154 quarters ¼ ¾ 0x00BC, 0x00BE fifths ⅕ ⅖ ⅗ ⅘ 0x2155-0x2158 sixths ⅙ ⅚ 0x2159-0x215A eighths ⅛ ⅜ ⅝ ⅞ 0x215B-0x215E
Urhixidur 17:32, 2005 July 29 (UTC)
If the fraction appears in a formula, especially if it is in a math article, then I think people are more likely to endorse the Meta:Help:Formula (your [itex]) approach. If it appears in prose, then probably the ASCII and some of the Unicode precomposed fractions.
However it should be noted that many fonts are based on either Windows Glyph List 4, which only supports ½ ¼ ¾ ⅛ ⅜ ⅝ ⅞ (no thirds, fifths or sixths), or the Adobe Glyph List, which only supports ½ ⅓ ⅔ ¼ ¾ ⅛ ⅜ ⅝ ⅞ (no fifths or sixths). Thus I would not recommend using the thirds, fifths or sixths characters. Both glyph lists support U+2044 (fraction slash), but not the super/subscript digits. Using the super/subscript digits is inappropriate anyway; The Unicode Standard section 6.2 says you are supposed to use regular digits with the fraction slash. — mjb 19:47, 29 July 2005 (UTC)
The "Unicode" method above uses characters that are deprecated when superscript and subscript markup are available. See Unicode in XML and other Markup Languages, Section 5.6 Superscripts and Subscripts. The same technical report recommends against precomposed fractions when alternatives exist: Section 5.4 Fractions. I think the style guide should recommend against these two methods, but otherwise leave the formatting up to the authors and editors: what works or looks best will depend on the article. Indefatigable 20:52, 29 July 2005 (UTC)
Note that the fraction slash &frasl; is different from the regular slash: 1/2 vs. 12 - Omegatron 21:08, July 29, 2005 (UTC)
It would be great if the Wikimedia software detected fractions by the presence of the &frasl; character and automatically put in superscripts and subscripts. This is how the Unicode Consortium intends the fraction slash to work, but I don't know of any browser or other software that does this yet. Anybody know where to submit a "request for feature"? Indefatigable 01:32, 30 July 2005 (UTC)
You can file a bug in bugzilla and pick severity "enhancement". Rl 07:35, 30 July 2005 (UTC)

## Unknown Death

What is the Manual of Style recommendation for handling a biographical article where the topic's date of death is known, but not date of birth? --SparqMan 09:01, 31 July 2005 (UTC)

According to the current MoS, nothing is listed for an unknown date of birth (see Wikipedia:Manual of Style (dates and numbers#Dates of birth and death). —Wayward 09:50, July 31, 2005 (UTC)
Clarification to my original post: A date of birth can't be given if it's unknown. Therefore, give what is known, e.g., John Doe (d. June 2 1765). —Wayward 10:03, July 31, 2005 (UTC)

When the date of death is known but the date of birth isn't, use "(died 2 June 1765)". Don't abbreviate "died". When death is unkown, it's "(born 2 June 1765)". (the heading clashes with the first post). Neonumbers 10:19, 31 July 2005 (UTC)

The standard usage in reference works, and generally for birth and death dates in parentheses, is to use "b." and "d."; why do you reject this? --Mel Etitis (Μελ Ετητης) 16:28, 31 July 2005 (UTC)
The MoS generally frowns on using abbreviations where they are not neccessary. See, for example the admonitions against i.e., e.g., etc. Writing out the words makes the content more accessible, especially for people who may not be familiar with English language conventions. Kaldari 17:58, 31 July 2005 (UTC)

In general I agree (though we're discussing the MoS, so we can't be limited unquesioningly by its current strictures). When there's a conflict with normal usage, though, it's worth considering whether the abbreviation isn't preferable. --Mel Etitis (Μελ Ετητης) 20:44, 31 July 2005 (UTC)

## Numbers written as words

I can't find any guideline on how (or whether) to hyphenate numbers when written out as words, such as forty-two. My instinct is to hyphenate it, but I have seen it not hyphenated on WP so I don't know if this is a US/UK difference. If it is to be hyphenated, what should I use? The hyphen-minus, or a unicode character? It seems odd that Hyphen#Hyphens_in_computing says "Usage of the hyphen-minus character is discouraged where possible, in favour of the specific hyphen character" yet this MoS says to use it. PhilHibbs | talk

I am not aware of specific guidance on numbers written out as words. However, there are guidelines that encompass the issue. My summary is: A hyphen is an optional tool to be used when it fulfils a purpose. If that purpose does not exist, then do not use a hyphen. For example, an author can use a hyphen for the purpose of linking number word pairs when the reader might reasonably have more than one choice of interpretation of the text.
For example:
• 'forty two sheep' (only one reasonable interpretation -> no hyphen)
• 'forty-two millimetre thick wires'
• 'forty two-millimetre thick wires'
I quote the references on my style talk page in the Hyphens section. Bobblewik  (talk) 12:25, 13 July 2005 (UTC)
Numbers twenty-one through ninety-nine are always hyphenated. And it's "forty-two-millimetre-thick wires," but 42 mm thick wires (abbreviations always open). —Wayward 12:44, July 13, 2005 (UTC)
So would anybody object to this being written into the MoS, or should it be left out as it does seem to be a standard rule of written English, rather than a matter of personal or national preference? (p.s. the Times's Style Guide is an interesting read, but I'm surprised that they don't clarify the posessive form of Jesus.) PhilHibbs | talk

I don't see any point in being prescriptive here. Incidentally, most prescriptive style guides say something completely different: namely to write out the numbers one to ten, but to write higher numbers in digits. So, for example, if you have five sheep and then get six more you have 11 sheep. If you then get nine more you get 20 sheep, but I don't see a need to prescribe this rule either, jguk 18:57, 13 July 2005 (UTC)

Yeah I was taught one through ten and 11–20. I'm sure in some articles it's useful to write numbers as numbers, too. I rarely see twenty-four written out. - Omegatron 19:48, July 13, 2005 (UTC)
About figures or words, the Chicago Manual of Style states the following:
"According to Press style, the following are spelled out in ordinary text:
• Whole numbers from one through ninety-nine
• Any of these followed by hundred, thousand, million, etc."
The exceptions are scientific writing (45 pounds, 3 cubic feet)—which would apply in our scientific articles—and clusters of numbers (e.g., their ages were 69, 64, 54, 47 and 35). They note that newspaper style, as well as that of some general publishers and scholarly journals, decrees that only the numbers from one through nine and such multiples as one hundred or nine thousand are to be spelled out. Sunray 20:19, July 13, 2005 (UTC)
The problem with the Chicago Manual of Style is that whilst it has quite a lot of respect in America, it has very little respect elsewhere in the English-speaking world. Returning to the point in hand, though, it is clear that it is not appropriate to enforce any one particular style, jguk 20:38, 13 July 2005 (UTC)
Since you have no respect for Chicago, perhaps you respect Oxford?
Oxford Style Manual (2003), 7.1.2: "In non-technical contexts, OUP style is to use words for numbers below 100. "
Oxford Style Manual (2003), 5.10.5: "Use hyphens in spelt-out numbers from twenty-one to ninety-nine (twenty-three, thirty-fourth, four hundred and sixty-eight, fifty-three thousand), and in fractions, unless the numerator and the denominator are already hyphenated (one-half, two-thirds, three thirty-seconds, four and five-eights)."
I'd say hyphenating numbers 21–99 is near universal. Can you cite a guide that recommends not hyphenating them?
Wayward 02:44, July 14, 2005 (UTC)
Not a rule, just a comment that specific proper names such as Twentynine Palms, California are based on early 20th century U.S. Post Office rules. Gene Nygaard 13:23, 14 July 2005 (UTC)

With the OUP "non-technical contexts" rule, we should also include a technical contexts rule, such as this one from NIST: [2]

"This Guide takes the position that the key elements of a scientific or technical paper, particularly the results of measurements and the values of quantities that influence the measurements, should be presented in a way that is as independent of language as possible. This will allow the paper to be understood by as broad an audience as possible, including readers with limited knowledge of English. Thus, to promote the comprehension of quantitative information in general and its broad understandability in particular, values of quantities should be expressed in acceptable units using
• "the Arabic symbols for numbers, that is, the Arabic numerals, not the spelled-out names of the Arabic numerals; and
• "the symbols for the units, not the spelled-out names of the units."

I'd also say that much of the reasoning ("as independent of language as possible") would apply even in more general, less technical contexts here on Wikipedia, especially the English Wikipedia. Gene Nygaard 13:23, 14 July 2005 (UTC)

But if people come here for a guide because they don't know what to do otherwise, the least we could do is have some advice for them.
There'll be some places where it's better to use words, and we should have a guide for those places, even if they are few. I say numbers from 20 to 99, when written out in words, are hyphenated. (I normally put eleven to twenty in words and 21-99 in numbers, but I think I'm the odd one out)
Which style we choose here isn't that important, but the manual of style exists to guide editors stuck over style issues, and it should live up to its job, so something about this should be on the manual. Neonumbers 01:47, 14 July 2005 (UTC)
Taken together, Oxford and Chicago style manuals are considered absolutely authoritative. I move we standardize on the above and spell it out for our users in the Wikipedia Manual of Style. Sunray 05:34, July 14, 2005 (UTC)
We should not over-specify the English language. There will be an obvious exception to this rule, but literal-minded editors will point to the guideline as gospel and start revert wars over it. Michael Z. 2005-07-14 06:30 Z
MoS is not enforced, it's guideline only. I would be astonished to see a revert war over a subject like this. PhilHibbs | talk 12:01, 15 July 2005 (UTC)
It will definitely happen. - Omegatron 15:22, July 15, 2005 (UTC)

I appreciate Sunray will be unfamiliar with British usage, so I should point out that he is quite mistaken to say that Oxford and Chicago style manuals are considered absolutely authoritative. The Oxford style manual does not represent the most common form of the language used in Britain (let alone the rest of the English-speaking world). For instance, very few British publications other than those produced by the OUP use -ize rather than -ise (similarly for writers). The mandatory Oxford comma is rare outside OUP's publications - and these are just two examples of Oxford house style which you will not see common usage outside the OUP, jguk 19:14, 14 July 2005 (UTC)

Style guides often do not reflect the most common usage—it is not their purpose. Style guides present a set of commonly acknowledged practices that are accepted in their respective academic fields. —Wayward 03:06, July 15, 2005 (UTC)
I've never seen any respectable source that doesn't hyphenate numbers; I can't see any mention of one above (but there's a lot there, and I might have missed it). --Mel Etitis (Μελ Ετητης) 22:58, 14 July 2005 (UTC)
General British usage (unfamiliar with which as I may be) is surely not our main concern in an encyclopedia. We tend to follow academic standards and style for our articles (albeit written as readably as possibly), rather than newspaper or magazine style guides. Perhaps Jguk could refer us to an appropriate style guide for academic British usage if the Oxford Style Manual is insufficient as a resource. For North American usage, the Chicago Manual of Style is considered authoritative, as the Wikipedia Manual of Style makes clear here. Sunray 02:37, July 15, 2005 (UTC)
Ditto above. Neonumbers 12:58, 15 July 2005 (UTC)
I agree. This isn't a magazine. - Omegatron 15:22, July 15, 2005 (UTC)
True, but we aren't an academic text either and certainly shouldn't be using an academic style guide. IMO we should imagine our readers have the knowledge and reading skills of a reader of a broadsheet newspaper. I don't mean we should copy newspaper style (as generally we shouldn't), but in choice of words and language style, we shouldn't be too different. I don't think there's one single British English style guide that can be considered wholly authoritative. As a useful guide though, if you stick to what The Times has, you won't go far wrong, jguk 06:22, 16 July 2005 (UTC)
I strongly disagree with basing the Wiki MOS on any newspaper house style. Newspapers cram much more text into each square inch than do other publications, and therefore they are extremely concerned with saving space (space + ads = \$). A few examples from The Times style guide: No degree symbol in temperatures, 15C (-2F); no spaces with other symbols/abbreviations, 5ft 7in (1.7m), 1,343m (4,406ft); no space with AD/BC, 350BC; no space or points in times, 6pm; no points with Latin abbreviations, ie, eg, etc; days abbreviated, Mon, Tues, etc.
Getting back to the original posters question about hyphens in spelt out numbers, The Times guide uses a hyphen in "twenty-three" here under "hour and a half, an." —Wayward 09:46, July 17, 2005 (UTC)
I wasn't intending to suggest we adopt The Times style guide wholesale, merely that it's a reasonably fair indication of the most common usage in modern British English - though there are a few exceptions. In the examples you give, if I was writing I'd probably put a degree symbol in temperatures, have no space in 5ft 7in (though I'd probably write it 5'7''), have a space between AD and BC and the year, write 6pm and omit points after abbreviations (ie, eg, etc). I think the only real difference you'd see with others is that some do continue to put points in abbreviations.
As a general point, the tendency in the UK has over the last 60 or so years been to reduce the amount of punctuation used. This tendency has also been adopted throughout the English-speaking world, except in North America. Ironically, if you look at a nineteenth century British English text, you'll see it punctuated in the same way as a twenty-first century American English text is punctuated! The difference between twenty-first century British English punctuation and American English punctuation can be very marked indeed, jguk 10:23, 17 July 2005 (UTC)

This has been an interesting discussion and I think we have consensus (though not unanimity) on a number of points. There is obviously a lack of clear guidelines on numbers as words in the Wikipedia Manual of Style. The commentary here seems to suggest that we should have guidelines to assist editors who have questions. Most who have commented agree that:

• Double numbers from one to ninety-nine should be hyphenated.
• Numbers from one to ninety-nine should generally be written.
• An exception to this is scientific writing, in which Arabic numerals should be used.

Have I got that right? If so, do you agree that we should add a few sentences to the MoS along these lines? I assume we all take the point that the MoS is a guideline, not a law. Sunray 16:02, July 17, 2005 (UTC)

Er, no. I very, very strongly disagree that the MOS should recommend spelling out numbers from 1-99. And I don't see much of a consensus on that point above. I'd be OK with always spelling out 1-10, but above that it should depend on the context and the preferences of the editor. I've no problem with the hyphenation rule. olderwiser 16:24, July 17, 2005 (UTC)
Here's my tally:
• Hyphenization: For: Wayword, PhilHibbs, MelEtitis, Sunray. Against: None. Optional: Bobblewik.
• Written words in text: For: Wayward, Omegitron, Sunray. Against: Jguk, Bknorad. Either: Neonumbers.
• Numbers in scientific contexts: For: Gene Nygaard, Sunray. Against: Nil
Consensus on talk pages is usually a two thirds majority. So we have consensus on hyphenization, but with Bkonrad's nay we do not have consensus on writing out words in text. Sunray 19:32, July 17, 2005 (UTC)
That tally does not represent my position. My position is:
• Hyphenization: Against. Bobblewik. As I said above, the hyphen is only a tool to resolve ambiguity. Otherwise it adds no value. If there is only one reasonable interpretation there is no benefit to a hyphen. See the quotes from styleguides on my page.
• Written words in text: Against. Bobblewik. Digits yes please, they are easy to read and translate. Words no thank you.
• Numbers in scientific contexts: For. Bobblewik. Digits yes please, in all contexts.
I am quite comfortable with the situation as it is. The Manual of Style should only contain advice where there is: (a) confusion that causes a problem or (b) conflict that causes a problem. I don't see either in this case, we merely had one question.
Bobblewik  (talk) 20:23, 17 July 2005 (UTC)
I have no preference on whether to spell out numbers or not; however, I do feel that numbers 21–99 when spelled out should be hyphenated, as it is common practice. —Wayward 00:46, July 18, 2005 (UTC)
I'm not sure we need to add anything about hyphenation. Isn't that at least close to being universal? Also, I prefer to use numerals for numbers of 10 or more. Maybe we should just leave it to the users' choice. Maurreen 19:56, 17 July 2005 (UTC)
None of the style guides you cite on your page say to leave out the hyphens in numbers not divisible by ten from twenty-one to ninety-nine, Bobblewik. Gene Nygaard 21:02, 17 July 2005 (UTC)
The quotes from styleguides describe the use of the hyphen as a tool to resolve ambiguity. The phrase forty two sheep has no reasonable ambiguity so a hyphen adds no value. Some people like to use hyphens in such cases. That is their choice but I don't want to be told to do it. According to Fowler: If you take the hyphen seriously, you will surely go mad. Bobblewik  (talk) 21:41, 17 July 2005 (UTC)
The fact that this encyclopedia is edited by many editors does not take away from the fact that it is one encyclopedia. One encyclopedia, one style, with precious few exceptions. This is guideline, not law, not policy - even if this is a small issue, people who need guidelines on it shoudl get guidelines on it. If people come around inserting hyphens, then all they're doing is bringing that article in like with the rest of the encyclopedia (assuming that the guideline is to insert hyphens). I think hyphens should be there - but so long as there is a guideline, it doesn't really matter which way we go.
I'd just like to correct my view in Sunray's tally, I'm for all three points: hyphens, word and digits in science. Neonumbers 01:38, 18 July 2005 (UTC)

## Abbreviations for timings

A problem has cropped up regarding how timings (for albums, tracks, and singles) should be expressed. At the moment there are various styles, the main three being:

The first is used in the Template:Album infobox; all three are used by various editors in various articles.

The official SI abbreviations are "min" and "s" and they are never written with fullstop or plural s (see also ISO 31-0).

The first seems to me to be odd: the Wikilinking is surely unnecessary, the abbreviations (lacking either plural or full stops) are incorrect, and I've rarely if ever seen it used elsewhere.

The second is often used on album covers, etc., but it's the same form as used in Wikipedia for times of day, and thus could conceivably cause confusion.

The last is pretty standard, being used not only on many album covers, etc., but in other contexts too.

OK, my preference is pretty clear — but I'd rather that one of them were chosen as the recommended style even if it weren't my preference than that we continue with the free-for-all that we have now. I was certain that I'd seen something on this in the MoS, but I can't find it; is there anything? If there isn't, shouldn't there be? --Mel Etitis (Μελ Ετητης) 18:25, 17 July 2005 (UTC)

The modern rules for abbreviations for units of measure are never change them in plural, and never use fullstops. The modern symbol for seconds, of course, is "s" rather than "sec".
The second form, used on album covers, is reasonable. There is no real difference between this and times of day (it's like the difference between a temperature reading and a temperature interval, which can both be expressed by using °F, etc.), and little likelihood of confusion in any case. I say use that one.
That prime and double prime format is common for minutes and seconds of arc (as fractions of a degree, but not of an hour), uncommon and unusual for minutes and seconds of time, and IIRC improper according to some style guides. For time (and also for hour angles), you are much more likely to see formats such as 3h 18m 25s. Gene Nygaard 19:27, 17 July 2005 (UTC)
See the discussion of abbreviations for timings that took place last year in WikiProject Albums. Bobblewik  (talk) 20:32, 17 July 2005 (UTC)

Thanks for the link. The trouble for me is (as with Gene Nygaard's message above) the discussion is full of people claiming that certain forms are unfamiliar and confusing when they're in very common use. The main one here is: (mm' ss"). Just plucking CDs from my shelf at random, they're almost equally divided between those using mm:ss, those using mm' ss", and those offering nothing (none use the peculiar "mm min ss sec" — and even without the unnecessary links, that still looks odd to me). With regard to G.N.'s claim, by the way, where do we find "the modern rules for abbreviations"?

In my opinion, the examples at the top of this section are listed in descending order of clarity. Of course, there's nothing wrong with adding periods or spelling out the words either. Maurreen 22:22, 17 July 2005 (UTC)
I don't think SI units are an issue here. This isn't science. I think either of the first two will do - doesn't matter which. The third is a bit, well, less common in my opinion. My personal opinion with dots for abbrevs here (and only here) is don't, but I wont' complain if everyone else thinks otherwise. Neonumbers 01:53, 18 July 2005 (UTC)

I'd not mind the first one so much without the links (which are included in the Wikipedia:WikiProject Albums version); the last one is in fact very common on CDs (as are 45.34, 45'34, and 45:34 — I've never seen the first one on a CD).