Wikipedia talk:Manual of Style/Dates and numbers

Guidance needed

It's easy to edit the text to include {{xt}}, but I'm having a bit of trouble working out where, and would appreciate some confirmation before I start:

• Change each phrase in italics at Numbers as figures or words and Typography?
• Change the examples in Uncertainty?
• Change the "double quoted text" in Percentages? What about the examples like "71%" which are not in quotes?
• What about the quoted text in Avoiding ambiguities?

I cannot see any other quoted text that needs conversion, although there are probably some more phrases in italics or just plain text that I'll notice later. Johnuniq (talk) 11:58, 11 September 2009 (UTC)

• This is why this was so time-consuming for me. There is some gray area as to what text should be given the “xt-treatment”. I suggest you approach the issue this way: The name “xt” stands for “example text.” So, generally speaking, if the verbiage you are going to convert is preceded with language such as the following…

1. Editors should write…
2. …or write this…
3. …but not this…
4. …such as…
5. …for example,

…and the verbiage that follows clearly can be categorized as “example text”, then give it the {xt} treatment.

The whole purpose of {xt} was to solve the problem of what to do when we were giving examples of what to do with quotes, such as “When giving a quote that itself contains quoted text, change the double-quote to a single-quote as follows…” Thus, the entire bit of example text that followed those words couldn’t itself be set off within quotes without serious limitations and complications. The alternative, of course, was to use italic text to set off example text, but then that screwed up examples where we touched upon how to properly use italics. MOSNUM previously used both techniques to avoid this limitation. So…

We have {xt} to give us complete freedom to no longer need either italics nor quotes to set off example text. Once we are using {xt} for those, we should finish the job and use it for all instances of example text. I know there are some gray areas where what can be considered “example” text isn’t entirely clear. However, I’m sure you will find the proper answer in each case. I really appreciate, Johnuniq, that you 1) stepped up to the plate and volunteered to perform this service, and 2) take your responsibilities so seriously that you saw fit to check in and clarify the scope and definition of the task, as you did above.

I suggest that you not worry too much about goofing. So long as you break up your edits into small, clear, logical sections and frequently post, it should be easy for you or someone other editor to go back and just hit [undo] in the history and keep the reverted text (and wasted effort) at a minimum.

In advance, thanks. Greg L (talk) 19:28, 11 September 2009 (UTC)

P.S. I use {xt} when quoting editors because debate here on WT:MOSNUM frequently contains quotes and type-style changes. So the use of {xt} avoids my having to switch back and forth from using italics and quotes when quoting other editors. It also more distinctly sets off the quoted passage. For MOSNUM itself, {xt} is used as an alternative to all other ways of setting off example text; namely, quotes, italics, and bolding. So, to (finally) answer a question you posed above, here is how the following text currently on MOSNUM would be addressed:

There is no such ambiguity with recurring events, such as January 1 is New Year's Day.

Greg L (talk) 20:04, 11 September 2009 (UTC)

Thanks for info. I made quite a lot of changes and will wait at least 24 hours before doing more (I'm up to "Scientific notation, engineering notation, and uncertainty"). In the line starting "The ordinal suffix", I used xt, but (owing to the font), it is not entirely successful in appearance. I tried xt for the italics at "Non-base-10 notations" but it looks bad with xt, so I did not use it. Johnuniq (talk) 04:32, 12 September 2009 (UTC)

• I think your judgement was spot-on on how to evaluate the suitability of {xt} for given pieces of example text. Indeed, the Georgian font that {xt} uses treats its numbers with a, uhm… stylistic flourish (abcdefg - 1627384950). Georgian was chosen because it is found on Mac, Windows, and Linux with native installations—or, at least, exceedingly common ones. Perhaps one day, we can again take a look at using some other font. Times proved too problematic because it is smaller than its same-size, sans-serif counterparts. Giving Times a proportional percentage boost in size (like 107.5%) proved unsatisfactory for use across widely different platforms.

  I am really impressed with your conservative, stepped approach to improving MOSNUM. Your style should be widely emulated. Greg L (talk) 18:09, 12 September 2009 (UTC)

P.S. I just realized that if you edited numeric examples so their digits, when they are used in significands, superscripts, and subscripts, are segregated into—and chosen from pallets of—three families: 012, and 34579, and 68, then that ought to fix the appearance problem you are wrestling with. For instance: The chance that 68 editors on Wikipedia will have 100% agreement on any one issue on WT:MOSNUM is 9.7435×10⁻²¹. Greg L (talk) 18:51, 12 September 2009 (UTC)

Good idea for most examples; but in one example the fact that 1, 2, 3, and 4 are aligned together is the point. It is "the Unicode characters ² and ³ ... are not aligned with superscript characters (e.g., x¹x²x³x⁴ vs. x¹x²x³x⁴)". I'd fall back on not using {{xt}} on that example, or on using a kluge such as "the Unicode characters ² and ³ ... are not aligned with superscript characters (e.g., x¹x²x³x⁴ vs. x¹x²x³x⁴)". (If it weren't the case that so many people's default serif font is Times which has a ridiculously tiny x-height, I'd just propose that Xt just specify "serif".) --___A. di M. 19:19, 12 September 2009 (UTC)

• As the saying goes: If anything works in this world, thank an engineer.

  I see you coded the good-looking one as follows:

  {{xt|1=<span style="font-family: sans">''x''<sup>1</sup>''x''²''x''³''x''<sup>4</sup></span>}} vs. {{xt|1=<span style="font-family: sans">''x''<sup>1</sup>''x''<sup>2</sup>''x''<sup>3</sup>''x''<sup>4</sup></span>}}.

  Indeed, the example text appears small on my Mac because it uses Times, but it still looks nice, and—indeed—that is quite the kluge to code.

  So let me try the {{xt}} treatment: x¹x²x³x⁴ vs. x¹x²x³x⁴. I see. A big, stupid superscripted “3” and “4” because one must do a progression, where choosing from a segregated pallet isn’t possible.

  However, this is an situation where the example text is sufficiently clear from context since it is numerics surrounded by words. Greg L (talk) 19:35, 12 September 2009 (UTC)

• A. di M.: What do you think of {xtc} where the only change is to the custom green color? It would be used where you have to deal with alphanumeric formulas. As regards color blindness, it would be merely be an assistive technology that facilitates the distinction of what is example text and what isn’t. So I don’t see adding green as being a problem, such as Wikipedia’s practice of using a dark blue to indicate there is a link that can be clicked. And the prudent use of {xtc} on alphanumeric formulas on MOSNUM (and similar WP: uses) would be a lightyear from the same class as using color alone to differentiate important distinctions, like “this is good” but “this is bad”. Greg L (talk) 19:54, 12 September 2009 (UTC)

Update

My edit finished converting all "quoted examples" and italic examples to use {{xt}}. It turned out that there were very few changes in that last edit; I included changing "US" to "US gallon" at one point where it looked like an oversight.

In the line beginning "Use nautical mile or statute mile" (and some other places) I used xt for the examples which included links, so they are blue and look a tiny bit strange (but I think ok). I did not touch the examples in the "Quantities of bytes and bits" section because 1000³ renders unsatisfactorily as 1000³ and the 3 can't really be changed to follow Greg's above suggestion. The idea of an xtc template sounds useful for rendering the numeric examples mentioned above, and it looks very simple to do for a trial. Johnuniq (talk) 09:26, 14 September 2009 (UTC)

• Thank you so much (again) for stepping up to the plate and doing such a meticulous job. If it rocks your boat, you can give math expressions like 1000³ an “xt-treatment” that looks like this: 1000³. Here’s how to do a really useful trial until an {xtc} template comes along…
  • This is link-blue, which is coded <font color="#002BB8">fake link</font color>.
  • This is “xt”-green, which is coded <font color="#006F00">example formula without the accompanying Georgia font</font color>.

It would be ultra-easy when an {xtc} template becomes available to do a search & replace in a word processor to change occurrences of <font color="#006F00"> with {{xtc|. You know the drill; same on the back end.

Greg L (talk) 02:43, 15 September 2009 (UTC)

Hmm. Why not create Template:xtc? I'll give that a go if you like. Johnuniq (talk) 02:56, 15 September 2009 (UTC)

• I don’t know how to make templates. I usually look towards A. di M. for these sort of things. It seems his take is a good litmus test for when is the proper moment to make a template.

  My sense is that using <font color="#002BB8"> to fix the last few stragglers will be the perfect demonstration for a “trial,” as you say. It’s also been my observation in the past that having code like that in MOSNUM, if the practice sticks, is just the very sort of thing that prompts someone to make a template so they can get rid of the code. By dipping our toes into this with a color-call, we can all look at the result and see if it serves a good and valuable end.

  I know that my use of CSS spans when making really nice-looking scientific notation (like 9.743534579(35)×10⁻²¹ kg) precipitated the effort to make templates to accomplish the same end and that resulted in the {{xt}} template. People just couldn’t stand looking at my hand-written code for scientific notation, which looked like 6.022<span style="margin-left:0.25em">141<span style="margin-left:0.2em">79(30)</span></span><span style="margin-left:0.25em">×</span><span style="margin-left:0.15em">10</span><sup>–21</sup> kg. ;-) Greg L (talk) 19:29, 15 September 2009 (UTC)

I used Greg's font color="#006F00" in Quantities of bytes and bits. It looks good to me (although I now think I should have added a greeen period to each of the three "Acceptable examples include"). Following is what Template:xtc should be (needs documentation):

{{#ifeq:{{NAMESPACE}}|{{ns:0}}|{{FormattingError|Template:xtc is only for examples
of style and formatting. Do not use it in actual articles.}}|<span class="example"style=
"color: #006F00;">{{{1|}}}</span>}}<noinclude>
{{Documentation}}
<!-- If/when the <samp> element is included in WikiText, replace span with samp,
which would be more appropriate. -->
</noinclude>

Hmmm, interesting off-topic: Why does {{xt}} fail when used like this abc{{xt|font color="#006F00"}}def which renders as abcExample textdef? Johnuniq (talk) 04:53, 16 September 2009 (UTC)

Because when it sees an equal sign, it believes that it's a named parameter font color whose value is "#006F00". You have to use 1= to specify the value of the first unnamed parameter when it itself contains equals signs (font color="#006F00"). --___A. di M. 14:37, 16 September 2009 (UTC)

• I think Quantities of bytes and bits looks great. Thanks, Johnuniq. Having an equal sign anywhere within the {xt} template breaks it. There are at least three ways to work around this, depending on whether the = sign is functional or only for display. For instance…

• {{xt|1=''M''<span style="margin-left:0.15em"><sup>2</sup></span>}} → M² (one must use the “1=” technique because the = sign is functional in a CSS span… the span here is just an example that came to mind.)
• {{xt|2 × 3 {{=}} 6}} → 2 × 3 = 6 ( {{=}} (template replacement of the = as it is only display)
• {{xt|2 × 3 = 6}} → 2 × 3 = 6 (character reference to replace the = as it is only display)

Greg L (talk) 17:40, 16 September 2009 (UTC)

Thanks both. This is a classic case of induced stupidity because I have done a fair bit of work with templates elsewhere and it was the html (that is somewhat foreign to me) that made me overlook the blunder. I'm a bit hesitant to create xtc after that, but I'll do it if wanted and A. di M. doesn't feel like it. I don't think it is worth putting any more green manually into MOSNUM, so let's wait for xtc. Johnuniq (talk) 04:36, 17 September 2009 (UTC)

• Were it me, I’d sit back and see if the #006F00-made examples are met with bored acceptance (or interested acceptance). After maybe a month, it would, IMO, look more *formal* and cleaner to create an {xtc} template; that will make the examples look more like the etherial Wikipedia gods ordained that they be there. Greg L (talk) 00:26, 18 September 2009 (UTC)

• Problem: this edit replaced the green font color with {{xt}}. I might restore my version, but this time put a link to this discussion in the edit summary (or please do it yourself if you feel so inclined). Johnuniq (talk) 08:02, 18 September 2009 (UTC)

• While it might be unsuitable for a fine glossy-paged coffee-table book, having the bytes ‘n’ bits examples in {xt} doesn’t appear to be an abomination unto the eyes of the typography gods to me. More to the point, the {xt}-treatment doesn’t, IMO, get in the way or distract from the message. I think the whole page looks much better as a result of your exceedingly professional work and collaborative style, Johnuniq. Greg L (talk) 21:35, 21 September 2009 (UTC)

"0.57 second" or "0.57 seconds"?

What is the correct grammatical number for nouns preceded by number whose magnitude is strictly less than 1? It appears to me that the singular is used in American English and the plural in British English, but I've never seen a source explicitly dealing with this. ___A. di M. 23:33, 12 October 2009 (UTC)

In English there are few absolute rules. That said, to me as an American, "0.57 second" looks more "correct", but "0.57 seconds" looks and sounds more "natural". I would say that neither form is clearly incorrect, so you can use either form. Marco polo (talk) 00:16, 13 October 2009 (UTC)

I don't think the terminology is related to the magnitude of the number. E.g., "0 seconds" is usually written in the plural, isn't it? I think unit terms are usually count nouns, so the only time you use the singular is for a value of 1. But that's just my interpretation. You might want to look through the SI documents to see if there's an established international standard. Another option would be using "57 centiseconds" or "570 milliseconds", but I doubt that's what you want. :) Indeterminate (talk) 01:56, 13 October 2009 (UTC)

If there's a particular place where this is causing difficulty, perhaps someone here could suggest a sound turnaround (for example, and depending on context, 57 hundredths of a second, 0.57 sec. or 0.57 sec). And why is this being forked or duplicated to Wikipedia:Reference desk/Language?

¶ As for the style, even when it's something concrete (or that could be concrete) like a penny, a pound or a dollar, English-speakers always pluralize it in speech when giving decimal fractions, but not when giving normal fractions. Half a dollar, five-sixths of a shilling, and 25 pence (or cents), but 0.25 pounds (or dollars). I'd stick to idiomatic English that will not make Anglophone readers hesitate, and 0.37 second will just look stiff and weird, or hyper-grammatical. 0.37 seconds may not be technically correct according to the style-books, but it won't look wrong to 95%-plus of readers. They just won't notice it the way they might notice (and wonder at) 0.37 second. —— Shakescene (talk) 04:58, 13 October 2009 (UTC)

I've had to make that choice several times in the last month or so, for example in kilometre (is it 0.6214 miles or 0.6214 mile?). As for me, the plural sounds more "natural" (but I guess that's just because my native language, Italian, uses it), but I've seen some sources using the singular for such numbers (e.g. "1.782×10⁻²⁷ gram", "10⁻¹² second", "10⁻⁷ metre" in The Feynman Lectures on Physics), and the only other thing these sources appear to have in common is that they're written in American English, so I guessed that's an ENGVAR issue. (Also, on the reference desk I just intended to provide a pointer to here, because I thought the people hanging around there were the ones most likely to have clearer ideas about that; too bad someone replied directly there. Also, I shouldn't have used prescriptivist words (What is the correct grammatical number for), but rather descriptivist ones (Which grammatical number is used for); so I clarify the question, "Is it the case that "0.57 second" is more common in American English and "0.57 seconds" is in British English? Does anyone know about a source dealing with this matter?" ___A. di M. 10:15, 13 October 2009 (UTC)

Here's a text:

Names like 'metre' or coulomb' are the names of standard quantities and are used in sentences like ordinary nouns. We write 'twenty coulombs' or '20 coulombs', 'one tenth of a coulomb' or '0.1 coulomb'.

• Harrison, R.D., ed. (1972). "SI units". Book of Data: Chemistry, Physical Science, Physics. Harlow: Longman for the Nuffield Foundation. p. 4. ISBN 0 582 82672 1.

which in turn cites:

• McGlashan, M.L. (1968). Physico-chemical quantities and units. Monographs for Teachers No. 15. Royal Institute of Chemistry.

--Redrose64 (talk) 10:44, 13 October 2009 (UTC)

• I for one have never seen the singular appended to anything but the unit value (that is, one second). The reason for this is rooted in the meaning of the word "singular" – strictly speaking, something less than one is not singular. So it makes sense. —Anonymous Dissident^Talk 05:23, 13 October 2009 (UTC)
  • If it's 1.00000001 seconds, we use the plural because it's more than 1 second; but we also do the same when it's 0.99999999999 seconds when it's less than 1 second; I even write 1.0 seconds or 1.0000 seconds, rather than 1.0000 second. —— Shakescene (talk) 20:36, 13 October 2009 (UTC)
    • That makes good sense, because it's contextually clear with "1.0 seconds" that sub-integral numbers are being analysed. —Anonymous Dissident^Talk 00:57, 15 October 2009 (UTC)

OK, so it's not an ENGVAR issue; it's more an issue of grammarians recommending something without ever checking what native English speakers actually do, like "don't split infinitives" or "don't end sentences with prepositions". (OTOH, I think Feynman isn't the kind of guy who gives a damn about what grammarians say without ever checking what native English speakers actually do, and I guess his editor Leighton isn't, either. So, maybe there are speakers who do find it "natural" to use the singular in these cases.) ___A. di M. 10:31, 14 October 2009 (UTC)

It seems to me that it might be a prescriptivist thing. Pluralising anything but ±1 doesn't seem right to me (even "e^2πi second" seems a little awkward). I'm guessing that the prescriptivists may have thought that "0.5 second" is either nothing but a short form of "0.5 of a second" or a different way of writing "1⁄2 (a) second". English is not maths. JIMp _talk·cont 12:35, 15 October 2009 (UTC)

Well, the way that we were taught at school (in England) was basically: "half of a second" should be written "0.5 second", not "0.5 seconds", because "[of] a second" is singular. --Redrose64 (talk) 13:03, 15 October 2009 (UTC)

The style guides don't really seem to address the issue, but the most common rule seems to be: if a value is greater than 1, it should be plural and if it is 1 or less it should be singular. E.g. 1.01 seconds, 1 second, 0.9 second, 0.1 second. I think the basic idea is that 0.9 second is short for 0.9 of a second. However, for values of zero or less they seem to go back to being plural again: 0 seconds, -1 seconds. SI abbreviations avoid the issue by never using plurals. RockyMtnGuy (talk) 17:38, 15 October 2009 (UTC)

Apply WP:ENGVAR and common sense. The MOS is already overly large and increasingly a barrier rather than an aid - let's not seek to inflate it further where existing consensus (ENGVAR) already exists. --Dweller (talk) 06:06, 16 October 2009 (UTC)

Yeah, but my point was that not everybody knows which national variety of English uses which grammatical number in these cases. But now I see that nobody does, so I agree we'd better keep it off the MoS. ___A. di M. 13:13, 21 October 2009 (UTC)

Wtmitchell posted the following in Talk:Grammatical_number#.220.57_second.22_or_.220.57_seconds.22.3F:

The question, "What is the correct grammatical number for nouns preceded by number whose magnitude is strictly less than 1?" was recently asked at WT:MOSNUM#"0.57 second" or "0.57 seconds"?. The answer so far arrived at there seems to be "Nobody knows". The question doesn't seem to be addressed in this article. Perhaps it should be. Wtmitchell (talk) (earlier Boracay Bill) 04:49, 22 October 2009 (UTC)

To which I replied:

The answer varies from language to language even for languages that share a similar system of grammatical number.

There are other questions that also arise about how to match a grammatical number with a numeral that is not a positive whole number.

Even if the language has only "singular" and "plural" grammatical numbers, a question arises whenever the numeral is less than one (including zero or negative as well as fractional), or more than one but less than two (a mixed-fraction numeral).

For languages with a dual number distinct from singular and from plural, the question arises not only of what to do with 0.5, 1.5, 0.8, 1.2, etc., but also 1.8, 2.5, 2.2, 2.8 etc.

For English the answer appears to be usually "if it's not 'one', use a plural". For French, the answer appears to be different. I know this has been stated in print and in an online-searchable source, but I don't remember what the source was.

So perhaps the answer should be given in each specific language's section on grammatical number, and this cross-linguistic article on "Grammatical Number" should only mention that such sections contain such answers.

In English it may depend on how the fraction-numeral is written and pronounced. We might say "half a second" but "0.5 seconds"; "fifty-seven hundredths of a second" but "0.57 seconds".

It is less clear how grammatical number should match up with a mixed-fraction numeral greater than 1 but less than 2. We would probably say "1.5 seconds", but would we say "one-and-a-half second" or "one-and-a-half seconds"? I think we would probably say "one-and-a-half seconds" but "a second and a half".

I apologetically admit that the above doesn't satisfy the "verifiability" nor the "no original research" requirements of Wikipedia.

You say "The question doesn't seem to be addressed in this article. Perhaps it should be." I encourage you to do so. (Is that being WP:BOLD ?)

--Eldin raigmore (talk) 21:49, 22 October 2009 (UTC)

Request to unprotect Wikipedia:Manual of Style (dates and numbers)/Datestempprotectedsection

Really, there's no reason to leave it protected anymore. Please comment at the relevant request section. Dabomb87 (talk) 22:36, 16 October 2009 (UTC)

It's been merged back. Dabomb87 (talk) 00:06, 23 October 2009 (UTC)

Infoboxes

Could someone please enlighten me on what formatting to use when editing music articles infoboxes? 2005-Present? Or Since 2005? Or are both acceptable Thanks! GaudiumInVeritate (talk) 16:33, 19 October 2009 (UTC)

We had a discussion at Talk:Simple Plan. I commented on WP:OTHERDATE. No dog in the fight, just one of the editors on the Simple Plan article.--Wehwalt (talk) 16:58, 19 October 2009 (UTC)

I guess everyone else is confused too. GaudiumInVeritate (talk) 22:19, 19 October 2009 (UTC)

Non-breaking spaces

The NBSP guidelines make sense for the most part, but I'm a little concerned by them. Some articles are now getting tons of   tags added into them, especially those related to measurement... and even in some other articles, like in the "official breed standard" section edited in this diff by my AWB-bot. My feeling is that   s look pretty confusing mixed in with article text, especially for non-programmers and new users. Could you imagine someone trying to make their first edit to an article and coming across that kind of code... it would be worse than citation templates, IMO. So, to help reduce the number of hardcoded   s that are being used in article, I have a proposal. For things which commonly use hard spaces, templates should be created to help make it easier. For example {{cm}} would have the code: {{{1|}}} cm, so that you could use {{cm|5}} in articles, rather than 5 cm ... the former of these looks much more user-friendly to me. There are already "shortcut" templates like {{US$}}, so this may be a logical extension of existing templates and to help make editing more newcomer-friendly. Thoughts? Opinions? –Drilnoth (T • C • L) 18:25, 19 October 2009 (UTC)

My thought and opinion is that it would be a good idea to have a feature in MediaWiki which turns underscores to non-breaking spaces in the rendered page, except in URLs and within <code> or <source> tags. I think that more than 99% of underscores in typical articles are in template parameters, HTML attributes, etc. where they don't get rendered, and 99% of the ones which are rendered are in URLS or in computer code within <code> or <source> tags. If such a feature were implemented, a source of 5_kg would look much less cluttered than 5 kg and no more cluttered than 5 kg. ___A. di M. 19:12, 19 October 2009 (UTC)

To the new editor, it would seem that underscores turn into regular spaces; that's just too confusing. If you want a new MediaWiki feature, how about turning   into a hard space, but have the editor render the non-breaking space in a special color so it can be distinguished from a regular space. --—Preceding unsigned comment added by Jc3s5h (talk • contribs) 19:19, 19 October 2009

The fact is that some browsers convert literal hard spaces into regular spaces when submitting; so I'm not sure that'd work. (Maybe with some JavaScript-based magic, but that'd slow down the page for users with slow connections, so I'm not sure that'd be a good idea.) I don't think someone would be likely to accidentally find out that underscores turn to spaces (how often would they use underscores where they'd be rendered?), but if that's a concern, Noetica once proposed to use ,, for it. ___A. di M. 16:53, 20 October 2009 (UTC)

Another ammunition exception

Being discussed here. Chris Cunningham (not at work) - talk 10:09, 21 October 2009 (UTC)