Wikipedia talk:Manual of Style/Dates and numbers/Archive 96

the terabyte has become a unit of bandwidth
It seems that the terabyte is now a unit of bandwidth. Comments on the talk page are invited. Thunderbird2 (talk) 09:41, 7 March 2008 (UTC)

Quick NBSP question
Under the NBSP section, the MoS states "In compound items in which numerical and non-numerical elements are separated by a space, a non-breaking space (or hard space) is recommended to avoid the displacement of those elements at the end of a line." and as an example, gives "19&amp;nbsp;kg" as an example. My question is, if I wrote out "19" as "nineteen", would I still put a non-breaking space in-between the number and the unit of measure? In other words, which would be correct: "nineteen&amp;nbsp;kg" or "ninteen kg"? Thank you. Dlong (talk) 16:38, 19 February 2008 (UTC)
 * That (nineteen kg) would breach WP:MOSNUM. Also, if you prefer  to nbsps, you can use that. Sandy Georgia  (Talk) 16:42, 19 February 2008 (UTC)
 * Better use "nineteen kilogram", with a normal space. Full numbers go with full names. &mminus;Woodstone (talk) 17:52, 19 February 2008 (UTC)
 * In other words, irrespective of the non-breaking space (which I do not have a view on) use either "19 kg" or "nineteen kilograms". (surely not "nineteen kilogram" ?) Thunderbird2 (talk) 17:55, 19 February 2008 (UTC)
 * A normal space in "nineteen kilograms" (plural) will be just fine—a line break between those words won't look awkward or be hard to read. Neonumbers (talk) 03:43, 21 February 2008 (UTC)

I don't think the current wording makes that clear at all.

Followup. Suppose you have "is Cp = 38.171 J mol−1 K−1 at"

Tell me:
 * 1) Where you think the current MoS rule "recommends" non-breaking spaces. How many?  All seven spaces?  only some of them?  Five?  One? Three?
 * 2) Where this should logically be allowed to break.
 * 3) Whether they are different.

I'd say our screwball rules, read literally, clearly call for one nbsp in this case (before the J)—and that one is precisely at the place where it is most logical to allow a line break.

You may well think that the rules call for more than that. Explain where, and explain why that is so in terms of the current rule.

Now, don't peek at the coding until you've answered the parts above, and I'll put in the nbsp which to me seem absolutely essential: "is Cp = 38.171 J mol−1 K−1 at ..." Well, maybe not absolutely essential: there is an alternative for two of them; you can use a centered dot instead, but it should not be the dot on the line used in the actual article from which I borrowed this expression.

Obviously, there are a lot of people writing these rules who are incapable of grasping anything more complicated than "19 kg" when it comes to measurements, who might be blissfully unaware that people do use much more complicated numbers and much more complicated units than that. Our rules used to call for a nonbreaking space only in that situation, between a number in numerals and a symbol for a unit of measurement (metrologists like to distinguish these "symbols" from ordinary "abbreviations", which unlike the symbols are usually language-dependent and might change in the plural be followed by a dot or be italicized and things like that).
 * Why was it changed, without discussion, to even apply to spelled-out names of units of measurement in any case, whether they are preceded by a number in words or by a number in numerals?
 * And why hasn't it ever been changed to recommend non-breaking spaces in the places where it really matters, such as within a number itself (e.g., 14&amp;nbsp;3/16 in) and within the symbols for a unit (which are not "numerical and non-numerical elements ... separated by a space"), such as the "J&amp;nbsp;mol−1&amp;nbsp;K−1" in my example. Gene Nygaard (talk) 10:43, 28 February 2008 (UTC)

Well, I don't know what the page said when you read it but my own extrapolated conclusion from the current page text and my personal point of view is that the whole formula should be prevented from wrapping. However the current page text recommends using nowrap for the more complex cases which doesn't work in your case since your formula contains characters that is interpreted by the MediaWiki template engine. So I would instead use the new nowrap begin + nowrap end, like this: "is Cp = 38.171 J mol−1 K−1 at"

Which renders this:

"is Cp = 38.171 J mol−1 K−1 at"

Neat, isn't it?

You can read up on line wrapping at the brand new how-to guide Line break handling.

--David Göthberg (talk) 22:10, 12 March 2008 (UTC)

Misleading and difficult to parse advice about dates
The guidance says:
 * Full dates, and days and months, are normally autoformatted by inserting double square-brackets, as for linking.

I find that to be misleading and difficult to parse. It can be interpreted as requiring square brackets around solitary days and solitary months. People often quote the MOS as requiring links to solitary years (a user has just now made this claim on my talk page). It takes effort to explain that the MOS does not require that. I think that the phrase above needs changing. It also need supplementing by a specific statement that it does not require links to fragments such as centuries, years, months, days etc. Lightmouse (talk) 09:35, 7 March 2008 (UTC)
 * The present phrasing is indeed deplorable; but in rephrasing, we do need to say more than "no fragments". We do autoformat 2 April; whether this is a good idea is another question, which should be decided by another appeal to the developers, not here (because this page can't solve it). Septentrionalis PMAnderson 16:16, 8 March 2008 (UTC)


 * I agree with both your sentences. With regard to your first sentence, you are right, we need strong simple sentences that mention specific fragments. We could build on the negative sections that are already there. Would anyone like to propose wording?


 * In addition, we probably need a full time bot to tackle these fragments that sometimes are merely annoying and sometimes actually break autoformatting. Lightmouse (talk) 21:38, 11 March 2008 (UTC)


 * Does anybody have any suggested wording? Lightmouse (talk) 09:29, 13 March 2008 (UTC)


 * I have fixed some of the grammar of this new entry. I do think you should have put up a draft here first. This way, we can avoid points of view seeping into the text. Woody (talk) 13:52, 13 March 2008 (UTC)


 * In the absence of objections, I have eliminated the misleading and difficult to parse text. Explanations of the nuances does not appear to be effective, even many experienced editors make errors because nobody checks all date links in all preference modes. I have attempted to make it simple and shorter so that confused users can be directed to a directly worded appropriate bullet point about what is not required. Lightmouse (talk) 13:55, 13 March 2008 (UTC)


 * I know what you have tried to do, I simply thought that you should have brought the text here for discussion before implementing it. There were no objections to the principle of rewording the text; but the text itself had not been discussed. Other editors should have seen the text first. As it is, it needs the fixes in situ. Woody (talk) 14:04, 13 March 2008 (UTC)


 * I take your point, it is well made. Sorry about that. Lightmouse (talk) 14:11, 13 March 2008 (UTC)

Whose turf?
People are making changes in MOSNUM issues at the main MoS page, changing the wording there with respect to spelling out of numbers (which numbers are spelled out, adding "end of a sentence" stuff, etc.), then coming here and claiming in their edit summaries that this is an issue which needs to be discussed THERE, rather than HERE where it belongs. Lets make that clear at Wikipedia talk:Manual of Style, and move the discussion here where it belongs. It is nonsense like that which leads to the inconsistencies in various MoS pages which some editors like to complain about. Gene Nygaard (talk) 16:23, 6 March 2008 (UTC)


 * Note in particular that they have conducted that discussion there, relative to this quintessential "numbers style" issue, at that different forum, without even having the decency to post a notice here—the appropriate forum for the discussion in the first place—that that discussion was taking place. Then they hae changed the guidelines not only on that WP:MOS page, but on this WP:MOSNUM page as well, without ever having discussed it here.  Gene Nygaard (talk) 16:59, 6 March 2008 (UTC)


 * Calm down, please. I'm sure it was an oversight, rather than a deliberate slight. People will ask questions about it there, because the content is on the main MOS page, and people will answer there, and discussions will crop up there. The oversight is failing to either move it here or notify here, and there's no need for it to end up a big row, which your tone feels like it's heading towards. SamBC(talk) 17:46, 6 March 2008 (UTC)


 * Yes the subject needs to be talked about here before changes are made to the page. Fnagaton 12:21, 7 March 2008 (UTC)


 * The changes that have been objected to now were reversions of a previous undiscussed change. We need to discuss reversions of undiscussed changes now? It happens that it was discussed, somewhere else, but there was no forward motion in the change that was made and objected to, merely a reversion. SamBC(talk) 12:53, 7 March 2008 (UTC)


 * I reverted the change that has the edit summary starting with "Revert; see WT:MOS; it doesn't need to be discussed at WT:MOSNUM...". It makes the assumption that changes don't need to be discussed here when actually they do. Fnagaton 13:06, 7 March 2008 (UTC)


 * Furthermore, it the edits by User:Centrx which were reverted here were themselves a reversion of an earlier undiscussed change of the longstanding "ten" to "nine" by User:Tony1. Gene Nygaard (talk) 17:50, 7 March 2008 (UTC)

Talk of "jurisdiction" is off-topic wikilawyering; content is at both places, discussion was there, consensus was there, broader readership and input is there. Sandy Georgia (Talk) 12:39, 14 March 2008 (UTC)

Non-breaking spaces
In the current version of this MOS page there is a section called "Non-breaking spaces" that talks a little about how to handle line wrapping. We now have a detailed technical how-to guide about line wrap handling at Line break handling.

I would like that the section "Non-breaking spaces" of this MOS page in some way link to Line break handling. Perhaps as a "see also" link at the top of that section?

--David Göthberg (talk) 22:36, 12 March 2008 (UTC)


 * Since no one protested I boldly went ahead and added "See also: Line break handling and Manual of Style" to the top of the "Non-breaking spaces" section.
 * --David Göthberg (talk) 15:35, 13 March 2008 (UTC)

Why was this raised on this page rather than at WP:MOS, which gets broader attention? Sandy Georgia (Talk) 12:36, 14 March 2008 (UTC)


 * SandyGeorgia: Oh, I see you misunderstood what I was asking. With "this MOS page" I meant "this Manual of Style page", that is Manual of Style (dates and numbers) = "MOSNUM". I see that my way of writing it wasn't very clear, sorry. I wanted to add those links to THIS page.
 * But anyway, WP:MOS also has such a section so I asked the exact same thing there. So I did raise the question on both affected pages. :))
 * --David Göthberg (talk) 16:24, 14 March 2008 (UTC)

Remove edit blocking please
The page has been blocked from editing since 6 March.


 * Edit blocking of pages should only be used if there are many editors doing bad things. If there are only a few editors doing bad things, then the target should be those editors, not the page.


 * Edit blocking should only remain in place if the bad things would still happen without it.

I do not see any justification for the edit block today. Please remove it. Lightmouse (talk) 09:27, 13 March 2008 (UTC)


 * That's a good suggestion, Lightmouse. Unfortunately, when the protection was applied it locked in an edit that made a guideline inconsistent with the corresponding guideline at WP:MOS. The protection explicitly does not aim to endorse or entrench that edit. Since the content of the edit was undiscussed and non-consensual, and contradicts WP:MOS, reverting it would be a Good Thing. Don't be surprised if such a Good Thing happens, at the first opportunity.
 * – ⊥¡ɐɔıʇǝoN oetica! T– 09:58, 13 March 2008 (UTC)


 * That would not surprise me at all. What surprises me is the edit block. Lightmouse (talk) 10:27, 13 March 2008 (UTC)
 * Yes, I'm sure it raised more than a few eyebrows. We'll see. Let's just continue to act in good faith, with respect for discussion and consultation towards consensus.
 * – ⊥¡ɐɔıʇǝoN oetica! T– 10:40, 13 March 2008 (UTC)
 * I agree with these calls for the block to be removed. It might have been appropriate for a few days to allow things to cool down, but it's unacceptable for a longer period. We need to agree here on a strategy to avoid the revert-wars. This need should not stand in the way of an immediate unblocking: the issues should probably go to arbitration while other ongoing issues are dealt with in MOSNUM. Tony   (talk)  12:22, 13 March 2008 (UTC)
 * Even though I see no resolution in sight, I do think that we can all be civil here. Do not keep on reverting, discuss it here or WT:MOS, otherwise I won't hesitate to protect this again. Woody (talk) 12:31, 13 March 2008 (UTC)
 * Thanks Woody. In fact, the issue has been discussed extensively at WT:MOS recently. Any proposal for change to the version that had been in place for several months should indeed be raised there, or here if the proposer prefers.
 * – ⊥¡ɐɔıʇǝoN oetica! T– 10:08, 14 March 2008 (UTC)

large ranges of numbers
I'm not sure how to format a range of large rounded numbers. An article I saw expressed the figure of 'twenty to thirty thousand' as "20-30 000" Since commas and not spaces are supposed to be used in large numbers, I changed it to "20-30,000". However, to me it still doesn't look right. Can anyone confirm this is probably the best way, or suggest an alternative?? Thanks a lot, Chebyshev (talk) 03:51, 12 March 2008 (UTC)


 * When this is discussed in style guides, the normal decision is that you should not abbreviate the first number. If your intention is 20 000 to 30 000, have exactly that, or 20 000 – 30 000 with a spaced en dash (which unfortuantely might look like a minus sign!). Spacing is a separate matter, much discussed recently. If you do not use spaces between the digits of a number, you should use an unspaced en dash: 20,000–30,000. This is all sound advice generally, but others here may have different advice more finely attuned to Wikipedia.
 * Note that even in words the meaning has to be clearly presented. If your meaning is 20,000–30,000, in some contexts your words (spoken or written!) may need to be twenty thousand to thirty thousand, without any shortening. One such context:
 * "The ion-emission drive enabled the craft to accelerate from twenty thousand to thirty thousand miles per hour in one week."
 * Without the first thousand, the meaning would be ambiguous. Both interpretations would be realistic. To disambiguate the other way, you might even need to have this:
 * "The ion-emission drive enabled the craft to accelerate from twenty miles per hour to thirty thousand miles per hour in one week."
 * No wonder we invented figures, and abbreviations. :)
 * – ⊥¡ɐɔıʇǝoN oetica! T– 05:36, 12 March 2008 (UTC)
 * Thanks a lot for your help Chebyshev (talk) 11:11, 17 March 2008 (UTC)

Trials and disambiguations
Opinions are invited on how best to disambiguate the megabyte at DEC 3000 AXP. I would like to avoid an unnecessary strand, so please reply directly on the talk page and not here. Thanks. Thunderbird2 (talk) 17:31, 16 March 2008 (UTC)

Scientific notation (aka Standard Form) discussion
There's a discussion at the main MOS talkpage about the prescribed formatting of exponential notation/scientific notation/standard form. Just to make sure interested folks know. SamBC(talk) 22:02, 21 March 2008 (UTC)

Delimitnum revisited
I am no mathematician, but I am a crazy software engineer and researcher that like to think outside the box. Or rather look at problems from many angles, like for instance backwards.

So delimitnum did get a "deadly failure". (See discussion in an earlier section.)

So lets do it the other way around. Instead of chopping up a number using maths and put in commas etc, instead lets put the number together only using strings and no maths.

Here is a basic example:



Which would render something like this :


 * 1.123 45 × 1025 kg

And a monster example:



Which would render something like this :


 * &minus;1,500,000.123 45(30) × 1025 kg

$-150,000.123 kg$

Such numbers should be pretty simple to input since the user only has to input normal dashes "-" and a normal "x", not the special maths symbols. I let them input the base 10 too, since that makes the input more readable and is more flexible. Oh and this of course means that inputting "+" and "000" is no problem either, they will not be dropped when the number is rendered, since it isn't a number that is rendered but rather a string.

I think I know how to code up such a template. The only maths involved here would be to detect that the string "(30)" is not a number, since it seems you guys don't want any spacing between the 45 and the (30). But if you allow a spacing there the template code gets much simpler.

With spacing I of course do not mean "spaces" but the narrow spacing trick mentioned before that makes the numbers copy-and-pasteable to other programs.

Would such a template be useful?

--David Göthberg (talk) 23:43, 22 March 2008 (UTC)


 * David, it’s great to hear from someone who’s really interested in what delimitnum has to offer. As you’ve no-doubt witnessed at the Delimitnum sandbox, using math-based parser functions within a template is a recipe for banging one’s head against a wall. Your proposal sounds like it would be perfectly workable. As you may also have seen on the initial proposal, here on Archive #94 of Talk:MOSNUM, the delimiting is done with  tags. That itself shouldn’t be a problem except that spans following the digit 1 are specified slightly narrower so they have the same visual appearance. The magic word-based version of delimitnum will use character-based parser functions as you are proposing. With any luck, we should have it within about a week. Since it will used character-based parsing, it will be imune from rounding issues. For instance,  will return 1.1200(25) kg, not 1.12(25) kg.  Where I could really use some support right now is on an issue of computer nomenclature. It involves abandoning the use of terms like “250 GiB file size” and proposes to adhere to the better-recognized units (e.g. “250 GB file size”). What do you think about this position statement and this specific MOSNUM policy (in light green)?  Greg L (my talk) 00:23, 23 March 2008 (UTC)


 * Ah okay. So the delimitnum magic word is not that far away. Right, then no use in spending time on making a template for it. I noticed some days ago that you were banging your head in the wall with delimitnum, just had to think for some day what to say.
 * Ok, I'll take a look at the GB vs GiB issue although I have to admit I dislike style issues. I am a software engineer, not a stylist. But I guess in the GiB issue you need some engineers.
 * --David Göthberg (talk) 01:21, 23 March 2008 (UTC)

Silly, I always forget to ask this, been thinking about it for days: Greg L: Why didn't you use MediaWiki TeX to render the numbers? It does have text output for simpler formulas. Or does it render the text to bad?

--David Göthberg (talk) 01:30, 23 March 2008 (UTC)


 * David, the biggest advantage of the upcoming magic word is it will automate the job of a grouping the digits. Many users will simply forget that you don’t leave a single dangling digit, like 1.234 456 8. No matter what the editor inputs, the magic word knows to parse as follows:

2.1 2.12  2.123  2.1234  2.123 45  2.123 456  2.123 4567  2.123 456 78  2.123 456 789  2.123 456 7890  2.123 456 789 01  2.123 456 789 012  2.123 456 789 0123


 * Regarding the use of MediaWiki TeX, I do if it’s absolutely necessary, such as for $$V = \sqrt{\frac {{K_b\over 2} \cdot T}}$$ But for straightforward numeric equivalencies, like this:

…the quantity known as the Avogadro constant, is an experimentally determined value that is currently measured as being $6.022$ atoms (2006 CODATA value).


 * …I use either hand coding or the template to avoid the change in text associated with , which interrupts the reading flow. Regards. Greg L (my talk) 02:54, 23 March 2008 (UTC)


 * 1: Ah, I didn't know about how to handle those dangling digits. We delimit numbers differently in my country (Sweden). So yeah, automating it is nice. He, come to think of it, that puts that magic word into trouble, since preferably it should be able to delimit numbers in the right way on all the different language Wikipedias. That coder has a big task ahead of him... Or of course, he could be "lazy" and just do English delimiting and the other languages have to do hand delimiting. Ehm, perhaps I should code a dnum template for my language instead.
 * 2: I think you might have misunderstood what I meant by using MediaWiki TeX. The default is that it outputs text, not images, for simpler "formulas". So it doesn't interrupt the reading flow as you thought. Perhaps you have set TeX to only show images in your Wikipedia user preferences?
 * --David Göthberg (talk) 03:19, 23 March 2008 (UTC)


 * Regarding your point #1 above, Americans don’t usually delimit the fractional side of the significand. Professionally done work does and the only proper way I know originally came from the ISO and is done as illustrated here at the NIST and here at the BIPM (SI Brochure: 5.3.5 Expressing the measurement uncertainty in the value of a quantity). Here is the NIST’s |search_for=Planck+constant+in+eV+s+ last two-digit grouping, their |search_for=Planck+constant+over+2+pi 3-digit and their |search_for=Planck+constant+over+2+pi+times+c+in+MeV+fm four-digit last group.As for your point #2, I don’t understand. I’ve set my Wiki prefs to “Recommended for modern browsers” and it shows the above example as large symbols. You seem to be suggesting that simple formulas show as straight text. Please provide an example. I am inclined though, to continue to use hand-coded text or use of magic words and templates, as I am certain what the appearance will be for all readers irregardless of their browser and user prefs.  Greg L (my talk) 05:02, 23 March 2008 (UTC)


 * P.S. If by “ differently in my country” you mean, the Euro-way: narrow spaces on both sides of the decimal marker, and the decimal marker is a comma, not a full-stop, then yes, it’s done differently in America. And that’s the convention the en.Wikipedia standardized upon: comma delimiting to the left of the decimal marker, which is a full a full-stop. The magic word does not even pretend to change any of that; advocating to do so just would never happen. All it does is add much-needed narrow-space delimiting to the fractional side of the decimal marker. Greg L (my talk) 05:25, 23 March 2008 (UTC)


 * 2: Ah, the point is moot. I tested my user settings for math (MediaWiki TeX rendering). As I remembered I can get it to show as HTML for simple formulas and as PNG when more complex. But in both cases it doesn't delimit numbers. So 123456789.123456789 just shows up like this when using TeX: $$123456789.123456789$$ So doesn't help us at all. (Of course, I might be using TeX the wrong way.)
 * 1: Well, the way I learnt to delimit numbers in school in Sweden is like this:
 * English: 1,234,567.123 456
 * Swedish1: 1 234 567,123456
 * Swedish2: 1.234.567,123 456 or was it 1.234.567,123456
 * My bank statements and my Swedish MS Windows use Swedish 1, and that is the more common way we do it. I have mostly seen Swedish 2 in really old printed matter, as in early 1900s and older.
 * I also took a quick look at the Swedish Wikipedia, and there MediaWiki TeX clearly is set to use the decimal comma ",".
 * See what I mean with that the delimitnum magic word will need localisation if to be used in other language Wikipedias? I am not advocating to change the delimiting in English Wikipedia, why would I? But magic words are part of the MediaWiki software and as such is supposed to work on all language Wikipedias. And that is why I mentioned it since you seem to be in contact with the dev that is coding that magic word.
 * --David Göthberg (talk) 06:07, 23 March 2008 (UTC)


 * Yes. I understand enough about the details of the magic word to know that it should be almost trivial to customize it for any language. The hard part is all the parsing logic and the rules governing span width. I was familiar with Swedish1 (and a variant that delimits the same way on the fractional side too). I didn’t know about Swedish2. I like Wikipedia because it is such an awesome way to learn. It’s my bed time. Bye. Greg L (my talk) 06:14, 23 March 2008 (UTC)


 * Yeah, I agree. Since using narrow spaces seems to be the most popular in most languages it should be trivial to switch between using a decimal "." or ",". And narrow spaces is anyway the delimiting that has the lowest risk of misunderstanding. Anyway, sleep well. Bedtime for me too I think.
 * --David Göthberg (talk) 06:26, 23 March 2008 (UTC)


 * It certainly can be done.  is language dependant. Jɪmp 04:00, 28 March 2008 (UTC)

Uncertainty vs. repeating decimal
I have just edited the Conversion of units article because it used parentheses to indicate a repeating decimal. For example, it used "3.3(3) × 10−4 m" to mean that the rightmost 3 repeats forever. I changed it to read "3. 3 m" because this notation is also used for repeating decimals, and it will not be confused with uncertainty.

An example of an expression of uncertainty is in Seidelmann's Explanatory Supplement to the Astronomical Almanac (1992, p. 693) where the Newtonian constant of gravitation is given as 6.67259 (85) m3kg-1s-2. A longer expression for the same thing would be 6.67259 ±0.00085  m3kg-1s-m 3kg-1s-2.

I believe this guildeline should have a section added to specify that repeating decimals are expressed with overbars, not parentheses, to avoid confusion with an expresion of uncertainty. I also believe that it should have a section recommending that uncertainty should be expressed with ± notation rather than parentheses because people unfamiliar with this notation would not know the order of magnitude of the uncertainty. That is, in the previous example, it isn't obvious whether the uncertainty is ±0.00085, ±0.000085 , or ±0.0000085. --Gerry Ashton (talk) 19:22, 8 March 2008 (UTC)


 * Good point. This looks like sensible advice to me. Thunderbird2 (talk) 19:32, 8 March 2008 (UTC)
 * ±8.5 may be easier to read, as long as we're recommending; but the deprecations should include a caveat: 3.3(3) × 10−4 m would be fine, if it were clear what was meant. Septentrionalis PMAnderson 21:40, 8 March 2008 (UTC)
 * BIPM advocates use of parentheses in this way. And that would be fine if it were clear to everyone, but that's a big if isn't it? Thunderbird2 (talk) 21:50, 8 March 2008 (UTC)


 * Note above that the plus-or-minus sign requires a space, and en dashes or minus signs are required, not hyphens. Tony   (talk)  00:51, 9 March 2008 (UTC)
 * Disputed, as invisible nonsense. Septentrionalis PMAnderson 16:19, 10 March 2008 (UTC)
 * Is it the span tag to create the overbar that you dispute? If so, do you have a solution to the ambiguity? --Gerry Ashton (talk) 19:36, 10 March 2008 (UTC)
 * No, I dispute the mandatory space around ± (we should do what is clearest in the circumstances), and the mandatory use of a endash instead of a hyphen in an exponent, where the difference is invisible to the reader. Septentrionalis PMAnderson 22:47, 10 March 2008 (UTC)
 * The substantive question of ambiguity should be dealt with by suggesting explanatory phrases at first use of any of these notations. Any of them will be clear to some readers and unknown to others. Septentrionalis PMAnderson 22:49, 10 March 2008 (UTC)
 * I still feel we should recommend different typography for uncertainty vs. repeating decimals, but I think PMAnderson's suggestion to use explanatory phrases near the first use is wise. Since this is a matter of typography, it would be particularly difficult for readers who don't understand the convention to think of keywords to search for in their favorite search engine. --Gerry Ashton (talk) 23:23, 10 March 2008 (UTC)

The problem with overlines for repeating digits is that there are at least two variants and both have their issues. — Christoph Päper 20:30, 13 March 2008 (UTC)
 * 1) Inline CSS:   – 0.1 37
 * 2) Unicode:   =   (U+0305) or   =   (U+0304) – 0.13̅7̅ or 0.13̄7̄
 * The span approach displays properly on Windows XP, both Internet Explorer 7 and Firefox. The unicode approach does not display properly in either case; it displays as a square after the 3 and the 7. --Gerry Ashton (talk) 02:35, 14 March 2008 (UTC)
 * Yes, the problem with the Unicode approach is its state of support, but the CSS approach is more fundamentally flawed in that the style, which conveys the information, is lost in plain-text environments, such as copy and paste, text or non-CSS browsers, search engines. Using a different HTML element that, unlike, has a default styling only helps in few of these cases. Years ago I used a non-combining overline (or macron) in front of the first digit to repeat:   – 0.1¯37. It doesn’t look as good, but is very safe. One could also imagine using brackets or braces instead of parentheses: 0.1[37] or 0.1{37}, but I think this would be a WP-specific convention, which we try to avoid. — Christoph Päper 09:09, 14 March 2008 (UTC)
 * I've never seen a non-combining overline before; I didn't even know this character existed. I suspect few people would know how to enter it. I think it would be just as much a WP-only convention as square brackes or curly braces. --Gerry Ashton (talk) 00:30, 22 March 2008 (UTC)

Major objection: "3. 3 " is completely illegible-as-intended in Mozilla-based browsers (Firefox, Mozilla, SeaMonkey, Camino) under MacOS X using default fonts (i.e. for probably half of all Mac users, since many have abandoned Safari as a slow, feature-poor and crash-prone piece of frak). For those who will even notice the difference at all, it does not look like a three with an overline, it looks like a three with a flat top, like those found in many serif fonts. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93; ‹(-¿-)› 00:43, 5 April 2008 (UTC)

Ga, Ma, ka preferred to bya, mya, tya
Wikipedia talk:Manual of Style (dates and numbers)/Archive 78 danced around the question, but never quite addressed the question of whether to discourage use of mya, focussing instead on discouraging MYA. Everyone seemed to prefer ka, Ma, Ga although there was some confusion over capitalization (SI usage for multipliers is quite unambiguous: k for kilo, M for mega, G for giga, T for tera, but it seems this wasn't universally grasped). Once archived, the discussion ended without addressing it. As phrased now, it leaves the impression that mya is an equally desirable choice to Ma, yet I don't believe that was the intention of those discussing the question. Reopen? LeadSongDog (talk) 07:40, 14 March 2008 (UTC)


 * I prefer to spell it out in full as "million years ago" wherever it occurs in the article, and use Ma where abbreviation is necessary (e.g. in taxobox fossil ranges). I feel this looks smarter and saves people translating it in their heads each time they see it.  I agree that mya is obsolete though.  That said, I've not got a very firm opinion yet... Verisimilus  T  08:36, 14 March 2008 (UTC)
 * My feeling from working with a number of geophysicists is that Ma is preferable, and I'm quite willing to agree to a prescription here that it should be preferred. I can't agree with Verisimilus that the whole three-word phrase should be trotted out many times in an article, unless used only twice or three times, perhaps: tedious to read. Besides, the fact of the abbreviation itself helps to focus the readers' minds. Tony   (talk)  08:47, 14 March 2008 (UTC)

The present text reads :*Abbreviations indicating long periods of time ago—such as BP (before present), as well as various annum-based units such as Ka (kiloannum) and kya (thousand years ago), Ma (megaannum) and mya (million years ago), and Ga (gigaannum or billion years ago)—are given as full words and wikilinked on first occurrence.
 * BP: Do not convert other notations to BP unless you are certain of what you are doing. In some contexts the unit BP is actually defined as "years before 1950 CE/AD", not "years before the literal present", and the conversion may introduce an error if the date being converted is not a wide approximation (18,000 BP) but a more narrow one or an actual known year. BP years are given as 18,000 BP or spelled out as 18,000 years before present (not 18,000 YBP, 18,000 before present, 18,000 years before the present, etc.)

I would suggest a change to read :*Abbreviations indicating long periods of time ago—such as BP (before present), as well as various annum-based units such as Ka (kiloannum), Ma (megaannum)and Ga (gigaannum or billion years ago) are given as full words and wikilinked on first occurrence. Where older source quotations use the now-deprecated abbreviations kya (thousand years ago), mya (million years ago), or bya (billion years ago) this should be explained to the reader as in "a measured Libby radiocarbon date of 35.1 mya" (million years ago, 35.1 Ma in modern usage) had to be calibrated against then newly available stratigraphic dating references in order to estimate a Cambridge standardized date of 30.2 Ma BP.
 * BP: Do not convert other notations to BP unless you are certain of what you are doing. In some contexts the unit BP is actually defined as "years before 1950-01-01", not "years before the literal present", and the conversion may introduce an error if the date being converted is not a wide approximation (18,000 BP) but a more narrow one or an actual known year. BP years are given as 18,000 BP or spelled out as 18,000 years before present (not 18,000 YBP, 18,000 before present, 18,000 years before the present, etc.)
 * Caution: Some source materials will indicate whether a date is calibrated or not simply by a change in capitalization. This is often a source of confusion for the unwary user.

Would something like this work for us? LeadSongDog (talk) 17:37, 14 March 2008 (UTC)


 * Is it really necessary to wikilink each term if you've explained what it means? I find excessive links distracting.  What more could a user want to know about "mya" than what it means? Verisimilus  T  18:18, 14 March 2008 (UTC)
 * Fair enough. I had them each wikilinked to annum anyhow, the one time should be enough.LeadSongDog (talk) 19:53, 14 March 2008 (UTC)

Now we have :*Abbreviations indicating long periods of time ago—such as BP (before present), as well as various annum-based units such as ka (kiloannum), Ma (megaannum) and Ga (gigaannum) are given as full words and wikilinked on first occurrence. Where older source quotations use the now-deprecated abbreviations kya (thousand years ago), mya (million years ago), or bya (billion years ago) this should be explained to the reader, as in "a measured Libby radiocarbon date of 35.1 kya" (thousand years ago, or 35.1 ka in modern usage) had to be calibrated against then newly available stratigraphic dating references in order to estimate a Cambridge standardized date of 30.2 ka BP cal.
 * BP: Do not convert other notations to BP unless you are certain of what you are doing. In some contexts the unit BP is actually defined as "years before 1950-01-01", not "years before the literal present", and the conversion may introduce an error if the date being converted is not a wide approximation (18,000 BP) but a more narrow one or an actual known year. BP years are given as 18,000 BP or spelled out as 18,000 years before present (not 18,000 YBP, 18,000 before present, 18,000 years before the present, or similar.)
 * Caution: Some source materials will indicate whether a date is calibrated or not simply by a change in capitalization. This is often a source of confusion for the unwary user.

Absent further comment, I'll make the above change in a day or two.LeadSongDog (talk) 18:10, 17 March 2008 (UTC)


 * emend: strike "and wikilinked"; RC only gd 4 ¬50000a. --Verisimilus  T  19:56, 17 March 2008 (UTC)(no kb!)
 * How on earth did I miss that??? Good catch!LeadSongDog (talk) 21:56, 17 March 2008 (UTC)

Done. LeadSongDog (talk) 18:00, 18 March 2008 (UTC)

Just a followup note to the geophysicist point: That may be true, but I know from equally direct experience that anthropologists definitely (and paleontologists sometimes but not always) prefer kya, mya. I'm happy with the current text since it isn't FORCING one or the other. I strongly suspect that anthro types editing anthro articles will initially use, and reject reversal of, the format they use, and geo/astro types will do likewise with their preferred symbols, so it shouldn't be a big deal. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93; ‹(-¿-)› 01:39, 5 April 2008 (UTC)

Years
Start of discussion moved from Lightmouse talk page You had best stop un-bracketing years until you get some consensus. I'm reporting this to WP:ANI. Baseball Bugs What's up, Doc? 12:59, 20 March 2008 (UTC)


 * You were already blocked once for this activity. STOP IT. Baseball Bugs What's up, Doc? 13:03, 20 March 2008 (UTC)


 * Right I don't know what is going on here but please stop for now. Contribute to the discussion on the | AN/I but stop delinking the dates. Theresa Knott | The otter sank 13:10, 20 March 2008 (UTC)


 * As I understand it, the place to debate policy on dates is Wikipedia talk:Manual of Style (dates and numbers). Feel free to raise it there. Lightmouse (talk) 13:13, 20 March 2008 (UTC)


 * First rule: Don't cop an attitude with an admin. Baseball Bugs What's up, Doc? 13:17, 20 March 2008 (UTC)


 * That's fine I don't care. Lightmouse I do not wish to debate the policy. I only wish to make sure that you actually have concensus for your changes. Can you point me to the discussion that shows this please. Theresa Knott | The otter sank 13:20, 20 March 2008 (UTC)


 * At the very least, the user did not get consensus from anyone to clobber the "year in baseball" template entries. Baseball Bugs What's up, Doc? 13:41, 20 March 2008 (UTC)

Lightmouse, you continued making controversial edits after you've been contacted abouts this. I've withdrawn your AWB approval for duration of discussion at WP:ANI. Please respond there. Max S em(Han shot first!) 14:02, 20 March 2008 (UTC)
 * Baseball Bugs, I am not sure what 'cop an attitude' means but that phrase along with the word 'defiant' used elsewhere sounds like an accusation of incivility. I have no incivil intentions in my responses. I try to be polite and expect the same from others. Please assume good faith.


 * There are popular misconceptions about date links so it does not surprise me that it is being questioned. I would prefer not to describe the policy here. Quite apart from anything else, you may not feel comfortable with what I say about it. There are plenty of other editors there that have extensive experience with this issue at Wikipedia talk:Manual of Style (dates and numbers). I would be happy to see you there. I hope that helps. Lightmouse (talk) 14:36, 20 March 2008 (UTC)

End of discussion moved from Lightmouse talk page


 * Hard to discern what this is about. If Lightmouse is delinking trivial chronological itmes, such as 1998 and 1970s and 20th century, I applaud it. He has the weight of MOS behind him. Read the guidelines, please. Tony  (talk)  02:57, 23 March 2008 (UTC)


 * I randomly selected about 25 of Lightmouse's recent edits of this nature, and they all seem well within current date policy to me. -/- Warren 04:25, 23 March 2008 (UTC)


 * I have the impression that Lightmouse was, as usual, implementing MOSNUM consensus, when an admin over-reacted to a change that was not to his liking. The discussion of the entire non-incident is archived here. Thunderbird2 (talk) 15:31, 23 March 2008 (UTC)


 * I've just noticed that Lightmouse has not done any editing since posting the avove message. I hope he is just taking an Easter break. Thunderbird2 (talk) 15:36, 23 March 2008 (UTC)


 * I have been watching the discussion. I appreciate the contributions made by all. I would appreciate your further assistance in getting my AWB approval reinstated. See the statement above by MaxSem (I've withdrawn your AWB approval for duration of discussion at WP:ANI#Units and Years) and my request for reinstatement at Wikipedia_talk:AutoWikiBrowser.
 * Lightmouse (talk) 11:03, 24 March 2008 (UTC)


 * If Baseball Bugs is the admin in question (I'm not sure I followed the rant correctly), he should actually be severely sanctioned at ANI. It is unbelievably inappropriate for an admin to attempt to intimidate another editor into yielding on a legitimate editing dispute by pulling rank and very obviously implying a blocking or other admin-level retaliation threat, especially a borderline-incivil one at that. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93;  ‹(-¿-)› 02:13, 5 April 2008 (UTC)

Sea of blue
The template name nowrap is linked four times (on every occurrence) in the section Non-breaking spaces. I would like that we do as we teach, so I would like to fix that. That is, only have the first occurrence linked and the other as normal text. Easy enough to do, just use instead of, which will render like this:. Any one that disagrees?

--David Göthberg (talk) 18:41, 27 March 2008 (UTC)


 * So fix it. Nobody's going to worry about as minor an edit as that.  If anyone even notices, they'd probably wonder why they hadn't noticed such a glaring oversight. Jɪmp 23:30, 27 March 2008 (UTC)


 * Well, personally I think that any edit to the MOSxxx pages should be discussed first (except really minor things like fixing some spelling etc). And even though tradition and recommended style is to not wikilink a word many times in one page or section, tradition up until now has been to wikilink template names on every occasion in documentation. So I don't like to just be bold and make an edit to the MOSNUM that goes against tradition.
 * Of course, the reason people have been wikilinking the template names all the time might have been that up until now they did not have a simple way to write a . But now we have the brand new tlc, tld and tlf. They produce output like this:, and.
 * Anyway, I did the edit to the MOSNUM since it seems no one disagreed.
 * --David Göthberg (talk) 18:21, 28 March 2008 (UTC)


 * Generally I'd agree that edits should be discussed first except minor ones. I'd call that minor. Jɪmp 15:32, 31 March 2008 (UTC)
 * WP:BOLD, WP:BRD. While the R in BRD is more likely to happen here than on many other pages, the B is policy, and the D is the only way that WP works in the first place. Major changes certainly should be discussed, but I've seen a great number of really good ideas introduced into MOS* pages by bold edits, often reverted, discussed, modified and reinstated, less often but still pretty frequently simply accepted because they were wise. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93;  ‹(-¿-)› 02:20, 5 April 2008 (UTC)

ISO 8601 dates
I'm seeing a rash of these in various articles within the text. Any chance there is a bot that can go in and change these when not used with a reference? Vegaswikian (talk) 06:52, 31 March 2008 (UTC)


 * Any such bot would have to be semi-automatic, with a person approving every change, because there is no automatic way to detect whether a piece of text is a direct quotation. --Gerry Ashton (talk) 17:45, 31 March 2008 (UTC)

Decades fix rationale
Sometime in the intervening long while since I looked at it, someone added an "okay" to abbreviate decades as long as the century was obvious. This is clearly untenable, because we (people who write and speak, I mean) simply don't do this outside of our own immediate past and future (with exceptions, like "Gay '90s" or "'49ers", that have become stock phrases and even re-used for other purposes like sports teams). When's the last time you saw something like "In 1075 BC King Grognar was ousted from his throne by his brother, but by the '60s had regained his kingdom with the help of the neighboring Fnord Empire"? I can't imagine ever seeing that in a scholarly or even vaguely serious work. Furthermore, the text as of a few hours ago ignored the fact that we use abbreviated decades most often, most pointedly, to refer to social periods that roughly coincide with decades, and with such consistency that they can become (sourceable) buzzwords or even signifiers for everything about society at a vaguely-defined range of time, and that this really isn't related in any way to identifiability of the century (see previous 2 examples). When I say "I grew up in the '80s" that conveys an enormous amount of information about me and my probable world-view. If I say "that model of our product was discontinued in the '80s", referring to a literal decade, this (lazy and unencyclopedic) usage conveys no such additional latent information. The usages are radically different, and the current (as of this writing) re-draft solves these problem. The redraft clearly also reflects consensus current practice, as it is a very regular AWB/bot or gnome edit to convert things like "the next phase of her career began in the early '40s" to "...1940s" (as far as I know I have never been reverted, ever, on such a correction), but articles about counterculture, civil rights, etc., can and do use "the '60s" when appropriate because it has nothing to do with the literal span of 1960 to 1969, but the societal changes that began happening in the mid-1960s and carried through to the early 1970s. No one would go to the article on the 1890s and change the phrase "popularly called the 'Gay '90s'" to read "'...Gay 1890s'". Anyway, I don't think anyone would revert me on this, but I like to provide rationales for major changes, especially here. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93; ‹(-¿-)› 02:55, 5 April 2008 (UTC)
 * We might want to add a clarification along the lines of "Decades should not be abbreviated in this manner even if they can be unless there is some connection between the usually-abbreviated social period and what is being discussed in the sentence (She grew up in the 1960s, and moved to Boston in 1971. Growing up in the '60s, her attitude about civil rights differed significantly from that of her parents.)" Maybe someone can come up with shorter examples that get the point across as clearly.  The lead sentence could probably be tightened too. Season 3 of Battlestar Galactica is about to start so I'm losing focus...  Anyway, I think the clarification is important; I really have seen articles consistently abbreviating "'60s" but not other decades, because someone is not quite realizing that every reference to the "special" time period is not a reference to what was defining about the period and its societal goings-on.  There's a world of difference between a '60s rock band in San Francisco and a 1960s classical quintet (even in S.F.!), or a '60s civil rights activist in Chicago and a 1960s accountant in the same city (or a 1960s activist for better funding of Chicago elementary schools). The distinction is important.  I'll see if I can think of a way to stick the idea in there in just a few words, but may need help.  —  SMcCandlish  &#91;talk&#93; &#91;cont&#93;  ‹(-¿-)› 03:18, 5 April 2008 (UTC)
 * After a dozen or so attempts, I think I finally got it. I managed to compress (I think) every single idea above into a passage that's only slightly longer than when I raised the "meaningful connection" issue above, and also took the time to tighten up, clarify, flow-improve and bulletize. I think it is quite clear and readable now, and hints at all of the guidance that it needs to.  Yes/no? —  SMcCandlish  &#91;talk&#93; &#91;cont&#93;  ‹(-¿-)› 03:54, 5 April 2008 (UTC)

Units
Which system to use * For US-related articles, the main units are US units; for example, 23 miles (37 km). * For UK-related, the main units are either metric or imperial (consistently within an article). * For other country-related articles, the main units are metric; for example, 37 kilometres (23 mi).

Let us end the confusion by adopting the Metric first, Colonial/Imperial second rule (Option 3). It will make editing easy when the U.S. metricates. I must also add that some items like broadcast antenna heights are always listed in meters and the current rule is a very ambiguous burden to subjects such as these. SirChan (talk) 17:18, 25 March 2008 (UTC)


 * I agree - it does seem confusing and unnecessary to make a special exception from the general "use SI units" rule in this instance. Verisimilus  T  21:07, 25 March 2008 (UTC)


 * I diagree. There is no guarantee that the US will ever metricate, and for US articles it makes sense to have the nationally preferred system of measurement listed first. Cheers, Rai - me  22:48, 25 March 2008 (UTC)


 * To Verisimilus: The American has to tilt their head to read the Colonial units on a non-US or international article, the Brit might have to change mindsets every time he reads an article, and the others have to tilt their heads in order to read the metric units in a US specific article or might not have an appropriate metric value in a British article.

To Rai-me: My mom has a maxim, "Always be prepared." The U.S. government has stated since 1866 that this is the preferred system but has been busy with other matters. Metrication has been increasing in recent years--just look at bottled water! This is on the internet; any English-speaking person can find out information about something U.S. or U.K. specific. (To all:) These rules are Byzantine! I might as well include Julian dates in articles also to accommodate historians and Eastern Europeans. SirChan (talk) 02:30, 26 March 2008 (UTC)


 * By your mother's philosophy, should we also create an article for the 2096 Summer Olympics to be prepared? Wikipedia is not a crystal ball. It is very unlikely that the US will metricate completely at any point in the near future. Whether the US government has stated that the metric system is the preferred system or not has no bearing on this situation. What matters here is what readers will understand, and for Americans, that is obviously imperial units over SI ones. Metrictaion may be increasing slightly (bottled water is actually still mostly in imperial units - soda bottles, however, are in litres), but the system used by far the most often is imperial. I fail to see how the feet (m) system is "Byzantine" when used in American-related articles, as the imperial system is the one that Americans, who are the most likely to read about American topics, understand the best. Wikipedia is optimized for readers over editors, so the guidleines should remain as is. Cheers, Rai - me  01:35, 29 March 2008 (UTC)


 * By the same token, it is very much true that many measurements in U.S.-related articles were originally made in metric units. The present rule is ludicrous.  Original units should be first; for articles related to the United States, that will often mean English customary units—but that is by no stretch of the imagination an invariable truth, nor something to aspire to in our MoS.  Gene Nygaard (talk) 03:47, 29 March 2008 (UTC)


 * I am not arguing that we should continue to use the "imperial (metric)" system because it was used first in American-related articles, and shouldn't be changed now. The rules are not, nor should be, the same as those outlined in WP:ENGVAR. The present rule is no more "ludicrous" than the rule to use US$ for American articles, £ for British articles, € for French articles, etc., and not US$ throughout. It makes sense to use the system in an article that is used by the readers of the country relating most to the article topic, so the status quo should not be changed. -- Rai - me  17:06, 29 March 2008 (UTC)


 * "I fail to see how the feet (m) system is 'Byzantine' when used in American-related articles, as the imperial system is the one that Americans, who are the most likely to read about American topics, understand the best."

I'm not talking about the measuring systems here, I'm talking about the rules on the units. (BTW, the Imperial system is not used in the U.S., only in the U.K. and the pre-metric system for the Commonwealth realms.)SirChan (talk) 02:02, 31 March 2008 (UTC)


 * I am talking about the rule as well. The rule is not "Byzantine" if it makes sense to use in order to optimize conditions in American-related articles for American readers. See the article about United States customary units, "Imperial units" is often used to describe US units. It is not used in the same context as English units, but it is still correct. -- Rai - me  02:14, 31 March 2008 (UTC)

←outdent← I'd been meaning to put my thripence in but Gene's pretty much stated my main point. Currently we've got the following. I've replaced the bullet points with numbers for ease of reference. Suppose we replace all of that with "Original units should be first." What have we not covered? The choice would be arbitary if the source(s) give(s) both metric/SI and imperial/US/other. Consistency within an article is a minor concern when stacked up against fidelity to the sources. If the order is changed, this should be noted as a footnote, unless the original was a rough approximation to start with. Jɪmp 07:41, 31 March 2008 (UTC)
 * 1) For US-related articles, the main units are US units; for example, 23 miles (37 km).
 * 2) For UK-related, the main units are either metric or imperial (consistently within an article).
 * 3) For other country-related articles, the main units are metric; for example, 37 kilometres (23 mi).
 * 4) American English spells metric units with final -er (kilometer); in all other varieties of English, including Canadian English, -re is used (kilometre).
 * 5) In scientific articles, use the units employed in the current scientific literature on that topic. This will usually be SI, but not always; for example, natural units are often used in relativistic and quantum physics, and Hubble's constant should be quoted in its most common unit of (km/s)/Mpc rather than its SI unit of s−1.
 * 6) 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.
 * 1) For US-related articles, the original units will generally be units US units.
 * 2) For UK-related articles, the original units will generally be either metric or imperial units.
 * 3) For other country-related articles, the original units will generally be metric.
 * 4) How a unit is spelt in this or that dialect is irrelevant here.
 * 5) In scientific articles, the original units will be use the units employed in the current scientific literature on that topic.
 * 6) If editors cannot agree on the sequence of units, the original units should be first.
 * the case where the choice of units is arbitrary
 * consistency within an article
 * Regarding Option 2 (UK Articles), an American or anyone else who reads a UK-centric article that's only in Imperial units is going to be a useless article if the metric value isn't included in parenthesis. The site is worldwide, so the information must be available to the widest possible audience. This Babel of units narrows audiences to a specific group and restricts the flow of information. SirChan (talk) 06:22, 1 April 2008 (UTC)
 * Similarly, an article with only US units is going to be useless to a non-American. In fact, as an Australian, I have a better feel for imperial units than I do for US units.  Conversions to metric should generally be given. J Ѧ ρ 08:07, 7 April 2008 (UTC)