Wikipedia talk:Manual of Style/Dates and numbers/Archive 111

Concealed links
The music project says:
 * Do not use piped links to years in music (e.g., do not write: The Beatles Please Please Me came out in 1963 ). Instead, sparingly use parentheses after years mentioned in the article, such as The Beatles released Please Please Me in 1963 (see 1963 in music ).

The film project referred to the music guideline and suggests formulas like:
 * Foo is an American political thriller that was released in 2001 and starred...
 * Snatch is a 2000 film ...

Now the aircraft project is discussing the same issue at: Wikipedia_talk:WikiProject_Aircraft. The music project and film project suggestions are not specific to just music or films. I think they have good suggestions that MOSNUM readers may find useful as options. Does anybody agree? Lightmouse (talk) 20:01, 5 September 2008 (UTC)


 * I'd be pleased for advice to be written into MOSNUM that the practice of piping to what looks like a single year generally be avoided. A few examples of how such pipes can be reworded might be provided, so that editors see that there are more skilful ways of linking. I'd also like MOSNUM to point out that it's unnecessary to link more than one or two "year in blah"s, since all years in blah can be accessed through just one. Tony   (talk)  04:43, 6 September 2008 (UTC)

A similar issue is present in Brian Boru, where years link to YYY in Ireland. I think it would be better to write in a style that gives readers a better idea of what they will find when they click on the link. --Gerry Ashton (talk) 20:26, 5 September 2008 (UTC)


 * A good general rule is to make sure a masked link includes more than one word or term in the text, so that users have some reason to believe they're not just going to be transported to the article 963 or whatever. Septentrionalis PMAnderson 22:34, 5 September 2008 (UTC)

Good point, Anderson. Maybe that should be included as advice in MOSNUM. Tony  (talk)  04:39, 6 September 2008 (UTC)


 * Strong support, as it were. I do this all the time (see my edits last night/this morning (depending on your time zone) to Steve Davis for an example. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93;  ‹(-¿-)› 21:29, 6 September 2008 (UTC)

Would anyone like to add this now? Lightmouse (talk) 09:54, 13 September 2008 (UTC)

Support Tony's, Gerry's, and Pmanderson's suggestions. Teemu Leisti (talk) 05:04, 14 September 2008 (UTC)

Text formatting math section merge proposal
Manual of Style (text formatting) serves no purpose at all, as it is simply rehash of WP:MOS (and in one place WP:MOSNUM). To the extent it may say anything distinctive that point should be added to MOS/MOSNUM, but otherwise this is just a blank-and-redirect-to-MOS. See also the closely related discussion at WT:MOS. The merge-from page is inconsistent on many points with both target pages, and its talk page is evidence of a great deal of confusion being sown among editors as a result of this break-away "guidelines existence. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93; ‹(-¿-)› 05:31, 4 September 2008 (UTC)

PS: I have edited a few bits of it to comply better with MOS/MOSNUM, but much of it is still messed up. There are probably a few points in it not presently in either of the controlling guidelines (which is why I suggested merges instead of just wiping it). One of the most important of these (one that I just added) is that variables should be marked up with  ( variable ), not   (variable). They both will typically visually render the same (depending upon user-side CSS), but the former actually has a semantic meaning, while the latter is just presentational hooey (notably, by the time it hits the user's browser, MediaWiki has converted the latter into, not  , because it has zero meaning at all from a content/semantics point of view). As with much else in MOS*, the average editor will ignore it and do what is convenient, but math editors (definitely not me) and cleanup gnomes (definitely me) should get this right. —  SMcCandlish  &#91;talk&#93; &#91;cont&#93; ‹(-¿-)› 06:32, 4 September 2008 (UTC)


 * Any objections? —  SMcCandlish  &#91;talk&#93; &#91;cont&#93; ‹(-¿-)› 21:59, 6 September 2008 (UTC)


 * It doesn't look like it. I support the merge, and so do all others who have commented on the MOS talk page. Teemu Leisti (talk) 06:09, 7 September 2008 (UTC)
 * Moving the provision here seems reasonable. I have seen a good many math articles, and I strongly expect that any requirement of will be ignored; math editors will contine to use  in-line and $$$$ for full-line formulas as they always have. The chief purpose of wiki markup is'' presentational hooey; and any effort which depends on regularity in the production of such hooey is likely to fail. Septentrionalis PMAnderson 19:00, 12 September 2008 (UTC)

I object to this undiscussed change of a policy that belongs in Manual of Style (mathematics) and not in Manual of Style (dates and numbers) and I will revert it if I don't get a reply very very fast to my objection expressed on the manual's talk page. "SMcCandlish", you really need to take this discussion to Wikipedia talk:Manual of Style (mathematics) and also to Wikipedia talk:WikiProject Mathematics. It's already being discussed on the latter page. Michael Hardy (talk) 21:13, 12 September 2008 (UTC)

Archived /Date format as /Archive 110
I've done some archive-related changes in the last hour or so.

The newly created subpage "/Date format" had some discussions that were unrelated to the subject. Some of them seemed finished, with no updates in the last couple of days, so I moved them to /Archive 109. Some others had updates in the last couple of days, so I moved them back to this main talk page.

Even after these moves, "/Date format" remained huge. I thought it would be best to archive it as /Archive 110, even though the latest updates to some of the discussions were within the last couple of days. If anyone finds it necessary to continue on any of the individual discussions of that page, perhaps they could copy the text they wish to discuss back to this talk page. In any case, the poll among four choices had been closed, and the runoff poll hadn't been updated for over a day.

D7 in the archives box on the right also points to /Archive 110.

In any case, it's perhaps better to just discuss things just on this main talk page, to avoid any "hiding" of discussions, as a couple of people have pointed out above. Teemu Leisti (talk) 17:11, 13 September 2008 (UTC)


 * Way to go. Multiple active discussions were just magically archived. Yet another reason most editors can't be bothered with this page. Gimmetrow 17:44, 13 September 2008 (UTC)


 * Not magically, but manually, and clearly noted here. And as I noted above, before I archived the page, I made sure to move discussions unrelated to the formatting of unlinked dates either to the previous archive (109), if they were inactive, or back to this talk page, if they had had updates in the last couple of days. The discussions now on /Archive 110 (507 kB of it) are all concerned with the formatting of unlinked dates. The contents can be easily referred to by linking, or by copy-pasting text from there to here.
 * Another reason for archiving: I have noticed that people do not generally want to discuss a subject on a subpage, whether it's a subpage of this talk page, or the talk page of a subpage of the project page. So, it's better to move the old discussions to an archive rather than cause confusion about which page the discussion should be conducted on.
 * The discussion on the formatting of unlinked dates is still going on here on the main talk page (and might well go on for a few more years, it feels like). However, if anyone feels that I've unfairly archived a particular topic that should be here, please feel free to delete it from the archive and restore it here, or to ask me to do it.
 * By the way, I notice that (a) no one else is complaining about my archiving, and (b) quite a few editors are bothering with this page. Teemu Leisti (talk) 04:47, 14 September 2008 (UTC)

Proposal: E-WANDaF
I don't actually advocate this proposal, but since some people seem to want to do it, I wish to bring the issue to a head and accept or reject it.

Resolved

The community of English Wikipedia editors retroactively recognizes the all-numeric date format used by the date autoformatting software (as of 13 September 2008), and names it the English Wikipedia All Numeric Date Format (E-WANDaF).

The format consists of a four-digit year, between 0001 and 9999 inclusive, a hyphen-minus (Unicode hex 002D), a two digit month, a hyphen-minus, and a two digit day-of-month. If necessary, any element is padded on the left to make up the stated number of digits.

Example: 2008-09-12

The format may be used to represent dates in the Gregorian calendar, the Julian calendar (either the proleptic Julian calendar or the calendar as actually observed in Rome in 45 BC to AD 8 inclusive), or the old Roman calendar. If necessary, any conventional notation may be placed near the E-WANDaF date to indicate whether the date is AD or BC. The calendar used shall be deduced by the reader from the context of the article.

Please express your support or opposition below. --Gerry Ashton (talk) 21:26, 13 September 2008 (UTC)
 * No.  Corvus cornix  talk  21:40, 13 September 2008 (UTC)
 * What do you mean by "may be used"? Note that JimWae explicitly only wants this format as an autoformatting preference, not in the source of articles. If and when a new autoformatting system is implemented, the inclusion or exclusion of this quasi-ISO format could be discussed, but otherwise I don't see the point (unless it's simply a POINT). -- Jao (talk) 21:53, 13 September 2008 (UTC)


 * By "may be used" I mean that a date in any of the mentioned calendars may be represented in this format under whatever circumstances this format is considered appropriate (perhaps on the screens of those who have expressed a preference for this this format in some future date autoformatting facility that is not deprecated, or as plain text in tables and other areas where space is at a premium). I also mean that the format may not be used for dates in calendars other than those listed, such as the traditional Chinese calendar. As for whether this should be supported in a new autoformatting system, such proposals do exist, and an acceptance or rejection of this proposal could serve as guidance for those seeking to advance those new autoformatting systems. --Gerry Ashton (talk) 22:15, 13 September 2008 (UTC)


 * By the way Jao, I certainly hope you didn't really mean, when you wrote "if and when a new autoformatting system is implemented, the inclusion or exclusion of this quasi-ISO format could be discussed" that a new system should be implemented FIRST, and THEN the format should be discussed; surely formats should be decided upon before implementation. --Gerry Ashton (talk) 22:25, 13 September 2008 (UTC)


 * No, I see I was imprecise on that point, but of course I meant after it has been decided to implement such a feature, not after it has actually been done. As for support or opposition of the clarified wording (that is, including "as plain text in tables and other areas where space is at premium"), that's an obvious oppose. I personally like the looks of yyyy-mm-dd in tables, but we probably shouldn't use it, and we definitely shouldn't use it where it's not actually ISO 8601. -- Jao (talk) 23:04, 13 September 2008 (UTC)


 * What is this talk of more techno-garble autoformatting in the future? The community is very happy, overall, without such. Tony   (talk)  05:00, 14 September 2008 (UTC)


 * While I'd be the last one to force any particular format on our readers, I'd like to note that the yyyy-mm-dd format has served me well in resolving a couple of date formatting disputes before. The articles in question were on Japanese subjects, and while none of the parties involved could be persuaded to give in to the US or International date format, they could be convinced to go with a subject-relevant third option. It's also a necessity in sortable tables, of course. --erachima talk 07:34, 14 September 2008 (UTC)
 * I thought the entire point of was so that numerical date formats would not be a necessity in sortable tables. -- Jao (talk) 10:41, 14 September 2008 (UTC)
 * Exactly. Teemu Leisti (talk) 21:10, 14 September 2008 (UTC)
 * The instructions for are deficient in that they don't explain what range of years is valid. The range is specified, I just didn't see it the first time I looked. --Gerry Ashton (talk) 21:41, 14 September 2008 (UTC)
 * All right, you lost me. The instructions say that any astronomical year number between -9999 and 9999 may be entered, and that the year number 5 with an additional BC parameter set may be used as an alternative to -4. What's the ambiguity? -- Jao (talk) 22:17, 14 September 2008 (UTC)

these polls
If we are to have any further attempts to measure opinions on this difficult issue of date format choice for articles related to non-English-speaking countries or to no country at all, I strongly feel that they should involve just two, maybe three, simple, narrow questions, each requiring just a yes or a no. The problem has been that fully-fledged proposals have been put; no wonder people baulk at having to declare a preference for a set of complex principles rather single issues.

On Pete Skyring's recent edit to MOSNUM: I wholly support Kotniski's reversion and edit summary. Your "country-driven" idea has been rejected, having been given a very reasonable airing. No one appreciates the disruption you are causing by slipping it back into the text. Tony  (talk)  13:01, 14 September 2008 (UTC)


 * How anyone can claim to be able to decipher a consensus for anything (related to date formats) in all the mess above and in the archives is beyond me. I, for one, certainly don't have the time to verify that what you say here is true.  Powers T 15:49, 14 September 2008 (UTC)


 * Tony, see my response to Kotniski above. And I take your point about the "national ties" wording being rejected, so I'm putting in the wording that actually won both the poll and the run-off. If any version has community support, this one does. I agree with you on the "simple questions" poll. Asking people to choose between "fully-fledged" versions is a big ask, especially when they aren't fully formed but whipped up on the fly for the purpose of putting them up as poll options. Much better to choose between a range of principles and then work on the precise wording as a community. --Pete (talk) 18:38, 14 September 2008 (UTC)
 * Please read WP:Polling is not a substitute for discussion. None of these polls is a vote; we are not majoritarian: it isn't "winning" that counts.


 * The runoff poll was between two options, A and C; neither of them the present text. By my count, there were 19 unvotes. Of those, 8 approved of C and disapproved of A; 2 approved weakly of both and preferred C to A; 4 approved A  and dispproved of C; 5 disapproved of both. That's 10  approvals at all, to 9 disapprovals, which is not consensus, even in our approximate usage. Both A and C were universally prescriptive; and neither is "widely approved of" as guidelines should be.


 * The present text is neither A nor C; it's what A and C have in common. So far it is widely approved of, and is much more popular than C; we will see. It may be possible to win a wider consensus by shortening it still further, but only if there is some interest in that by the Not Contents. Septentrionalis PMAnderson 19:22, 14 September 2008 (UTC)

Julian dates and templates
I have an interest in dates on Wikipeida, in relation to birth (and death) dates for hCard and start and end dates for hCalendar microformats.

Is it not possible that the existing date templates could be used,; modified so that, if a date before a certain point is entered, a prominent warning is generated, requiring a "calendar" flag be set, and, depending on the flag, the date be rendered as "DD MM YYYY (Gregorian) or "DD MM YYYY (Julian), using any DD-MM-YYYY order/format as suits the user? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 14:51, 10 September 2008 (UTC)


 * Not everyone edits articles with the edit box, so it wouldn't be possible to raise an interactive warning flag. All that could be done is to refuse to save edits that don't meet requirements. Also, indicating Gregorian or Julian with every date in an article would be excessive. A single statement of the convention followed in a certain article would suffice. --Gerry Ashton (talk) 15:03, 10 September 2008 (UTC)


 * There can be on-page warnings, like those generated, say, when coordinates are entered wrongly. Perhaps the indication could be shorter - say "DD MM YYYY (J)"? We don't seem to have a problem with repeating "BC" (or whatevr0 in articles. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 15:06, 10 September 2008 (UTC)
 * We probably should not; on such articles as Pericles, where there is no possible question on almost all of them, one BC per section or less would be enough. But it is sufficiently short and commonplace (and, above all, does not interrupt the syntax) that it is not deeply intrusive. Septentrionalis PMAnderson 15:15, 10 September 2008 (UTC)


 * Possible, yes: but that's in the sense that a Turing machine can perform any computable operation. Several widely supported requests have been made to the developers to change autoformatting in various ways; none have been acted on.
 * Prominent warnings can be produced without affecting the software, although one which is triggered by an unlinked July 15, 1581 would be difficult to imagine.
 * But is this desirable?
 * It would not affect dates between 1582 and 1752 in the English-speaking world, which are a large proportion of the problem, and which may actually need clarification.
 * It would not affect dates which mention only October 14 with the year understood,
 * On the other hand, it would force parenthetical explanations on every date in Battle of Hastings which mentions the year. This is bad writing, especially if there is a footnote explaining the calendar, as some articles have. Even without, who ever imagined that October 14, 1066 was not Julian? (Some readers have never considered the question, but they are merely confused; they will in all likelihood remain confused by the parenthesis.) Septentrionalis PMAnderson 15:10, 10 September 2008 (UTC)


 * "Several widely supported requests have been made to the developers to change autoformatting in various ways; none have been acted on." &mdash; for the record, I have acted upon every single technical request that has been pointed out to me, usually within a few hours. I have submitted three patches to modify the behavior of date autoformatting, all in response to requests made here or on the bugzilla site.  The WikiMedia system administrators (who are not properly called "developers" though most of them also happen to be developers as well) will no doubt be happy to put in place whatever patches the community decides are appropriate.  The problem is (and always has been) getting a broad consensus on what should be done.  Stop blaming the developers and/or system administrators for the inability of editors to come up with a sensible plan.  --UC_Bill (talk) 23:22, 10 September 2008 (UTC)
 * But why bother with all of this palahva? What is wrong with WYKIWIG (what you key-in is what you get)? Tony   (talk)  13:28, 15 September 2008 (UTC)

Using templates for dates and date ranges
In line with the concerns over mass-delinking of dates that would leave such dates difficult to relocated via computer searches (not impossible, just difficult), I really think we should consider the replacement of dates in articles with a templated version which provides many benefits. The template itself should be simple/stupid:, for example, as to make it easy to type as well. Use of the template is not required, but as shown by the benefits below, it can easily help an article maintain an article-consistent date format per MOS. An equivalent template can be made for date ranges. (Note that this is not an ISO date, the date is entered as described in the correct calendar format per the MOSNUM section).

The template would not have to link dates so we don't have over-linking, and the template would have a field for the date format specifier so that regardless of how it is determined what date format to use for an article, the template can put out dates in either format; such a format could also be easily changed in one shot in an article via automated tools like AWB and so forth (just by changing the template parameter in all dates in an article). No DA would be used at all, so the end page results are still the same for anon user and logged in user. The template can be used in main and footnote areas as to normalize the date format (the "cite" template family would need modification for this, but it needs modification anyway for date format equilivalence between text and footnotes). Bots and script tools that are already stripping dates can likely be easily modified to replace linked dates with the template version.

The key benefit is that very likely, if a DA solution is found that addresses all the concerns that others have against it (nonlinking result, anons are shown date format best suited to them (geographical-based or article-based), etc.) only the template has to be modified to bring in the DA solution. Now, there's a likelihood that a proper DA solution may not work with the template, but now we have the other benefit of using a template: a bot can go through and convert the templates to whatever format the DA needs.

The only drawback for this is that we will have a very widely used template, assuming full usage, at least 2 million times (once per page if not more). It would likely need full protection to prevent IP vandal harm, but that's not a huge concern. --M ASEM 13:26, 13 September 2008 (UTC)


 * Templates are useful in some cases, such as when displaying dates in sortable date columns; otherwise the "date" sorting would actually become alphabetical sorting. For instance, in a table in an article Schengen Agreement, I used the dts template in this manner:  . That produces "28 April 1995", doesn't produce a link, and sorts chronologically when used in a sortable table.


 * I understand that you're proposing a new, simpler-to-use date template, which doesn't take a "link" or "format" parameter, since it would never produce a link, and would format the date according to the MOSNUM. Is this correct?


 * I do see drawbacks to using this method: (a) It's harder to use it than just typing the date directly in the text. (b) Since no link is produced, nor does it have any other visible benefits, such as sortability in a table, I suspect many editors will not see the point of using it in normal text. (c) If a MOSNUM guide on date formatting that everyone agrees with ever arises, and it says that formatting depends on the article context&mdash;or heaven forbid, on what some earlier editor decided was the best date format&mdash;then the template will either have to have some sort of articifial intelligence, or a "format" parameter for outputting the date in other than the default format. Some of the simplicity is thus lost, and editing the use of the template takes as much effort as just editing a date expressed in normal text. (d) It will be hard to write a template that can cope with things like "the night of 21/22 September" or "July 15/25" (if giving a date in both Julian and Gregorian), and I suspect using such a template will be difficult, too.


 * In conclusion, I think it's a good idea to use date templates in tables, citations, and such; but we already have such templates,  and  . For dates in normal text, we should just write out the date without markup. I don't really see the extra effort involved in template-ifying every single date being worth it.


 * What would be a good idea, in my opinion, would be to make the  template produce by default no links and an output in the dmy format, making it unnecessary to specify the parameters  . Also, there should be another output format,   or such, that would produce standard month abbreviations in case of the longer months, so the date field wouldn't date up so much space in tables: for instance, "23 Sep. 1879" instead of "23 September 1879". Teemu Leisti (talk) 19:40, 14 September 2008 (UTC)


 * To address the drawbacks:
 * Yes, unfortunately it is harder to use, but, again, it would be optional, and only what would result with a bot-assisted DA-delinking manner. I would also believe that save for date ranges, as long as bare dates are in the two expected manners, they themselves can be converted to the template form via bot or script-assisted tools; and also vice versa.
 * Same issue; which is why I'm pointing out that this would typically be output from a bot or script that could be done on a cleanup pass or the like when reviewing an article. It may become second nature to some later.
 * I address this point already. The template right now needs a format parameter since it does not relay on DA.  Now, say I wrote an article, all dates with this template using intl format, and the MOSNUM decision decides that it should be US format for this page.  Changing this is only a matter of going to a regex-aware editor (AWB or the like), and swap the format parameter in each date, a very simple automated step compared to reversing day and month in every date.  (Yes, it would be nice if WP's template system allowed a page-persistent parameter, so that I could have a lead template set the format, and all others get from that, but I don't see that happening without a heavy dose of JS).
 * Date ranges (ignoring diffent calender aspects) are not difficult and only require a few more #if checks to reduce characters produced. The calender differences is a tougher solution, and in such cases, may be better just written out directly, or at least see how many such instances are around to warrant a template need.  (Technically, it is possible that a template could be written to spit out a date in the G/J calanders - just a bunch of #if statements again, and then add in the proper formatting.  Not sure how date ranges would then interfere with all that). --M ASEM  13:39, 15 September 2008 (UTC)

---

Could we have a date template which, rather than require the specification of YYYY|MM|DD and format as four separate parameters, instead simply wrapped around already fully MOS formatted dates.

The point of this would be for those users who have date preferences set up, so that:


 * If user preference was for International date format, then would render as 9 May 2001.


 * If user preference was set for US format, then would render as May 9, 2001.

- Ideally I'd like the template name "date" rather than "dte", but "date" is already taken with this buggy thing. Perhaps this could be resolved in future, but I'd encourage keeping things simple for now.

To clarify all the combinations, and to suggest keeping it simple, I'd recommend allowing only the following MOS formats in the template:

Users without set date preferences would of course just get what's in the template.

This way both the user's preferences and the editor's intent (with respect to MOS) would be taken into account.

--SallyScot (talk) 13:14, 15 September 2008 (UTC)
 * This would return us to the most unsatisfactory situation where it's almost impossible for WP to manage its date formats properly. Editors need to see what their readers see. And as has been pointed out above, it's a lot of work to key in, and not intuitive. Why all the trouble over whether month or day come first and second in order? Tony   (talk)  13:26, 15 September 2008 (UTC)

I don't follow the argument of how such a template would make it almost impossible for WP to manage its date formats. Could you explain? As I see it such a template's intent would simply be to give editors the option of catering for user date preferences. I'm not suggesting it ought to be mandatory, but I would ask why you'd deny such choice? It's not as if the question of why all the trouble over whether month or day comes first or second is a newly invented issue. The template would simply resolve linking context issues around the pre-existing date autoformatting approach. If anything, the pre-existence of that autoformatting approach, the fact that it was being used despite its linking issues, clearly indicates that some editors would like to be able to cater somehow for user date preferences. --SallyScot (talk) 14:42, 15 September 2008 (UTC)


 * Some, perhaps, but the vast majority used it (or had it inserted for them) just because it was the recommended thing to do at the time. Had there been a widespread desire to cater for readers' preferences, editors would have been banging on the devs' doors all these years asking them to implement something along those lines for IP users (which nearly all readers are).--Kotniski (talk) 18:24, 15 September 2008 (UTC)

Again, I don't think I quite follow the argument. I don't see how editors would've been asking for date autoformatting preferences for IP users to be taken into account, because I don't see how such requests would make sense. Maybe I'm missing something, but as I understand it, you'd need to be a registered user in order to have set up preferences. I think this is how most editors would understand the situation, and as such they wouldn't have been banging on devs' doors asking for the implementation of something that couldn't be done for IP users. --SallyScot (talk) 19:47, 15 September 2008 (UTC)


 * There are two ways that I could see IP users and users without a preference set handled. One is to support a per-article default. The other is to base it on the country associated with the IP address of the user. PaleAqua (talk) 20:02, 15 September 2008 (UTC)


 * What Tony meant is that if there's a way to format dates according to preference, then many editors will use it. I know I did, for four years. That's not a huge problem if there is a default format for unregistered users (possibly on a per-article basis), as PaleAqua suggests. But if "users without set date preferences would of course just get what's in the template" (just as they do in the present, deprecated, system), then we are in a lot of trouble. If an article contains the text "The meeting was held between and  " and we, the editors, see "The meeting was held between 9 May 2001 and 13 May 2001" (or the other way around), then who would fix it? Who would know there was even something that had to be fixed? That's exactly the mess where auto-formatting has left us in the first place. If somehow editors could be prohibited from using the feature, then catering for non-editor reader preference might be nice; but how many non-editors have a date preference set? And there would still be the date-range problem and the possibly needed extra comma after middle-endian dates to be addressed. And we would effectively have four date formats to choose between: now each article must either consistently use "9 May 2001", "May 9, 2001", "  " or "  ", as they would all be incompatible, any mixing leading to inconsistent results for at least some users. -- Jao (talk) 20:24, 15 September 2008 (UTC)

Date format preferences
Thanks for the clarifications here. I wasn't really considering the issue of article default date formats. I'm just making the point that IP users can't expect to have their own overriding date preferences without becoming registered users.

The date template I've suggested would simply resolve the linking context issues around the pre-existing approach to date autoformatting (now apparently deprecated). That pre-existing autoformatting approach was to cater for registered user date preference rather than article default date preferences.

Tony's point that "Editors need to see what their readers see" is at the expense of having user date preferences. It necessitates the removal of a long-established and pre-existing choice.

I would rather suggest that registered users (readers/editors) continue to be allowed this choice.

Perhaps some users would rather see dates in their preferred format. That is, rather than feel they're under any particular obligation to spend their time 'fixing' mixed date instances.

With it working this way, if an editor feels on the other hand that the issue of mixed dates formats is their overriding consideration, then they can choose to specify "No preference" as date format option on their preferences page.

Editors need to understand the implications of having their date preferences set, but they certainly ought not to be "prohibited from using the feature" as Jao suggests.

Another thought that occurs is that a date template may even be unnecessary for fully-formatted dates (ones that include day of month, month and year). It ought to be possible to recognise dates in article text; therefore it ought to be able to recognise and format 9 May 2001 and May 9, 2001 in accordance with user date preferences (i.e. without the need to wrap either full date format in a template in the article text).

--SallyScot (talk) 13:56, 20 September 2008 (UTC)

Final up/down vote on guideline for writing fixed-text dates
This poll runs for a week, and closes at 15:48, 20 September 2008 (UTC).

This is the final step in choosing the MOSNUM guideline to assist editors in determining the most suitable fixed-text date format to use in Wikipedia’s articles.

The results of the runoff poll, as of the vote by Pete, are as follows:

A = 1.31

C = 2.48

e = 1.48

So option C advances with a clear super-majority among the two options that laid out a specific proposal. Option e (each editor suggest an alternative solution) had a variety of opinions and no consistent theme. However, many editors voted with a non-zero value on e but did not add a ref-comment as to what they specifically had in mind. It seems that many treated e as “do nothing”, which would be applied as “keep the current wording.”

The next step is an up or down vote. Now that subsections on Talk:MOSNUM has been archived and the past voting moved to a subpage, we have freed up a lot of room here. The best venue for this next, critical step, is to keep this poll in this high-profile venue '' so the maximum number of editors can participate. '' One of the shortcomings of Wikipedia’s procedures is how controversial discussions have in the past been moved to remote backwater venues where it tends to drop off editors’ radar. That’s not good. The more controversial the issue, the more we need to foster the greatest participation by the Wikipedia community to ensure we are getting a good measure of the community’s mood. So let’s keep the voting here, well out in the open where the maximum number of editors can voice their opinion and discuss this matter.

Most everyone in the runoff poll did a great job of registering nuanced votes (a surprising number of 1, 2, and 3 votes), posted thoughtful and constructive vote comments, and debated without rancor. '''The general consensus in the previous voting, debate, and discussion was that option C was preferable to the other new options. But is option C better than what we currently have? Let’s see if we can push this to a natural conclusion and arrive at a general consensus now.'''

The options are as follows:

(C) Default to international unless U.S. and its territories—listed countries for editors’ convenience:


 * For articles on, or strongly associated with, the U.S. or its territories (or countries listed in this guideline that use U.S.-style dates: Micronesia and Palau), editors should use the U.S.-style date format (“February 2, 2008”), otherwise, editors should use the international date format (“2 February 2008”) in articles.
 * New articles on or strongly associated with Canada should use the international format but, for existing articles related to Canada, whichever format was used by the first major contributor shall be retained.

(R) Retain present guideline: (This is the current wording. There will be no editwarring on MOSNUM to fight battles over this poll.)
 * Articles on topics with strong ties to a particular English-speaking country should generally use the more common date format for that nation; articles related to Canada may use either format consistently.
 * If an article has evolved using predominantly one format, the whole article should conform to it, unless there are reasons for changing it on the basis of strong national ties to the topic.
 * In the early stages of writing an article, the date format chosen by the first major contributor to the article should be used, unless there is reason to change it on the basis of strong national ties to the topic. Where an article that is not a stub shows no clear sign of which format is used, the first person to insert a date is equivalent to "the first major contributor".

'''Is the proposed text better than what we have now? This is an up-or-down vote. No “0–4” values for voting; just an “X”.'''


 * Please cooperate. Do not add new columns and options to this table. We started with a four-option poll, then a three-option poll. A clear consensus can not be determined unless we have a simple two-option poll here. The question is whether or not to replace the current guideline on this issue with option C. If you have another idea, start your own poll. Alternatively, wait for the outcome of this poll and, if you don’t like what gets posted to MOSNUM, come here to this discussion page to rally editors to your way of thinking to go with your even-better idea.

This poll runs for a week, and closes at 15:48, 20 September 2008 (UTC).


 * Vote solicited by User:Skyring/Pete.

Vote statements
could just as well be formulated as. "
 * This might be too confusing though. Editors would probably misunderstand and do that literally.
 * In any case, these would all also cover the 1776 of the examples others have provided.
 * B) Not to muddy the issue, but something I haven't seen discussed is what happens when only one part of the date/date-range meets the criteria. Say, " " for Saint Val. Should then nothing be linked here (since the event is also referred to later in the text), or only "February 14" be linked, or the whole date, or both dates or what? -- Fullstop (talk) 07:47, 27 September 2008 (UTC)
 * B) Not to muddy the issue, but something I haven't seen discussed is what happens when only one part of the date/date-range meets the criteria. Say, " " for Saint Val. Should then nothing be linked here (since the event is also referred to later in the text), or only "February 14" be linked, or the whole date, or both dates or what? -- Fullstop (talk) 07:47, 27 September 2008 (UTC)