Wikipedia talk:Manual of Style/Dates and numbers

Years and dates archives

Archive D1 1.1 A new parallel syntax for autoformatting dates Archive D2 2.1 Date preferences and Jogersbot's massive undoing of edits Archive D3 3.1 What about date ranges? 3.2 UTC± Notation 3.3 Question 3.4 Clarification 3.5 Proposed 3.6 Wikipedia:Manual of Style (dates and numbers)#ISO date formats 3.7 AD/BC CE/BCE issues 3.8 Ranges of years 3.9 When did the year-without-date recommendations change? Archive D4 4.1 Comma in dates 4.2 Linking dates 4.3 Link only years? Archive D5 5.1 Default date display setting 5.2 Just following the rules, folks 5.3 Piped 'Year in xxx' links 5.4 Deleting year links 5.5 Year ranges 5.6 Wikipedia:Manual of Style (dates and numbers)#Years, decades, and centuries Archive D6 too many topics to link here, all related to date autoformatting, date autolinking, whether either of these should be used, or which date format to use in different contexts

Clarity on auto-formatting dates?

Centralized discussion

As suggested above, this discussion has been added to the {{Cent}} template in order to gather a wider consensus. This has been on ongoing issue for a long time now, and there have been several discussions on the matter which have been archived. I support the proposal to discourage the auto-formatting of dates, and in my experience when people have been given the explanations of the disadvantages of auto-formatting, they tend to side with discouraging its use. In my experience people only tend to be in favour of keeping auto-formatting when they are not fully aware of all the disadvantages. I do hope this proposal is carried forward and MOSNUM is re-written to clearly discourage the use of auto-formatting. SilkTork *^YES! 12:34, 9 August 2008 (UTC)

I also came here via notice from the {{Cent}} template. I've always used DA, as I figured it was the best way to avoid the wrong date format, because the dates would be displayed according to user preferences. However, I had not heard the valid argument that very few readers are actually registered users, or if they are registered users than they have not set their preferences (come to think of it, I waited well over one year to set it). Now, looking at the disadvantages (and embarassed that this thought escaped me: most readers of this encyclopedia are actually readers), I have to support the deprecation of DA. Lazulilasher (talk) 16:20, 9 August 2008 (UTC)

I must say that I am against the removal of links from articles and would be against the removal of the DA or any weakening of this in the MOS. Some articles that have had the links removed that I have looked I would put the links back on but that would I am sure would end in edit wars.

I would rather that we seek to have a software solution implemented that allows for users to see dates in the format they wish and that a default wiki-wide or a localised format be used to present dates to those who are not registered or those who are registered and have not set their preferences. Also the display of a link or not should be controlled by user-preference, and a link to the corresponding article only made when explicitly required. May-be this could be achieved by the use of extra mark-up. It would be good if a proposal could be put together for a software change that would satisfy all camps and that could be presented to the devs with a large backing that may get them to act. Keith D (talk) 17:30, 9 August 2008 (UTC)

I agree that the wanton removal of date links wouldn't be best. If it were possible to have the devs work on something that could automatically format dates according to a user's location, that would be fantastic (but, I know nothing about that aspect of the project). Maybe there's something that could be implemented based on a user's ip address (again, I don't know what I'm talking about -- just an idea). My point above was that requiring of auto-format dates is not something I support. I would support their deprecation for the time being, with the caveat that we not wholesale reformat articles (as you are correct, I could forsee many edit wars). Lazulilasher (talk) 18:01, 9 August 2008 (UTC)

The problem we run into is that there are two kinds of changes that can be made in Wikipedia: editor changes and developer changes – and there is no way to force the developers to make changes they’re disinclined to implement. A number of proposals have been made that would require developer implementation, but these have not been successful. That leaves editors with only the tools at their disposal: encouragement or deprecation of usages via MOS.

I think what would satisfy most editors in this issue would be for autodates to be displayed highlighted as “links” – whether by boldface, underlining, coloration or any combination of those. It would also be great if the date display format option could be added to the menu (under “Toolbox” perhaps) for both non-registered and registered users to choose from. Registered users would retain a default preference, but then they could switch back and forth more easily – which would remedy the problem that registered editors using a preference currently miss, namely, the ability to readily see what an article looks like to those selecting no preference. I don't know how easy or difficult this would be to implement, though. Askari Mark (Talk) 00:30, 10 August 2008 (UTC)

It may be useful if you could give links to the failed proposals so that we can see what has previously been proposed and why the developers declined to implement them. Keith D (talk) 17:45, 10 August 2008 (UTC)

It would indeed – and I would try to do it if I had several days to waste going back through several pages’ archives full of mile-long threads full of contentious (and often vicious) discussion like those currently on this page, going back some 3 years. Sorry. Askari Mark (Talk) 23:18, 10 August 2008 (UTC)

Here's a shortlist to review.LeadSongDog (talk) 14:04, 11 August 2008 (UTC)

Thanks, LeadSongDog! Just for reference, there have also been discussions on the VP, the main MOS, and elsewhere, as I recall. Askari Mark (Talk) 19:04, 16 August 2008 (UTC)

I think the changes to date have been positive but would oppose discouraging the use of autoformatting. This choice should be left to the editors of individual articles, who are in the best position to assess the pros/cons in specific cases. I would prefer MOS to remain silent on the issue and leave it as an optional style choice. Christopher Parham (talk) 05:23, 10 August 2008 (UTC)

I hope that date autoformatting will eventually go away. Scripts or bots that strip out autoformatting will probably cause ill will. So I suggest that autoformatting be handled like citation templates currently are. If an article has been systematically laid out using one style, don't change it without local consensus. I would not mind if WP:MOS discouraged date autoformatting, but any removal should be gradual and should respect local sensitivities. EdJohnston (talk) 20:14, 19 August 2008 (UTC)

No consensus

The main advantage of seeking consensus regarding any issue is that in doing so, many alternatives get discussed. Have many alternatives to date-linking as a means of automatic date formatting been discussed? Where? Straw arguments like, "IP users or users who haven't set their prefs see a mish-mash" should be countered with, "Let's assign (even at random) a pref for these users so they see consistent dates." If that discussion hasn't already taken place, there can't be a true consensus on de-linking dates. (sdsds - talk) 18:07, 9 August 2008 (UTC)

Tony, I think this is now the right approach, as the topic has generated a lot of interest. I agree with you that a straight forward easy to read date system makes the most sense, that is why I also favour a written out date rather than the ISO dating presently in use. FWiW, I have now begun to re-edit some of the older articles I have authored and then lack of autoformatting of dates is seen in all of my newest efforts So far, no one has complained. Bzuk (talk) 18:16, 9 August 2008 (UTC).

FWIW, as a result of this discussion I have now redoubled my efforts to use only wikilinked ISO dates in articles which I start or to which I contribute. (sdsds - talk) 23:16, 9 August 2008 (UTC)

On Wikipedia, "no consensus" usually means "no change from status quo", so since linking all full dates has a long history of being the de facto date formatting standard on Wikipedia, this discussion doesn't seem to give anyone justification to start (especially wholsesale) de-linking full dates. Shawisland (talk) 00:38, 10 August 2008 (UTC)

On wikipedia "no consensus" usually means "let's put up up with the current shit because I don't understand why anyone would want to change this best of all possible worlds". As in this case of date autoformatting. --Malleus Fatuorum (talk) 00:47, 10 August 2008 (UTC)

Exactly: Shawisland assumes that "be bold" consensus is necessarily inferior to "keep it this way forever" consensus. Tony (talk) 00:45, 11 August 2008 (UTC)

Sdsds, please see MOSNUM's deprecation of ISO 8601 dates. Tony (talk) 00:47, 11 August 2008 (UTC)

Tony, is that an invitation to begin a discussion about why wikilinked ISO dates should become the new standard for the English language wikipedia? The guideline's assertion that they are "uncommon" in English prose assumes they will be presented to readers in that format. So perhaps un-wikilinked use of ISO dates might currently be "deprecated," although that term isn't used in the guideline. It is clear when they are wikilinked, it is only because enwiki chooses it to be so that they are shown in that format to readers who haven't expressed a preference for it. The obvious solution is to choose a format into which wikilinked dates are consistently transformed for readers who have no expressed preference, but allow readers with a preference to have them consistently transformed into the format that makes sense to them. You are of course correct that some dates shouldn't be autoformatted. Those are the only ones that shouldn't be wikilinked. (sdsds - talk) 20:01, 11 August 2008 (UTC)

Consensus

My understanding of the situation, taking into account previous discussions on this and other talkpages, is that there is a consensus to discourage and or change the auto-formatting of dates - but there is no consensus to keep the existing situation. What we are looking for here is both a wider consensus and for that consensus to be grouped in one place so that there are no objections when the MOS is changed. Objections above turn to agreement when the matter is fully considered. I do recall some kind of list of signatures which it might be useful to find, which indicated concern with the autoformating system. To say that there is no consensus here is to misdirect people arriving to take part in the discussion. SilkTork *^YES! 09:31, 11 August 2008 (UTC)

Hi, thanks for this summary of previous discussions. Could you please indicate where those discussions can be read? I have no desire to "misdirect", and certainly apologize if I have inadvertently done so. I share the general concern about the current autoformatting system. Is there someplace a proposal like the one I have floated (choosing a format arbitrarily for readers who haven't expressed a preference) has been discussed. If that discussion hasn't taken place yet, it may be premature to assert a well-reasoned consensus has been reached on the question! (sdsds - talk) 20:06, 11 August 2008 (UTC)

I've responded on Sdsds's talk page with the link to the notorious Bugzilla page. Tony (talk) 05:28, 12 August 2008 (UTC)

Full date formatting

The major reason for insisting on two date formats is for compliance with autoformatting. If autoformatting is not going to be used then providing editors are consistent in an article, most of the other prescriptions on dates can go as well:

For example there is no reason why one should not write:

• The Battle of Hastings was fought on the 14th of October 1066.
• The Battle of Hastings was fought on October the 14th, 1066.
• The Battle of Hastings was fought on the 14^th of October 1066.
• The Battle of Hastings was fought on October the 14^th, 1066.
• The Battle of Hastings was fought on 1066-11-14.
• The Battle of Hastings was fought on 14/11/1066.

After all the arguments recently expressed on this page for removing autoformatting that readers can understand both the format "October 14, 1066" and the format "14 October, 1066" is equally true for the other formats. To accommodate the concerns those editors who do not agree, it could be suggested in this guideline, that a footnote explaining the format should be inserted next to the first date -- as is done with Old Style dates. --Philip Baird Shearer (talk) 10:01, 15 August 2008 (UTC)

The first ones to strike are #3 and #4; unnecessary superscripted text is discouraged because it makes reading harder and poses accessibility concerns. We then move on to Nos. 5 and 6, which are highly ambiguous because they do not spell out the month, causing confusion to European and North American readers alike (and yes, that includes the ISO format, with which not all readers are familiar—far from it). That leaves us with the first two, which constitute sloppy writing for reasons I am sure the honourable colleagues here can analyse better than I can. Even so, it's more or less common knowledge that these little words (the and of) fell in disuse quite a long time ago, and are hardly anywhere to be seen in written text, less so in a formal register like the one we use in an encyclopaedia. This doesn't apply exclusively to dates, but can also be seen in offices and titles like Director, Royal College, London. The ofs and thes are added mentally while reading such a title, but they don't have to be written, and are routinely omitted in text.

In any case, I don't find getting rid of auto-formatting a reason to abandon consistency in date-writing. Such consistency is one of the elements of good writing, and contributes to a more professional appearance for articles. Compromising this in order to introduce a needlessly complicated system, involving such additional clutter for intros like footnotes we could do without, does not sound like a good idea.

PS: You've slipped up: 14 October, 1066 is not a format—not with the comma, anyway. Waltham, The Duke of 10:22, 15 August 2008 (UTC)

So do you also object to "The Battle of Hastings was fought on 14th October 1066"? --Philip Baird Shearer (talk) 11:29, 15 August 2008 (UTC)

Predictably enough, I do. Waltham, The Duke of 13:42, 15 August 2008 (UTC)

Why? --Philip Baird Shearer (talk) 15:04, 15 August 2008 (UTC)

I have mentioned both ofs and thes, have I not? Waltham, The Duke of 21:59, 15 August 2008 (UTC)

• #1 and #2 are highly formal; but that may be desirable on occasion: either would do well in the lead, and the form works well in the colophon ("The chief effect of Harald Hardraadi's long and turbulent life was felt after his death in a place he had never been: the Battle of Hastings, on the 14th of October 1066.") We do not, and should not, adhere to a single monotonous level, imitating the worst of the Britannica's writing.
• We are not going to be consistent in date-writing. To say nothing of the practical matter that most articles will not comply with this page or care about it, we will encourage the use of two styles, radically and obviously different. Why jib at four?
• It would be worth warning editors that 14^th is often seen as old-fashioned; that "14th of October" is highly formal, and should only be used where the effect is intended; and that ISO can be confusing. Having deprecated them thus, is there any point to a prohibition? Septentrionalis PMAnderson 13:08, 15 August 2008 (UTC)

There are many ways to make an article boring, Mr Anderson, from which no variation in date formats can save it. There is engaging text, and there is annoying inconsistency; the former makes reading an article more worthwhile, while the latter simply distracts readers.

That we have not achieved consistency does not mean that we should not pursue it. We use two date formats because it is the minimum that can be achieved; each is the principal format for one of the two main components of our readership, and the two formats combined cover all English-speaking readers. We need no more formats.

Any style-related guidance of this kind is still a part of the Manual of Style, which is a guideline. It's optional to follow it, and using terms like "prohibition" creates the erroneous impression that it isn't. We don't block editors for using superscripted numbers. Waltham, The Duke of 13:42, 15 August 2008 (UTC)

I have always agreed that we should be consistent within an article. We will not be consistent between articles, and I would be surprised if a reader would be disconcerted by the change. He is not left with an unexplained inconsistency; she knows why it happened; xe clicked on a link. Septentrionalis PMAnderson 21:40, 15 August 2008 (UTC)

Let's drop the "little words" like "on" entirely:

• The Battle of Hastings was fought 1066-10-14.

Or alternately, let's allow editors to choose the number of "little words" as they see fit:

• The Battle of Hastings was fought on the fourteenth day of October in A.D. 1066.

These suggestions are not attempts at being silly. As we think about deprecating autoformatting, we really ought to think through the variations in formatting that will naturally arise (and be the subject of disputes). (sdsds - talk) 16:14, 15 August 2008 (UTC)

I support the latter. We are a collaboration; we should not go to enormous and fruitless effort to read like we aren't one. The wording of WP:ENGVAR could be usefully adapted, or we could just say leave well enough alone. Septentrionalis PMAnderson 21:40, 15 August 2008 (UTC)

We are a collaboration with the aim of taking articles to a specific standard of quality; the community has everything to do with the effort of taking the articles there, but this does not mean that the community should make its presence felt in the finished product. A featured article should avoid, if possible, every reference to itself, Wikipedia, and its community. Using writing formats incompatible with crisp, formal, encyclopaedic register just for the sake of doing so is ridiculous. And there need be no disputes if the guidelines of the Manual of Style are followed, guidelines the validity of which some people insist on undermining despite the need for them and the consensus upholding them. Waltham, The Duke of 21:59, 15 August 2008 (UTC)

I agree with the principle stated. None of these (first four, or ^th), however, is incompatible with a "crisp, formal, encyclopaedic register", and banning Ye Fourteeneth of Octobre is WP:BEANS. ;-> Septentrionalis PMAnderson 22:08, 15 August 2008 (UTC)

Hehe, good one. :-D But if all that the story's mother had said to her child was "behave yourself", he'd have fewer ideas to work with in the first place. We just need to state which styles we accept; leave it to the others to consider whether their own little ideas contradict the guideline or not. Waltham, The Duke of 22:30, 15 August 2008 (UTC)

No, what we need to do is treat our fellow editors like adults; say that this style and that one have disadvantages, and let them decide, based on what they happen to be writing, whether euphony or emphasis warrant the dangers. No one died and left us Jimbo's inheritance; to accept or refuse styles, or to belittle other editors' ideas. Septentrionalis PMAnderson 22:38, 15 August 2008 (UTC)

An orchestra cannot work without someone conducting, even with the mots experienced musicians. The same goes here: adults still need to be directed, and the greater the numbers, the more intense this need is. No matter how intelligent our editors are, without direction, chaos will ensue. (Actually, the more intelligent they are, the more bikesheds we'll end up quarrelling over.) Waltham, The Duke of 16:01, 18 August 2008 (UTC)

The first four examples above all seem quite compatible with crisp, formal, encyclopedic register. Depending on context, the two examples you disdain will often permit better-reading parallel structure. The example you give ("Director, Royal College, London") would be rather unusual in normal written prose; it would be "He was director of the Royal College" not "He was director, Royal College". Christopher Parham (talk) 22:42, 15 August 2008 (UTC)

It's so simple: "DD Month YYYY" or "Month DD, YYYY" (and maybe "YYYY-MM-DD" in certain specific cases) with the choice determined by ENGVAR-like rules. Nobody's belittling anyone: those who'd rather see "The Battle of Hastings was fought on the fourteenth day of October in the one thousand and sixty-sixth year of our LORD." are just as free as anyone else to come and see how their argument stands against the very reasonable call for consistancy such as that we see from Waltham. JIMp _talk·cont 02:50, 16 August 2008 (UTC)

Most solutions which begin with "it's so simple" are overlooking a large part of the problem. This is no exception. The four formats PBS began with are largely useful for special purposes; but they are useful for those purposes, and should not be deprecated for a spurious "consistancy". (Even ISO is sometimes useful; that's why it exists.) We should normally use the two formats Jimp prefers, but there is no reason to forbid the others. Septentrionalis PMAnderson 16:33, 16 August 2008 (UTC)

I don't understand this proposal; I thought we don't use ordinal suffixes on dates at all, so what is the proposal? SandyGeorgia (Talk) 03:08, 17 August 2008 (UTC)

Well, starting from the premise that consistancy is worth striving for where feasible it does become simple. Is the premise spurious? If so, much of the MOS can be deleted. JIMp _talk·cont 10:42, 17 August 2008 (UTC)

As I said at the start of this section: "The major reason for insisting on two date formats is for compliance with autoformatting. If autoformatting is not going to be used then providing editors are consistent in an article, most of the other prescriptions on dates can go as well." Your argument Jimp leads to total consistence across the project because it is feasible to implement such a style and spelling policy, but many would ague it is not desirable. --Philip Baird Shearer (talk) 10:48, 17 August 2008 (UTC)

Yeah, saw the claim at the top, repeated here ("The major reason for insisting on two date formats is for compliance with autoformatting.") I question that this was the case, Philip. And which came first, the chicken or the egg—the autoformatting mechanism that some developer cooked up or the rationale that the minimum number of formats the project needed to avoid significant disgruntlement was the big two, US and international? In any case, I think the project has moved on from whatever happened then. Dates, like spelling, seem to have happily evolved as a largely binary system in the main text. (Sure, the use of ISO dates in reference lists probably needs to be talked through over the next six to 12 months as we review the cohesion of the citation and other templates). But I think almost all WPians are content with US vs international dates, and would see no reason to engineer chaos in the absence of the double square brackets. Why would anything need to change? Tony (talk) 11:05, 17 August 2008 (UTC)

I don't see any advantage in letting loose and allowing whatever format whoever wants. Dates stripped of the thes, sts/nds/rds/ths and ofs are working fine. An argument for ISO dates is that they save space and are of consistant number of characters. An alternative we could use in lists/tables/etc. would be ordinar dates with the month abbreviated to three letters. Even the creators of autoformatting thought of this. JIMp _talk·cont 16:21, 17 August 2008 (UTC)

It is not our business to allow or disallow what our fellow editors choose to do; it is our business to advise what the consensus does do. If it were our business, it would be silly to issue an (unenforceable) blanket prohibition, ignoring all consideration of tone or euphony. Septentrionalis PMAnderson 20:29, 17 August 2008 (UTC)

Our business is to provide advice to those who should request it. If editors ask us "Should we use superscripted ths in text?", we advise them not to. If they don't like this advice, they shouldn't have asked for it, but they are free to ignore it. If they are unaware of the guideline, they are still free to continue their blissful ignorance. Terms like "blanket prohibition" are meaningless when a guideline is concerned, especially considering that "ignore all rules" is policy. Of course the individual needs of an article should be taken into account, but consistency has undeniable merits. This might be an international encyclopaedia, but it's still one encyclopaedia. It should also look like one. Waltham, The Duke of 16:01, 18 August 2008 (UTC)

Has anyone noticed how Wiki dates our contributions after the signature? Let's see... Saintmesmin (talk) 19:13, 21 August 2008 (UTC)

Proposal

As I have pointed out many times in the past, in spelled out dates, the distinction between the order being labelled by nationalities is at best weak, at worst fallacious. It has been suggested that the simplest solution is to adopt one preferred format fro spelled out dates, NN Month YYYY, or Month NN Year, it would be trivial to adopt this, and simple to impelment and probably not cause anyone any concern. Therefor in the next section I make such a proposal. Rich Farmbrough, 21:26 15 August 2008 (GMT).

Proposal

While any unambiguous date is acceptable fromeditors at large, the MoS should recommend either

NN month YYYY or where space is limited, the ISO format YYYY-MM-DD

Rich Farmbrough, 21:26 15 August 2008 (GMT).

No, we should not. We have already decided against this; it violates WP:ENGVAR. Septentrionalis PMAnderson 21:42, 15 August 2008 (UTC)

Why should it do that? To prevent American editors being confused by the international format? I think Tony has said many times that we're all quite able to understand either April 29, or 29 April. But what does 2008-03-04 mean? --Malleus Fatuorum (talk) 21:32, 15 August 2008 (UTC)

More likely to ensure that American readers are confused. Both will stand out like a sore thumb in a passage of idiomatic American. Septentrionalis PMAnderson 21:52, 15 August 2008 (UTC)

They certainly will, it's an absurd suggestion. --Malleus Fatuorum (talk) 21:56, 15 August 2008 (UTC)

Americans constitute an important part of our readership (and editing body), and their primary date format, unusual as it otherwise is, should be respected. (Besides, it's as clear as the international one, so there are no readability problems there.) I oppose this proposal. Waltham, The Duke of 22:05, 15 August 2008 (UTC)

I'm opposed to this idea. WP has matured to manage the largely binary system of spelling superbly well, with robust and workable guidelines. There is absolutely no reason that our guidelines for the (raw formatting of) binary date-formatting system don't work equally well. In both cases—lexicogrammatical and date formats—management of the binary system requires a little planning and checking. We should be planning and checking everything. Tony (talk) 23:53, 15 August 2008 (UTC)

I'm not sure that this an ideal solution. I agree with Tony that the dual spelling system works well. However, it should be pointed out that the US military uses dd Month yyyy. I'm also pretty sure that not even one US reader would be "confused" by that format. --Elliskev 01:13, 16 August 2008 (UTC)

Oppose—The adoption of one preferred format would by no means be trivial, simple to impelment or cause no concern. The fact is that there'll never be agreement over which this preferred format should be ... we might as well wait for the autoformatting to be fixed. JIMp _talk·cont 03:01, 16 August 2008 (UTC)

Thanks for your comment, Jim. May I point out that your last clause was intended to be ironic (cf, wait till the cows come home)? I had to read it twice to work it out. Tony (talk) 04:09, 16 August 2008 (UTC)

... home to the icy wastelands of Hades. JIMp _talk·cont 05:05, 16 August 2008 (UTC) ... singin' Dixie. JIMp _talk·cont 05:08, 16 August 2008 (UTC)

I think it should be as it is now - American subjects (or those where American date format is used, such as Canada, the Philippines, Taiwan etc) use American dates, subjects in countries which use international date format use international format, and with others where either is possible, consistency is to be observed regardless of the format adopted. As long as the full variety of the month is used, this should never be ambiguous, and should keep most readers happy. Orderinchaos 08:17, 16 August 2008 (UTC)

Whilst either format is acceptable - newspapers around the world tend to use American Dating and the world hasn't fallen apart (well, maybe it has, but not because of date formats) - editors are going to prefer one format over another and we would get some bitter disputes. As we have had in the past. The current system works, we manage inernationalisation issues in other things such as spelling and units of measurement well, and unless there's some technical solution on the horizon, we should keep things as they are. --Pete (talk) 11:33, 16 August 2008 (UTC)

The present arrangement is almost perfect and should not changed including the dualistic co-habitation of both date styles: In June 2005, the Arbitration Committee ruled that when either of two styles such as 14 February or February 14 is acceptable, it is inappropriate for an editor to change an article from one style to another unless there is a substantial reason to do so. Edit warring over optional styles is unacceptable. If an article has been stable in a given style, it should not be converted without a style-independent reason. Where in doubt, defer to the style used by the first major contributor. See: Wikipedia:Manual of Style (dates and numbers)--Thomaq (talk) 16:03, 16 August 2008 (UTC)

No, I disagree with this proposal. Long-standing convention on this page in this respect is fine. SandyGeorgia (Talk) 03:11, 17 August 2008 (UTC)

I'd like to point out that the problem would not lie with just American users being disconcerted with reading them. After all, they'll be inputting dates as well, and ISO or no ISO, it will be a natural habit for them to mistakenly enter the numerical digits in American date order, which will cause problems all around. Askari Mark (Talk) 18:19, 18 August 2008 (UTC)

It doesn't matter if people input dates in any format. The MoS is not a constraint on editors, it is designed to be something to which WP can move, giving a better result for all the reasons a MoS is usually used. Let me stress both formats are extensively used in most places, this has been portrayed as US vs international - it is not. Similarly with spelling and punctuation, we don't castigate editors or have flame wars over such things, just quietly and quickly bring them into line with the WP way. Rich Farmbrough, 15:37 20 August 2008 (GMT).

Another proposal

Reading the discussion above, it seems that some people think that (a) producing links from auto-formatted dates is bad because it results in lots of unneeded links from dates, and that (b) defaulting the autoformatting to ISO-standard is bad because that format is confusing to most readers.

I agree with both of these points. However, I think autoformatting dates is useful, so that people can see the dates in the format they prefer; and that for inputting the dates, the ISO standard is the best solution, since it makes dates easily searchable in articles.

Therefore I propose the following solution, which would (1) deprecate date-linking while (2) preserving auto-formatting and (3) making the default formatting understandable to the average person.

If the year is unnecessary or repetitive, it would also be possible for users to input only the month and the day-of-month. The output would, again, depend on the user's settings, and the default output for the non-logged-in would be readable.

However, it would be mandatory to input dates in ISO format, either as [[yyyy-mm-dd]] or as [[mm-dd]], depending on whether the year is specified or not.

What you must type	What logged-in registered users will see (settings on first row; other settings could also be available)						What others (i.e., not-logged-in and non-registered users) will see
	1999-12-31	1999 December 31	December 31, 1999	31 Dec 1999	31 December 1999	No preference
[[2012-02-29]]	2012-02-29	2012 February 29	February 29, 2012	29 Feb 2012	29 February 2012
[[02-29]]	02-29	February 29		29 Feb	29 February

(Table inspired by the one by Greg L in section "Autoformat quick-question" above.) Teemu Leisti (talk) 06:40, 17 August 2008 (UTC)

I don't think this would work... First of all, the ISO format is confusing for many people, and the fact that it would only be visible in the edit window does not help things. We call on our readers to be bold and edit, don't we? We don't want to confuse them even more than they already are. :-) And, in any case, much of our readership resides in the United States, and they wouldn't like seeing a default international date format.

All that said, even if it gains consensus as an idea, it remains unlikely to be applied. Apart from the clearly undesirable links, there are several issues with auto-formatting, including its inability to handle date ranges and slashed dates. We need more sophistication for anything like this to work. Waltham, The Duke of 09:05, 17 August 2008 (UTC)

I see your point. Just one quibble: the default display format in my proposal would be 29 December 2012, which should be understandable to everybody. Teemu Leisti (talk) 07:33, 18 August 2008 (UTC)

There's much talk of extending/improving the system. Here's the sorry truth. We can't fix the autoformatting. We've tried asking for it to be fixed on several occasions. The requests have been ignored. It appears that the creators of this mess believe it to be good. Here's a notion: if you want it fixed, why not boycot this autoformatting; that might get their attention? JIMp _talk·cont 11:56, 17 August 2008 (UTC)

Well, in that case there's really no sense in expending any more energy on this question. Teemu Leisti (talk) 07:33, 18 August 2008 (UTC)

Hey Teemu, thanks for your good work in preparing that case (I'll be needing to learn how to construct exactly that type of table, so this will be a model for me). As for the content, I'm sorry to say that I think it's too complicated and not intuitive—imagine having to teach every existing WP and every newbie how to key in and what is rendered. And again, there's no problem in the first place. Nothing beats WYKIWYG: What You Key in Is What You Get. Plain, intuitive, and simple, both for us and all of our readers. Tony (talk) 12:29, 17 August 2008 (UTC)

You're welcome. Perhaps you and Jimp are right, and we should just give up linking dates altogether. Teemu Leisti (talk) 07:33, 18 August 2008 (UTC)

As I've mentioned before, during the several times I've seen this debate rage, the general consensus has been that the best solution is one which the only the developers can implement. It sure would be informative to have one of them explain here why they cannot or will not. Askari Mark (Talk) 18:14, 18 August 2008 (UTC)

This proposal becomes a horrible mess for any date before 14 September 1752. --Gerry Ashton (talk) 18:27, 18 August 2008 (UTC)

Gregorian calendar

Dates like 1582-10-10 appear to be in ISO 8601 format. This is especially true for autoformatted dates, because the user preferences window indicates this format with the text "2001-01-15T16:12:34", which is unmistakably in the ISO 8601 format. The ISO 8601 format requires the use of the Gregorian calendar, and for dates before that calender was introduced, the proleptic Gregorian calendar. Browsing a few articles suggests that many editors do not understand this convention, and are therefore presenting incorrect dates to readers. Also, a date in a non-numeric format such as 10 October 1582 will generally be presumed to be in the Julian calendar (since it is before the introduction of the Gregorian calendar) so if it is autoformatted, the meaning of the date changes depending on the reader's preference setting.

Please note that discussions on this matter have been quite amicable and this RFC is only to attract a wider audience.

--Gerry Ashton (talk) 17:46, 18 August 2008 (UTC)

I have moved the recent addition:

Because readers who are logged in and have set a date and time preference to "2001-01-15T16:12:34" will see what seems to be an ISO 8601 format date, and because ISO 8601 format only uses the Gregorian calendar (or the proleptic Gregorian calendar before that calendar went into force in various areas), articles autoformatting any date before the adoption of the Gregorian calendar in the area discussed in the article should explicitly state that the proleptic Gregorian calendar is used in the article. A date in any calendar except the (proleptic) Gregorian calendar must not be autoformatted.

to here for further discussion, because it affects so many article and we should give careful consideration to whether this should be addressed separately from a general removal of links on dates --Philip Baird Shearer (talk) 11:55, 18 August 2008 (UTC)

---

The MOS contains the text:

• A date in any calendar except the (proleptic) Gregorian calendar must not be autoformatted.

Does that mean that autoformatting is not safe for dates prior to 1583? Lightmouse (talk) 08:56, 17 August 2008 (UTC)

The proleptic Gregorian calendar means this very thing: that it is used for dates in areas which did not have the Gregorian calendar then. The guideline refers to different calendars. But yes, editors should be more careful for these dates.

I suppose this complication is yet another reason why we would be better off without auto-formatting. Waltham, The Duke of 09:14, 17 August 2008 (UTC)

Hmm. This is worrying. Are you saying that all the current links for dates prior to 1583 are safe? It is just editors that create new links that need to watch out? Lightmouse (talk) 09:26, 17 August 2008 (UTC)

No; I cannot possibly claim that the Manual of Style is adhered to in every single article. Quite the contrary. What I am saying is that, as long as we use auto-formatting, it is not restricted to dates after 1583. If by "safe" you mean "foolproof", then yes, it is unsafe to use auto-formatting for dates in a calendar other than the Gregorian. But again, I don't think anything is foolproof in Wikipedia. (And we're aiming to remove auto-formatting anyway; we might as well start from the more dubious dates.) Waltham, The Duke of 10:28, 17 August 2008 (UTC)

I agree. There are general problems with autoformatting that apply to all dates. I was not aware until now that dates prior to 1583 have a specific error condition that will almost certainly remain uncorrected. I propose that the current text:

• Because readers who are logged in and have set a date and time preference to "2001-01-15T16:12:34" will see what seems to be an ISO 8601 format date, and because ISO 8601 format only uses the Gregorian calendar (or the proleptic Gregorian calendar before that calendar went into force in various areas), articles autoformatting any date before the adoption of the Gregorian calendar in the area discussed in the article should explicitly state that the proleptic Gregorian calendar is used in the article. A date in any calendar except the (proleptic) Gregorian calendar must not be autoformatted.

is changed to:

• Dates prior to 1583 must not be autoformatted. This is because autoformatting of dates that are not Gregorian or proleptic Gregorian can cause an error that is almost impossible to detect.

Lightmouse (talk) 10:41, 17 August 2008 (UTC)

The guideline is wrong, as is the proposal. For example William Shakespeare died on April 23 1616. If what you are saying is true then nearly every date article on Wikipedia will need changing, and the dates of events in the general articles will have different dates from those in the date articles. It will also effect categories for births and deaths of those born around January 1. Indeed for some years Christmas day will not be in the right year let alone on the right day.--Philip Baird Shearer (talk) 11:02, 17 August 2008 (UTC)

I think it was Gerry Ashton who inserted this yesterday. Gerry, can you assist in this discussion? Tony (talk) 11:08, 17 August 2008 (UTC)

The problem, which is what I had in mind when I edited the guideline, is that a reader is apt to presume that a date that looks like 1582-10-04 is an ISO 8601 date, especially after reading this guideline and viewing the date and time preferences window. Since ISO 8601 dates by definition use the (proleptic) Gregorian calender, 1582-10-04 is 11 days before the introduction of the Gregorian calendar, and would have been designated 24 September 1582 on the Julian calender which was in use on that day. Because the potential for errors is so great, and because readers who have chosen the ISO format have a definite statement about the calendar just by the format of the date, which other readers lack, it is wise to confirm to the ISO format readers, and to inform the other readers, that the proleptic Gregorian calendar is in use.

I am not aware of any fault in the automatic formatting that causes dates before 15 October 1582 to be displayed incorrectly.

Since ISO 8601 only uses the Gregorian format, any attempt to use that format and at the same time state that some other calendar is in use is an unacceptable contradiction.

I'm cheap, so I have not purchased the official version of the ISO 8604 standard, but here is a quote from a version that is on the web:

The introduction of the Gregorian calendar included the cancellation of the accumulated inaccuracies of the Julian calendar. However, no dates shall be inserted or deleted when determining dates in the proleptic Gregorian calendar. NOTE In the proleptic Gregorian calendar, the calendar year [0000] is a leap year. EXAMPLE The Gregorian calendar was introduced on 15 October 1582. In the calendar set by this standard the calendar day preceding that calendar day is referred to as 14 October 1582. In the Julian calendar that calendar day is referred to as 4 October 1582.

--Gerry Ashton (talk) 18:06, 17 August 2008 (UTC)

Philip, I'm not sure what you mean by "if what you are saying is true then nearly every date article on Wikipedia will need changing". Could you explain? What is a "date article"? --Gerry Ashton (talk) 20:06, 17 August 2008 (UTC)

By date article I mean under the article "date day" things that happened on that day will need adjusting if we are to say that such and such an event happened on that day. For example what was the date of the Battle of Agincourt? The article says "25 October 1415", as does the day date of October 25. We should not adjust such a date to the Gregorian calendar --Philip Baird Shearer (talk) 22:18, 17 August 2008 (UTC)

More to the point, because there is ambiguity, articles should say which calendar dates are expressed in. We must not assume that readers are familiar with MOSNUM.LeadSongDog (talk) 13:16, 18 August 2008 (UTC)

a reader is apt to presume that a date that looks like 1582-10-04 is an ISO 8601 date, especially after reading this guideline and viewing the date and time preferences window - I'd say that less than 1 in 1000 readers knows what an ISO 8601 date is, and less than 1 in a million readers will have read this guideline. So let's not do anything based on what readers know about either of these things - in fact, I'd give long odds that the vast majority of readers have no idea what the difference is between Julian and Gregorian calendars (and yes, that is certainly covered in one or more classes in high school, or earlier, but that doesn't mean that people remember what was taught). -- John Broughton (♫♫) 20:01, 19 August 2008 (UTC)

Random832 asked at the Village pump (policy) page "for what years should the Julian calendar be applied? Between 1582 and 1923 it is ambiguous." It depends on the subject matter of the article, but whatever is done should be explicitly stated in the article. If it were me, and the article was not connected to any particular location, I would make the transition at 1582. --Gerry Ashton (talk) 18:22, 22 August 2008 (UTC)

Shakespeare: a particular error

When reading Wikipedia's William Shakespeare article with date/time preferences set to ISO 8601 format, we read:

William Shakespeare (baptised 1564-04-26 – died 1616-04-23)^[a]
.
.
.
a. ^ Dates use the Julian calendar, used in England throughout Shakespeare's lifespan. Under the Gregorian calendar, which was adopted in Catholic countries in 1582, Shakespeare died on May 3.[187]

So the format of the date proclaims it is in the Gregorian proleptic calendar, but the footnote proclaims it is in the Julian calendar, so the article contradicts itself. --Gerry Ashton (talk) 18:17, 17 August 2008 (UTC)

The article does not use 1564-04-26 it uses "26 April 1564 – died 23 April 1616". If one sets date format in one's preferences one will see "baptised 1564-04-26 – died 1616-04-23" but there is nothing to say that that is a Gregorian date and the footnote makes it clear that Julian dates are being used -- as is the norm for dates in England before the Calendar (New Style) Act 1750 and the changes bought in in 1752.

Gerry Ashton, you are making the assumption that is not born out in fact of how dates are entered in Wikipeida. If an editor enters the date as 1616-04-23 and the event took place in England they are assuming the Julian calendar not the Gregorian calendar date, and the link goes to the common English events of that date one of which is the death of Shakespeare. This is made clear in this guideline in the section Calendars. There is nothing in the preferences that says that the style yyyy-mm-dd is an ISO 8601 format it only says:

• No preference
• 16:12, January 15, 2001
• 16:12, 15 January 2001
• 16:12, 2001 January 15
• 2001-01-15T16:12:34

To alter dates would be very time consuming and complicated, for example Charles I executed on 30 January the articles Charles I of England, High Court of Justice for the trial of Charles I and the day say January 30? The Battle of Hastings says October 14, as does the article October 14. If we were to alter this Wikipeda would be at odds with just about every secondary source that is published on these topics. And No I see no problem with autoformatting to October 14 from the Battle of Hastings article any more than November 5 from the Glorious Revolution (An auspicious date for the landings because of the Gunpowder Plot):

• 1605 - Gunpowder Plot: A plot led by Robert Catesby to blow up the English Houses of Parliament is thwarted when Sir Thomas Knyvet, a justice of the peace, finds Guy Fawkes in a cellar below the Parliament building.
• 1688 - Glorious Revolution begins: William of Orange lands at Brixham.

Or October 25 to the 1917 October Revolution. Much easier to add a warning to those using the format YYY-MM-DD in this guideline --Philip Baird Shearer (talk) 22:18, 17 August 2008 (UTC)

The association between dates in the format 2001-01-15 and ISO 8601 is quite strong, and the association between the format 2001-01-15T16:12:34 and ISO 8601 is so strong as to be practically unbreakable. Furthermore, many editors and readers seem to think such an association within Wikipedia does exist. If Wikipedia made a futile attempt to break the association by posting some text somewhere, it would amount to a truly idiotic and uncalled for novel standard. I suggest going through all the date articles with a bot and removing all the autoformatting. --Gerry Ashton (talk) 22:42, 17 August 2008 (UTC)

Also, I take the view that a date in an article is entirely wrong if it is wrong for any possible presentation using any possible user preference. If it is right for a user using no preference, and autoformatted into an error for a user with the 2001-01-15T16:12:34 set, it is wrong. --Gerry Ashton (talk) 22:45, 17 August 2008 (UTC)

There seems to be a conflict of what is expected of autoformatting:
1 - That it correctly interprets and formats the date, which is NEVER going to happen. Any date from 1582 until the date of Gregorian switchover in the relevant locale, you'd also need to know which calendar (Gregorian or Julian), which implies a need to incorporate location into the date. And which date is being used for new year, 25 March or 1 January?
2 - That it formats a text string into the format preferred by the reader. While I can see the point that it's not going to help most readers, it will help registered users, those who are more likely to get into edit-wars over date format than casual readers. If autoformat forestalls edit wars I'm all for it. Bazj (talk) 23:57, 17 August 2008 (UTC)

Again, Bazj, you're promulgating this fear of edit-warring without evidence that it has already occurred or is likely to occur. Apart from the power of the ArbCom ruling, emblazoned at the top of MOSNUM and elsewhere, to lead to the quick, clean hosing down of any dispute that did occur, I put it to you that the community has moved on from such systemic edit wars. I hope this is not fearmongering! Tony (talk) 00:17, 18 August 2008 (UTC)

Bazj's post illustrates the danger of the passive voice when he writes "what is expected of autoformatting". Just who is doing the expecting? In the absence of an explicit statement, I assume that since we exist to serve the reader, the reader is doing the expecting. A reader who registers and sets a preference to 2001-01-15T16:12:34 will either not know what to expect (due to ignorance of ISO 8601, the finer points of the Gregorian calendar, or both) or will expect dates to be in the (proleptic) Gregorian calendar. Any date that does not satisfy that perfectly reasonable expectation is wrong. --Gerry Ashton (talk) 00:31, 18 August 2008 (UTC)

"Any date that does not satisfy that perfectly reasonable expectation is wrong." The date is not wrong -- it would be wrong (as it would not be supported by most reliable verifiable sources) to use Gregorian dates for British events that happened before 1752. Wikipedia is no alone in making such compromises. This NASA source say "All eclipse dates from 1582 Oct 15 onwards use the modern Gregorian calendar currently found throughout most of the world. The older Julian calendar is used for eclipse dates prior to 1582 Oct 04. Due to the Gregorian Calendar Reform, the day following 1582 Oct 04 (Julian calendar) is 1582 Oct 15 (Gregorian calendar). ... here are a number of ways to write the calendar date through variations in the order of day, month and year. The International Organization for Standardization's ISO 8601 advises a numeric date representation which organizes the elements from the largest to the smallest. The exact format is YYYY-MM-DD where YYYY is the calendar year, MM is the month of the year between 01 (January) and 12 (December), and DD is the day of the month between 01 and 31. For example, the 27th day of April in the year 1943 would then be expressed as 1943-04-27. We support the ISO convention but have replaced the month number with the 3-letter English abbreviation of the month name for additional clarity. From the previous example, we express the date as 1943 Apr 27." Which is similar to the wording we use in this guideline. BTW that NASA source links to a page called Calendar Dates that links to a Swiss site which includes this page that says: "The ISO 8601 date format may be used both with the Gregorian and with the Julian systems (and with many other calendar systems). Dates in the Julian calendar are marked with 'J', and those in the Gregorian calendar (when this is made explicit) are marked with 'G'." --Philip Baird Shearer (talk) 10:46, 18 August 2008 (UTC)

What I mean by Wikipedia had warnings about ISO is in this guideline:

• Dates "ISO 8601 dates (1976-05-31) are uncommon in English prose, and are generally not used in Wikipedia. However, they may be useful in long lists and tables for conciseness and ease of comparison." (also given in WP:MOS#Dates
• Calendar "Dates before the adoption of the Gregorian calendar on 15 October 1582 are normally given in the Julian calendar. The Julian day and month should not be converted to the Gregorian calendar, but the start of the Julian year should be assumed to be 1 January (see below for more details). ... "Dates of events in countries using the Gregorian calendar are given in the Gregorian calendar. This includes some of the Continent of Europe from 1582, the British Empire from 14 September 1752, and Russia from 14 February 1918 (see the Gregorian calendar article). ... The dating method used in a Wikipedia article should follow that used by reliable secondary sources. If the reliable secondary sources disagree, choose the most common used by reliable secondary sources and note the usage in a footnote."

I think that this gives sufficient warning (similar to the NASA eclipse web article above) that autoformatted ISO looking date strings will use both Julian and Gregorian date depending on what is the normal method of describing dates for a given topic. --Philip Baird Shearer (talk) 12:08, 18 August 2008 (UTC)

The Swiss web site mentioned by Philip is the personal web site of Peter Meyer. Ordinarily it would be considered unreliable, although I have seen it mentioned favorably in a number of places. In any case, he has no authority to change an ISO standard that explicitly says it uses only the Gregorian calendar. Furthermore, section 3.4 of the standard, "Characters used in the representations" does mention single letters with special meanings, like "T" and "Z", but does not mention Meyer's "J" or "G". The instant you put a "G" or a "J" onto what would otherwise be an ISO 8601 date, it becomes a Peter Meyer date.

Wikipedia does indeed say we should ordinarily use either the Julian or Gregorian calendar, whichever was in general use in the area described in the article. That's the right thing to do. (What to do in an area that used some other calendar is not so clear.) Wikipedia also strongly implies that dates in the YYYY-MM-DD format are ISO 8601 format, and even if we didn't imply it, readers would infer it anyway. Thus, every article that presents a date in the YYYY-MM-DD format is proclaiming that it is in the Gregorian calendar. If the article has a note to the contrary, the article contradicts itself.

We could attempt to create the "English Wikipedia Standard Date Format" and try to explain how it might be either Julian or Gregorian, depending on the country discussed, but that would be a terrible idea. It would be better to just not use the ISO 8601 format, or autoformatting, for any date preceeding 14 September 1725, as well as any date in the Julian calendar after that date. I'm starting to think this needs wider attention, perhaps at the Village Pump. --Gerry Ashton (talk) 17:25, 18 August 2008 (UTC)

Straw poll on all-numeric dates, meaning & warning

Since three days have passed since the last comment in this thread, I think it's time to sum up editor's conclusions.

The purpose of this straw poll is to find out if editors opinions on two questions:

1. Within the English Wikipedia, are all-numeric dates like "1582-10-09" governed by ISO-8601?
2. Should the Manual of style (dates and numbers) contain a prohibition against using all-numeric dates like "1582-10-09", or autolinking such dates, unless the date is in the Gregorian calendar or proleptic Gregorian calendar (except direct quotations, combined with appropriate explanation to the reader)?

---

Note that the below chart does not allow ~~~~ signatures to be used. You must copy/paste or hand-edit your signature. For assistance in writing your signature, you may copy the time below in red from the preview window after refreshing while in edit mode:

03:39, 30 August 2024 (UTC)

---

Y = yes, N = no

Editor	ISO 8601 governs dates like "1582-10-10"	Prohibit such dates & autolinking if not in (proleptic) Gregorian calendar
Gerry Ashton 23:52, 21 August 2008 (UTC)	Y	Y
User:Blank	?	?
User:Blank	?	?
User:Blank	?	?
User:Blank	?	?

Comments for all-numeric dates, meaning & warning straw poll

ISO 8601 is the best-known standard for this format, and there are many mentions within Wikipedia MOSs and cite templates of these dates being ISO or ISO 8601. --Gerry Ashton (talk) 23:58, 21 August 2008 (UTC)

Large numbers

According to Wikipedia:Manual_of_Style_(dates_and_numbers)#Large_numbers, commas are used to break the sequence every three places left of the decimal point; spaces or dots are never used in this role (2,900,000, not 2 900 000).

However, according to

• AIP [1],
• NIST [2],
• English Style Guide. A handbook for authors and translators in the European Commission [3] (page 26),

three−digit groups in numbers with more than four digits are separated by thin spaces instead of commas (for example, 299 792 458, not 299,792,458) to avoid confusion with the decimal marker in European literature.

Taking into account inherently international character of Wikipedia, should this rule be applied to MoS as well? --texnic (talk) 16:28, 17 August 2008 (UTC)

ISO 31-0 also refers to a 'small space' for that purpose. Lightmouse (talk) 16:58, 17 August 2008 (UTC)

I prefer the thin space but only if it vanishes on copy & paste, doesn't wrap, can be produced with a parser function, appears properly on the vast majority of machines and would be applied WP-wide. As for European literature, though, just about the only subset we need worry about is that from the British Isles, ie.e that in English, where we never use the dot as a thousands marker nor the comma as a decimal point. JIMp _talk·cont 17:44, 17 August 2008 (UTC)

This has been discussed at length in the past. No consensus coud be raised to use thin spaces. Even if there was the illusion of a consensus among the editors of this page, it would be unenforceable. --Gerry Ashton (talk) 18:23, 17 August 2008 (UTC)

Can we allow use of spaces as decimal separators without making it a strict rule? The point is that if one writes a new article and knows that there are the above mentioned recommendations and is accustomed to them, why should s/he still use the old-style notation? --texnic (talk) 15:13, 19 August 2008 (UTC)

Linking to years, again

I am having a disagreement with another editor over the implementation of the de-linking years in articles, in one specific context: in the lead of a biographical article, where the practice has been -- & AFAIK still is -- to link the years of the subject's birth & death. Even if the month & day are not known. As in this fictitious example:

Harvey J. Wallbanger (1940 - 1995) was an American activist for Alcoholics Anonymous.

Now I've looked through the MoS, & it does not offer any guidance about this specific use of link. Then again, I can't imagine that no one has considered this specific use of linking to years, & that a consensus has been stated somewhere. (My hope, obviously, is that the consensus is one that I would agree with.) Thanks for the pointer. -- llywrch (talk) 06:14, 18 August 2008 (UTC)

[WP:CONTEXT makes it quite clear: "In general, do not create links to [low added-value items] such as, 1995, 1980s, and 20th century." WP:MOSLINK and WP:MOS are consonant with this. Tony (talk) 09:03, 18 August 2008 (UTC) PS I can't locate this article you refer to. Tony (talk) 09:09, 18 August 2008 (UTC)

Either this person is fictitious, or there is an interesting story to find here; my search has yielded a cocktail with that name. :-) Waltham, The Duke of 11:54, 18 August 2008 (UTC)

I seem to remember this issue as partly about 'aesthetic linking'. People sometimes like to link date fragments because there is an associated full date with a link. I think some people like 'aesthetic linking' in tables and lists too.

• Harvey J. Wallbanger (1940 - 12 March 1995) was an American activist for Alcoholics Anonymous.
• Harvey J. Wallbanger (1940 - 12 March 1995) was an American activist for Alcoholics Anonymous.

I am not aware of any other reason and I think I would know if there were. Lightmouse (talk) 13:04, 18 August 2008 (UTC)

Perhaps a more interesting story is that of Harvey Wallbanger Jr, the famous racing buffalo. But we digress. Biographies normally use Template:Infobox Person with a nested template for tombstone dates. See Template:Infobox_Person#Example. They should be and usually are templated, so autoformat shouldn't kick in. LeadSongDog (talk) 15:38, 18 August 2008 (UTC)

Well, the templates produce autoformatted (i.e linked) dates, so I guess it does kick in. When date linking is finally deprecated, these date templates will hopefully be changed to remove the linking.--Kotniski (talk) 16:44, 18 August 2008 (UTC)

The above was meant only as an example, not as the article in question. (Please note my words "as in this fictitious example". I am amazed none of you noticed those words, & will refrain from any further observation or comments on it.) Either the practice of linking birth & death dates -- even if only the year is known -- is permitted in all lead paragraphs, nor none. -- llywrch (talk) 16:08, 18 August 2008 (UTC)

llywrch, please note that the comments above were intended as jests, not to be taken seriously. FWiW Bzuk (talk) 16:23, 18 August 2008 (UTC).

I don't care. Good bye. -- llywrch (talk) 21:37, 19 August 2008 (UTC)

Again calling for date linking to be deprecated

Resolved

Although the discussion always seems to go off at various tangents when the issue is raised (like when I raised it above), I still see no argument of any weight against deprecating the linking of dates (except in special cases), and plenty of arguments (and apparent consensus) in favour. If no-one can come up with a valid counterargument, I'm going to start feeling bold and editing the guideline accordingly.--Kotniski (talk) 10:57, 16 August 2008 (UTC)

I do not think we are anywhere near the stage of changing the guidelines as yet there is definitely no agreement as far as I can see to depreciate the use of date linking. It looks like from the arguments that we are heading for a do it on an article by article basis. I still think that the way forward is a software change that would enable dates in articles to be put in which ever format people want and then for them to be formatted for display according to user preference or according to browser preference for unregistered users. The constant changing of dates by people as has been noted above causes a significant number of history revisions on articles, this will continue as people will put the dating in in their normal manner no matter what we prescribe here. By changing the software to correctly do the actual formatting for the user means that the format in the article source is irrelevant. This will cut out the need for changes to articles where the only change is from '2 August' to 'August 2'. Those that want to use ISO dates can do as people will not see them as such unless they chose that preference or have a browser set for that preference. The software could also be changed to display or not display the links to the date articles and so solve the problem of too many blue links that some people have. Keith D (talk) 11:29, 16 August 2008 (UTC)

I see no earthly point in doing it on an article-by-article basis. I mean the linking; obviously the style of date will be consistent only at article level. But one we've decided (as I believe we have) that there is no virtue and much vice in linking dates for the sole purpose of making autoformatting work, then this surely applies to one article as much as to any other. So let's change the guidelines so that edits which remove date links (which nearly everyone agrees are improvements) are not detrimentally reverted on ENGVAR-type grounds. What the developers may do with the software in the future is not foreseeable enough to be of much concern to us - personally I hope they don't waste any of their valuable time over this issue when there are many far more significant improvements waiting to be done.--Kotniski (talk) 12:29, 16 August 2008 (UTC)

Well, predictably, I fully support Kotniski's call. However, I don't think reversions are on any rationale basis, let alone ENGVAR. Perhaps leave a little room for autoformatting while generally recommending it not be used? That might save angst in quarters where folk will become very upset at losing their blue-splotch dates. Tony (talk) 13:15, 16 August 2008 (UTC)

Excellent idea Keith! Please go ahead and change the software. We're not heading for an article-by-article basis: this is where we're at. The delinking being done on this basis has been generally met with a positive response. The arguments for it are strong. Those not in favour of depreciating autoformatting seem to think it can be fixed. If you've been following the issue, you'd understand why those in favour of depreciation have given up hope in having it fixed. We're everywhere near the stage of changing the guidelines. Let them be changed but heed Tony's call about allowing a little room to move (at least for now) for those who still think it's a good thing. JIMp _talk·cont 14:00, 16 August 2008 (UTC)

Not sure I understand what is envisaged by this "a little room to move" - can you or Tony be more specific? --Kotniski (talk) 15:15, 16 August 2008 (UTC)

Well, that could be achieved in a number of wordings; for example, by including the word "generally", or "unless there is good reason to use it", or some such. This would soften what might otherwise be an imperative, and would avoid suddenly turning the vast majority of articles into MOS breaches while at the same time legitimising the change-over. Tony (talk) 15:33, 16 August 2008 (UTC)

Although there are still some "pockets of resistance", I wonder if Tony1's reasoned arguments behind the change can be posted somewhere as "pointer" to the direction being advocated. FWiW Bzuk (talk) 15:39, 16 August 2008 (UTC).

• Jimp, Tony: I think you’ve got some inertia going here in identifying what needs to be done and are doing good work in finally fixing MOSNUM. Loose the links to random trivia that is usually beyond tangential to the articles. Formatting by itself (without linking) seems perfectly workable tool when employed properly under current guidelines—but with one caveat: Except for the [[2005-06-08]] method, which isn’t for producing ISO formats, but was intended to take an ISO-formatted input with which to output dates with spelled-out month names *but it only does so for registered edtiors* (terribly unwise since 99.9% of readers see numeric-only dates). I suggest a bot be made that goes in to articles, looks for any other hard-coded dates to see how they are formated (14 Feb or Feb, 14), and which then converts the ISO-input dates to hard-coded ones. Greg L (talk) 20:54, 16 August 2008 (UTC)

Earlier on this page SandyGeorgia had indicated some formatting problems in the citation template that precluded the use of any other format than ISO dating, which is still an prospect of take this information "with a grain of salt" as previous efforts to have the templates adjusted have not been well received. An attempt to have MLA-style templates was simarily rebuffed, so don't hold your breath. FWiW, now, have I mixed enough metaphors in one statement? Bzuk (talk) 21:48, 16 August 2008 (UTC).

• Disagree - The work being done, albeit well intentioned, to remove markup that indicates some text may be automatically formated, is a step in the wrong direction. What wikipedia (enwiki, and elsewhere) needs is more markup indicating text which can be autoformatted, not less! (sdsds - talk) 00:43, 17 August 2008 (UTC)

Sd is concerned that meta-info will be lost forever if we say good-bye to double square brackets in dates; I've explained on his/her talk page that identifying raw dates is a piece of cake for a script if need be in the future. I don't know how this point was lost. I don't think there will be much enthusiasm for more markup in search of a problem that doesn't exist. Tony (talk) 00:54, 17 August 2008 (UTC)

Here's the text: I think this is the insertion we're looking for (in italics) in the "Date autoformatting" section. It will promote the simplification of wikitext, an increase in the salience of our high-value links, and the reduction of colour-clutter in our article text. It's a good example where a straight move towards simplicity in process immediately improves the product.

A combination of day-number and month can be autoformatted by adding square brackets ([[5 November]] or [[November 5]]; [[5 November]] [[1989]] or [[November 5]], [[1989]]). The square brackets instruct the MediaWiki software to format the item according to the date preferences for registered users who have chosen a setting and are logged in. This should not generally be used unless there is a particular reason to do so. Careful consideration of the advantages and disadvantages of the autoformatting mechanism should be made before applying it: the mechanism does not work for the vast majority of readers, such as unregistered users and registered users who have not made a setting, and can affect readability and appearance if there are already numerous high-value links in the text.

I've also inserted the US style into the example, for users who do have "a particular reason to do so"; I don't know why that was ever left out. Tony (talk) 00:51, 17 August 2008 (UTC)

Looks good. SandyGeorgia (Talk) 03:12, 17 August 2008 (UTC)

I would oppose this change. Aside from the addition being obviously false at the moment, since autoformatting is the current dominant practice, it is a step in the wrong direction to legislate personal preference in this manner. Providing the facts neutrally and leaving editors to use their best judgment is ideal. Christopher Parham (talk) 03:39, 17 August 2008 (UTC)

Christopher, I'm willing to change "is not generally used" to "should not generally be used" if people wish it. WRT your claim that the wording "legislate[s] personal preference", I, for one, am thinking only of our readers at large: if it's my personal preference to present the best formatting and linking environment for them, yep, I'm guilty as proven. I suspect that the "personal preferences" of the huge number of people forming a consensus here also place the needs of our IP readers uppermost. (I don't doubt that the opposition of such a skilled editor as you is well-intentioned, however.) Tony (talk) 05:58, 17 August 2008 (UTC)

Strongly support the change (though perhaps with additional information for the uninitiated to explain why legacy breaches of the rule are to be widely found - although hopefully bots will soon be getting to work eliminating these). It is not personal preference - it reflects widespread consensus reached on the basis of very strong arguments. And "providing the facts neutrally" is not what MoS is about - it's here specifically to make recommendations in matters where consistency across WP is desirable.--Kotniski (talk) 07:05, 17 August 2008 (UTC)

I also support the change. Opposers say that we could improve auto-formatting to a system acceptable by all; what we should be asking, however, is "Do we need it?" If you approach it as if auto-formatting were suggested now for the first time, you'd see that its benefits would not justify the trouble of applying it. As a proposal, it would sink. Features in Wikipedia should have a good reason for being there—let us not allow inertia to dictate our actions. Waltham, The Duke of 10:22, 17 August 2008 (UTC)

Moreover, those with the intrest to have it improved have not the power and those with the power have not the intrest. JIMp _talk·cont 11:39, 17 August 2008 (UTC)

Support It's time to rid WP of these links to nothing of worth. JIMp _talk·cont 11:39, 17 August 2008 (UTC)

Support Lightmouse (talk) 11:50, 17 August 2008 (UTC)

• Harmful, unless... As Tony and others well know, there are many ways to use markup other than wikilinking that allow dates to be auto-formatted. Confusing the two is a convenient simplification that is simply unwarranted. MOS should indicate that if an editor wishes to unwikilink dates, that editor should use appropriate markup, such as <span class="wpAutoDate">17 August 2008</span> to allow the date to be appropriately formatted for presentation to readers. This kind of markup is automatically generated by e.g. {{date}}, isn't it? How can there be consensus when this hasn't been discussed? (sdsds - talk) 19:41, 17 August 2008 (UTC)

The Date template documentation says "dates between 1901 and 1969 will be displayed incorrectly (see bugzilla:11686). " Therefore this template is unfit for use. --Gerry Ashton (talk) 20:02, 17 August 2008 (UTC)

Support, if editors haven't read the comprehensive discourse already taken place, maybe they should familiarize themselves with the discussions on this page as well as the archived material. This discussion has now moved along, autodate formatting offers few advantages to the vast majority of readers/users who use wikipedia. The proposal, immediately above, of using markup such as <span class="wpAutoDate">17 August 2008</span> would put the onus on editors removing wikilinks although the case was not made for why we are keeping them in the first place. No one is advocating removal, merely deprecating the use of a format that is not that useful. FWiW, I would still favour the wikilinking of dates to lists such as the "year in film or "year in aviation." Bzuk (talk) 19:52, 17 August 2008 (UTC).

They are unlikely to familiarise themselves with any discussions unless there is a link to them. Would it be sufficient to take His Grace's list, make it into a subpage, and link to it, with a statement like Many editors find autoformatted date-links a net harm to Wikipedia. ? That's a deprecation. Septentrionalis PMAnderson 20:38, 17 August 2008 (UTC)

Bzuk: Piped links such as "year in film" and "year in baseball" are quite a separate issue from DA; this proposal will not impinge on them. Tony (talk) 00:10, 18 August 2008 (UTC)

On the contrary, it will affect them positively; not only will the removal of linked dates make the important links more visible, but the general de-linking will also expedite the removal of the remaining single-year links, and with it the confusion between them and the piped links you describe. Waltham, The Duke of 14:28, 18 August 2008 (UTC)

• Strongly Oppose. Most of the arguments *against* auto-formatting are actually arguments against the blue-links, not the auto-formatting. As Sdsds says above, we're nowhere near consensus on changing the guideline. -- SatyrTN (talk / contribs) 22:02, 17 August 2008 (UTC)

Well, trivial blue links are bad, but there are at least six disadvantages, some of them highly significant. See His Grace's list. Tony (talk) 00:10, 18 August 2008 (UTC)

See also the talk page thereof for a few arguments against auto-formatting in general. Waltham, The Duke of 14:28, 18 August 2008 (UTC)

• Oppose. The removal of auto-formatting is just inviting edit-wars. Bazj (talk) 23:29, 17 August 2008 (UTC)

Bazj: Can you point to a single instance of edit warring over date formats, even though autoformatting has been removed from many articles? Please provide a link to these instances here. Just as for the two basic varieties of spelling, WP has come to manage the selection and maintenance of one type very well indeed. What evidence do you have suggesting that the binary system for dates would spark wars whereas the binary system for spelling doesn't?Tony (talk) 00:02, 18 August 2008 (UTC)

Hasn't got there yet but the edits on the 14th & 16th Aug on Pope Leo XIII were heading that way. Autoformat at least lets User:Thomaq and User:Skyring see what they want to see. Bazj (talk) 11:10, 18 August 2008 (UTC)

Right, thank you for that, Bazj. I see that this article was a bit of a mess long before this debate ensued, with multiple MOSNUM breaches (mixture of US/international, autoformatted / not autoformatted, unspaced em dash in the dob/dod range, if you please, and wrong nav-box ranges at the bottom, still). And I'm afraid that in your attempts to reinstate autoformatting, you've got the syntax all wrong (e.g., 31 December, 1903). We go back to the first 12 months of the article history and see that there was a sole US-formatted date; I think that wins out, since any preference in Italy at that time seems irrelevant now. Either way, it should have been worked out at talk according to our long-established guideline "In the early stages of writing an article, the date format chosen by the first major contributor to the article should be used." I'll deal with all of these issues now, retaining DA for the moment, in US format. However, your comment "Autoformat at least lets [the two editors involved to] see what they want to see" I believe misses the entire point: our readers see the raw format, and it is they who matter, not what these two editors see! Tony (talk) 11:54, 18 August 2008 (UTC)

For the last month I have created a number of new articles within two of the most "strictly-adhere-by-the-letter-of-the-law" project groups, WP:FILM and WP:AVIATION. First of all, in the nature of full disclosure,I am not a neophyte editor. However, the articles have been the subject of interest to other editors, both new and from the ranks of veterans. I did not detect any sign of edit warring over the format of "no autodate links", much the opposite, the two project groups accepted the format without any visible reaction. I posted extensively to both project forums, to find mild interest in the direction I was taking but no real opposition, either. FWiW, I believe it is time for a change, and the majority of editors will accept this format adjustment without any problems. Bzuk (talk) 00:45, 18 August 2008 (UTC).

Support. Autoformatting doesn’t benefit the vast majority of readers. There was simply never any proper justification for creating a tool where only registered editors can see the editorial effect so 99.9% of Wikipedia’s readership see only a default style that could have been set in fixed text. Further, using the tool introduces the disadvantage of generating links to random irrelevant trivia. Even though the tool is widely used, that’s no reason whatsoever MOSNUM can’t call for editors no longer *adding* them to new articles and for explicitly allowing editors to change from autoformat to fixed text. Further, bots can later be made to do that drudgery for us. Greg L (talk) 03:56, 18 August 2008 (UTC)

Support. I have yet to see any compelling reasons for auto-formatting. So a few users can avoid seeing "16 November 2003" in some articles and "November 16, 2003" in others? They (or rather, we) still are the exact same users who think there are no problems with seeing "colour" in some articles and "color" in others, so what harm could date format differences do? Among the drawbacks of auto-formatting, the most serious is the hiding of inconsistencies from precisely those who would be most likely to fix them. My suggestion is to apply WP:ENGVAR to date formats while wikilinking only extremely relevant dates, and then the whole date (November 16, 2003), which would make for meaningful backlinks. -- Jao (talk) 13:04, 18 August 2008 (UTC)

Comment As I pointed out before without autoformatting there is no longer a reason, other than personal preference, why other alternative date styles can not be used within an article, providing of course they are used consistently within the article. So we can look forward to articles with "16th November 2003" in them. --Philip Baird Shearer (talk) 15:57, 18 August 2008 (UTC)

I don't see any reason why deprecating linking/autoformatting need affect what date formats are approved in the MoS.--Kotniski (talk) 16:49, 18 August 2008 (UTC)

Support - Agree completely with the rationale of Jao above, and others. There are no compelling reasons to continue autoformatting, to my knowledge. —Mattisse (Talk) 15:27, 18 August 2008 (UTC)

Support - Tony's proposal (with the "should" modification) states it clearly and benefits the greatest number of readers. If and when the developers decide to introduce coding to allow everyone to select a preference and to display so automatically, we can get by without. Askari Mark (Talk) 18:52, 18 August 2008 (UTC)

Support - many good reasons already given above. I also support Jao's idea: "My suggestion is to apply WP:ENGVAR to date formats while wikilinking only extremely relevant dates, and then the whole date (November 16, 2003), which would make for meaningful backlinks." Teemu Leisti (talk) 13:20, 20 August 2008 (UTC)

Support Tony's proposal above. Autoformatting as it is currently implemented does not have the kind of benefit that might justify its use, given that the blue links in articles should be of high value to the reader. EdJohnston (talk) 13:44, 20 August 2008 (UTC)

Support - plenty of good reasons to get rid of auto-formatting and none to preserve it. The auto-formatting hackery (and hacks to work around hacks for those hacks) is superfluous. Editors ought to write dates without adornment and without ambiguity; 'August 22, 2008' or '22 August 2008', just like they might write any other three words. Its not rocket science. -- Fullstop (talk) 17:21, 22 August 2008 (UTC)

Change

I believe that a reasonable person would determine that there is overwhelming consensus to add the following sentence to the section “Date autoformatting”:

This should not generally be used unless there is a particular reason to do so.

The movement from mandatory to optional has occurred, in various forms, over more than two years. This has come to a head in the discussion here and elsewhere over the past six weeks, culminating in calls from numerous experienced editors for further movement. Their arguments are plainly set out here; in addition, I submit that the views expressed in favour of removing DA by some 50 other editors, most of whom are not regular participants at the MoS talk pages, be taken into account as secondary evidence of wider consensus in the project. Their comments have been cut and pasted into a central page here. In addition, the current seeking of consensus was flagged at the Village Pump.

Arguments have been put against the move, apparently based on the following good-faith motivations.

• A frustration that WikiMedia could be lobbied again to improve the DA mechanism rather than dispensing with it ("throwing the baby out with the bathwater"). This appears to be particularly prevalent in users with IT/programming skills who value templates in general on WP.
• A desire that WPians not have to read the "other" format.
• A fear that edit-wars will ensue.

I put it to you that there is ample evidence on this page and elsewhere of compelling rebuttals of the opposing views. A number of people began the debate on the negative side and switched to strong support. Users Ckatz, EncMstr, BillCJ and Francis Shonken have expressed their opposition to the move, but have not contributed to the debate on this page; except for EncMstr, whose views are of the first type above, it has not been easy to locate and encapsulate substantive arguments by these users (F Schonken would only refer, repeatedly, to this debate from two years ago).

I will insert the sentence into MOSNUM forthwith. Thank you all for your contributions. Tony (talk) 01:38, 23 August 2008 (UTC)

There is a good aspect and a bad aspect to the interaction between autoformatting of dates and citation templates. The bad aspect is that the people who code the templates will have to be pursuaded to stop autolinking dates. The good aspect is that if they are pursuaded, many of the dates in citation templates will change at once. --Gerry Ashton (talk) 02:45, 23 August 2008 (UTC)

Proposal by Seddon

Having read through the various discussions I would like to offer a middle ground outcome when it comes to auto formatting. There is still the option of improving the auto formatting methods but this is a an option until that is worked upon. This may suffice until that time.

• When using May 15 or 15 May no wiki linking is to be used as common sense prevails here that no matter where in the world you are you should be able to understand the date.
• When using May 15, 2008 or 15 May, 2008 and other variations based on the same style for the same reasons as above, no wiki linking should be used.
• When using ISO or US date standards in the format 2008-02-04 and 2008-04-02 and other integers between 1-12 used in this format, wiki linking should be used due to the easy confusion. For those who are not logged in, the link to the date will remain allowing this to be clarified.

This removes redundent wikilinking where it is not necessary but allows clarification where it is obviously needed. Seddσn ^{talk Editor Review} 01:51, 18 August 2008 (UTC)

I don't understand the third bullet. Are you suggest there is any standard anywhere in which the first group is the year and the last group is the month? I've never heard of such a standard. Are you suggesting there is a difference between the US and ISO interpretation of "2008-02-04"? I live in the US, and don't know of any difference in interpretation (except that some Americans might not know what to make of it).

Also, in the second bullet, I would not use a comma in "15 May, 2008" --Gerry Ashton (talk) 03:06, 18 August 2008 (UTC)

I don't think i quite understand your question in regards to the 3rd bullet. Could you clarify for me? Its a little late so my brain isn't working. Seddσn ^{talk Editor Review} 04:14, 18 August 2008 (UTC)

I understand your question now :) I also believe a mistake was made on my part as you have pointed out. There is no standard but i feel misinterpretation is possible. Seddσn ^{talk Editor Review} 05:03, 18 August 2008 (UTC)

The third bullet makes no sense to me at all. Also, please explain it from scratch. Please say what you think each example means, giving a spelled-out month. --Gerry Ashton (talk) 04:33, 18 August 2008 (UTC)

ISO dating should simply not even be in the question. Your first statements concur with others on the unnecessary need for autodate linking, leave it at that. FWiW, there is enough evidence that ISO dates cause too much confusion. Bzuk (talk) 05:07, 18 August 2008 (UTC).

So prehaps converting all ISO dates to some other format by the whatever it gets changed to first method, then leaving all other dates without formatting? Seddσn ^{talk Editor Review} 05:49, 18 August 2008 (UTC)

• Support - This proposal as written is clear and understandable, and should be implemented, especially as regards ISO dates. That is, MoS should discourage editors from removing wikilinking around ISO dates. Moreover, per enwiki best practices, if the dates in an article (or table, or whatever) are already in ISO form, other editors should make their edits conform. (sdsds - talk) 03:46, 19 August 2008 (UTC)

• Oppose - The date articles (for example, 18 August) are often very badly referenced; we should try to pretend they don't exist and avoid linking to them as much as possible. --Gerry Ashton (talk) 03:54, 19 August 2008 (UTC)

• Oppose , it again brings in the conundrum of ISO dating. FWiW Bzuk (talk) 12:18, 19 August 2008 (UTC).

Basically, little or no support for this proposal, but moving it here to keep up the continuity of the discussion. FWiW [[[User:Bzuk|Bzuk]] (talk) 12:18, 19 August 2008 (UTC)).

• Comment. For those who oppose autoformatting of ISO dates (or ISO dates altogether): how would you execute this edit? I believe it is warranted; it replaces a non-standard date format with a standard one. Automatically, with a (reasonably simple) regex. (Done through wikEd, incidentally.) I could have omitted the wiki links, but what purpose would that serve? In this way, editors can at least see their preferred date format, while the non-editors will see wikilinked ISO format, which is not a big problem. So, how would you do it? I'm interested. GregorB (talk) 12:48, 19 August 2008 (UTC)

Perhaps you have not seen the comments on this page about the confusion as to ISO dating. Those unfamiliar with the format will not be able to understand the sequence of mm/dd or dd/mm while the vast majority (one editor has estimated it at over 98% of Wikipedia users) do not have date preferences set on their browsers. What the autoformat date linking does, is accomodate the tiny proportion of experienced editors, but for the vast unwashed, it produces an undecipherable blob of numbers, or a blue link that gets them to date articles that often have no relation to the article context. FWiW, the entry that was provided as an example could have been written out in "clear" rather than ISO, sure it shows up properly on my browser but it won't on the typical user's browser who will shake their heads and wonder what is that? Bzuk (talk) 13:03, 19 August 2008 (UTC).

Once ISO dates are wikilinked, there is no confusion (i.e. the date is not ambiguous), which is true for both editors and non-editors. Other things being equal, marginally useful links are a small price to pay. You did not say what would you do instead. Leave it as it is? "01-02-2007" - that's a non-standard, ambiguous date. In this particular case it's February 1, 2007 - granted, this is the best solution. but I don't have a magic wand to change it, and I guess I'm not masochistic enough to do it by hand, introducing errors in the process. What to do with existing digits-only dates, ISO-style and non-standard ones? My position is that in these particular cases wikilinked ISO is the best temporary solution. GregorB (talk) 13:24, 19 August 2008 (UTC)

I did mention that I would have formatted the note in "clear" which is to say, written out as 1 February 2007. BTW, to someone unfamiliar with the format, reading "01-02-2007", is it "February 1, 2007" or "2 January, '07" and as a Polish editor pointed out in the WP"AVIATION PROJECT group as to the uncertain nature of determing the date, why have it all, and the entire project group now deprecates ISO dating. FWiW, digit only dates are very rare in English prose and look decidedly odd in Wiki text. At this point, wikilinking all dates into ISO would be as much a problem as simply writing out the dates in the first place, as to make the wikilinking work, all users would have to be reading the format accurately by having their preferences set. I'm not even going to bring up the differences in date format between regions. Bzuk (talk) 13:44, 19 August 2008 (UTC).

Yes, I'd write it in clear too, but as I said, I don't have a magic wand... I did not say that digit-only dates are not rare in English prose (they are) - they are not rare in Wikipedia articles, unfortunately. To summarize my point: 1) "clear" dates are better than 2) wikilinked ISO dates, which are in turn better then 3) non-wikilinked ISO dates, which are in turn better than 4) non-standard gunk (e.g. Kazushi_Sakuraba#Mixed_martial_arts_record). It's difficult to convert #4 to #1, but it should be reasonably easy to convert it to #2, with at least some benefit. GregorB (talk) 13:59, 19 August 2008 (UTC)

I would write your analysis as: 1) "clear" dates are better, use them. 2) See #1. There is no large-scale advantage to using a date system that the majority of users will not find as a benefit. Remember we are not writing for ourselves, and if readers do not use or need autoformat date linking, then we should not be using it. Bzuk (talk) 14:15, 19 August 2008 (UTC).

I have a feeling that we didn't understand each other. I got it, "clear" dates are the best, but are you actually volunteering to fix Kazushi_Sakuraba#Mixed_martial_arts_record with "clear" dates? I guess you're not, and I'm not volunteering either. Saying "don't use ISO for new stuff" is fine, but what about the old non-standard stuff? If the choice is leaving it as it is or converting it to wikilinked ISO - and let's face it, it is currently the only reasonable choice for countless existing tables in Wikipedia articles - then I'd opt for the latter. GregorB (talk) 14:48, 19 August 2008 (UTC)

Actually, I am at work right now in rewriting all dates in Wikipedia to a consistent style. I will get back to you in 20 years or so. No, of course, it's illogical to try to change everything that is in place but an effort to use common sense protocols will make a difference for the future. What is being advocated here is the deprecation of autodate linking as it is of limited (or dubious) benefit and a better system already exists. FWiW Bzuk (talk) 14:54, 19 August 2008 (UTC).

Oppose: If ISO dates are ambiguous, there's a better solution than linking: don't use them. There are only two advantages of ISO dates I can think of.

1. They are short and of a standard length. With ordinary dates, however, if instead we abbreviate the month name to three letters, it's not much longer and a leading zero can be added where necessary to standardise the length.
2. They sort chronologically. The only way to make use of this on WP is in a sort table. There are ways to get ordinary dates sorted correctly.

So don't link ISO dates, better just avoiding them. JIMp _talk·cont 15:43, 19 August 2008 (UTC)

Oppose. Instead, I support Tony's proposal, for reasons already enumerated. Teemu Leisti (talk) 00:53, 21 August 2008 (UTC)

Oppose. To the extent that DA is to be deprecated, Seddon's middle-ground proposal is OBE; as for ISO dates, the rarer they are, the better. Askari Mark (Talk) 02:42, 23 August 2008 (UTC)

Datestyle citation template proposal

Following MOS decission that linking dates not automatically the prefered approach, discussion had at the Cite XXX family of templates about removing autowikilinking, but allowing the option of a per-article only-if-a-consensus basis of selecting US or International style of dates approapriate for article topic and so make references use a consistant style of dates as appears in the actual text of an article. Discussion centred on Template talk:Cite web to add a datestyle parameter, but with plan to then role-out across the citation templates.

First attempt had to be nearly instantly reverted, as others not using the templates as we (i.e. largely myself) had appreciated and actual implementation proved not quite what others had expected. So after some very helpful discussions and a number of alternative suggestions (and large number of examples set out), I think the proposal is ready for re-implementation. However we lack a breadth of other editors' input, so seems sensible for me to post a heads-up here, as the cite templates clearly must be subserviant to MOSDATE :-)

As a few quick sandbox examples for an updated cite web:

With current default with no datestyle set

{{User:davidruben/sandbox4|author=Author |title=Title |url=http://example.org |date=August 24, 2007 |publication=Pub |accessdate=2008-08-18 |datestyle=}}

Author (August 24, 2007). "Title". Retrieved on 2008-08-18.

Where datestyle=mdy

{{User:davidruben/sandbox4|author=Author |title=Title |url=http://example.org |date=August 24, 2007 |publication=Pub |accessdate=2008-08-18 |datestyle=mdy}}

Author (August 24, 2007). "Title". Retrieved on August 18, 2008.

Where datestyle=dmy

{{User:davidruben/sandbox4|author=Author |title=Title |url=http://example.org |date=August 24, 2007 |publication=Pub |accessdate=2008-08-18 |datestyle=dmy}}

Author (24 August 2007). "Title". Retrieved on 18 August 2008.

see Template talk:Cite web#Review of current and sandbox coding - examples and now Template talk:Cite web#Proposal to go #2 - thank you David Ruben ^Talk 23:14, 18 August 2008 (UTC)

David, thanks heaps for your work on this. It looks good, but only as far as my untrained eyes can see. I do wish we could drop the "on", which in a list of many is quite unnecessary. I see a lot of non-citation refs that have no "on". Tony (talk) 00:01, 19 August 2008 (UTC)

• Thats all well and good for Cite web, since the publication date and especially the access date are very likely to be in the Gregorian calendar. It is not acceptable for Cite book, since a fair number of books will have publication dates before the Gregorian calendar was adopted. I believe your code should reject dates in the YYYY-MM-DD format prior to 14 September 1752, and require that any such date be in a format in which the month is written out in letters. --Gerry Ashton (talk) 00:14, 19 August 2008 (UTC)

  Sorry not an issue can address for time being as mediawiki can not format (using #time) dates outside of range 1970-2038. So for all dates outside of this range, all that any template can currently do is to leave the input unaltered. Hence we can process "|date=1984-10-24 |datestyle=dmy" as 24 October 1984 but for "|date=1884-10-24 |datestyle=dmy" just as 24 October 1884 and likewise if "|date=24 August 1884 |datestyle=dmy" this must be just shown "as is" i.e. as 24 August 1884, or even if an editor enters "|date=uncertain but after 1792 |datestyle=dmy" as uncertain but after 1792. David Ruben ^Talk 03:42, 19 August 2008 (UTC)

  So if 'datestyle' parameter was introduced at {{cite book}}, then for now the 'date' parameter will still be entered by editors in the suitable format for the article (US or International) as they currently do, and whether or not 'datestyle' parameter is set will make no difference if the date is before 1970 (hence "|date=12 January 1624 |datestyle=dmy" or "|date=January 12, 1624 |datestyle=dmy"). But {{cite book}}'s instructions do allow for 'date' to be entered in ISO format and so where this might be after 1970, then at least 'datestyle' might show in a style consistant to dates within the article text and with other references that might use {{cite web}} or perhaps {{cite journal}} etc. (hence "...<ref>{{cite book....|date=1981-12-20|datestyle=dmy}}</ref>....<ref>{{cite web ...|date=2008-07-26 |accessdate=2008-08-18 |datestyle=dmy}}</ref>" migh appear in a more uniform manner). If at some future time mediawiki gets improved, then dates within cite book that predate 1970 might be also be displayed in a fixed consistant style as set primarily by user preferences (if set) or else as a default style set by the article's editors. For now I agree cite book for older publications wont appear any different with 'datestyle' enabled, but it may at least nudge towards greater consistancy throughout an article and across different cite XXX templates appearing within an article. David Ruben ^Talk 04:02, 19 August 2008 (UTC)

• David, consider also a template that is not in the American Psychological Association (APA) style, but uses the Modern Language Association (MLA) style widely used in referencing works in the social sciences. The style consists of: Author (last name, first name). Title. Place of publishing: Publisher, date. Or failing that, at least a Harvard style template of Author (last name only) date, page. With these templates we are perpetuating the American Psychological Association myth of most-in-use. As for the need for citation templates, a great number of editors do not favour their use, and if anything, an effort to completely discard them would not be out of place. FWiW, I would support a outright deprecating of citation templates as they represent the worst of the "garbage in, garbage out" axiom. Bzuk (talk) 00:19, 19 August 2008 (UTC).

  Citation templates are of course optional, users are equally justified in choosing to manually mark-up references. The proposed 'datestyle' parameter discussion therefore only applies where an editor is choosing to use a citation template and aims to offer greater flexibility for them. That said, in biomedical articles where I mostly contribute, the use of PubMed and User:Diberri's markup generator tool are extremely useful. Citation templates have, IMHO, two major advantages, firstly the ease of initial markup creation (using automated tools described) and secondly, as you point out, if we could all agree to change citation formats (dropping for example the "on" in "accessdate" parameters, or deciding that 'author' or perhaps the 'title' should be in italics) then a simple template tweak takes care of this for us. However, as I say, using a citation tenplate is optional, and the 'datestyle' is just an attempt to help give greater flexibility for editors who elect to use the templates (a mediawiki feature to set a per-article variable that might be read and used by all templates appearing in the article would be a simpler option, but that is just mediawiki wishful thinking for now).David Ruben ^Talk 03:42, 19 August 2008 (UTC)

  Bzuk, it's the American Psychological Association, not Psychiatric; I meant to point this out before. I HATE their reference formatting with a vengeance. It hasn't changed in a long time, and contains illogical quirks and lots of redundant punctuation. Moreover, they make packets of money by changing a nut and a bolt in their manual and forcing all of the institutions to buy their new edition every few years. Not happy.

  More generally, I remain to be convinced that any citation template is superior to plain old manual WYKIWYG: What You Key in Is What You Get. The great advantage is local editor control and stability (whereas templates can suddenly change remotely, in the shadows. Tony (talk) 07:55, 19 August 2008 (UTC)

Tony1, thanks for the correction, the Psychological typo may have been inadvertent or as a subliminal aspect of my bias. I know that the dominance of the APA style dates back from my University days when professors literally enforced the style on legions of unsuspecting students. At the time, the MLA guide was prevalent but as a number of profs explained it, they were tired of having to explain the format and APA was so much easier. BUT, the overall question of citation templates remains, what they were there for in Wickywacky land was as an aid, not as the "be all and end all" that some editors have claimed. FWiW Bzuk (talk) 12:14, 19 August 2008 (UTC).

Well, I flatly declined my psychology supervisor's advice to toe the line with APA; my primary supervisor was in music, so that gave me a freer reign. I have no reason to disbelieve the explanation I've heard somewhere recently that citation templates were originally devised for newbies, and for WPians who weren't used to the demans of research references. My line is that compared with learning how to write prose, choosing a (manual) format for references and applying it consistently is a doddle; in fact, the relatively clerical task of checking through a manually constructed reference list is a good sop to the intensity of writing article prose. Tony (talk) 13:07, 19 August 2008 (UTC)

And as SandyGeorgia offered in a previous statement, one other consideration for "trumpeting" citation templates was that it would be a time-saver, which she concluded was not the case. In trying out the citation templates when I first encountered the Wiki world, I found them not only cumbersome, but rife with "bugs" that made them less than useful. I agree with you, the cite templates are not "cost effective" and their use should be more than entirely optional, but that's another story... FWiW Bzuk (talk) 13:16, 19 August 2008 (UTC).

A concern for the future

First off, I completely understand most of the reasons to remove date linking - the wikilink function is being overloaded with date autoformatting to distasterous results, as normally happens with badly thought out overloaded operators. Approaching the mediwiki devs resulted in no immediately plans to correct this. As a result, it does make sense to remove this function.

However, are we 100% that there will never be some extension or addition to Mediawiki that will give us the ability to autoformat dates via some other means (template/magic word, whatever)? My understanding is that if there was a magic template that took in any recognizable date format and spat out the date formatted to the user prefs (with some mechanism for non-logged in users and per page to see a consistent date format), we'd be all over that in a heartbeat. Let's assume the template also dealt with date ranges, the Gregorian calender per-1600 issue, IP geo-tracking, etc, in that it is a magic bullet for date autoformatting. (I don't know the programming for this extension, but I can see the template creation being rather easy as long as the extension is flexible as such). My concern thus is if there is going to be a possibility of date autoforatting via a sensible means, we are going to want to go back to that, and thus to do that we will need to re-iterate through all articles to find such dates. The mechanism being done now is stripping an easy (despite the illogical nature) method of computer-assisted detection of dates such that if we convert to this hypothetical scheme, it's going to take a lot of extra work.

Presuming that we haven't ruled out a future mediawiki change, my suggestion is that any dates stripped of wikilink presently need to be wrapped instead with a dummy template, "date" or "d" or something short. Presently, that template would simply regurgitate the date (no formatting at all), but the key is that now a script/bot can identify dates. This gives us a lightweight placeholder for all dates such that we can readdress the possibility of autoformatting in the future, with the turnaround to reconvert dates to autoformatted ones rather trivial - it would just be a matter of modifying this "date" template. Only if we are 100% sure that mediawiki will never have this ability can this aspect be ignored. However, my impression is that there's still room for such an addition. --MASEM 13:45, 20 August 2008 (UTC)

I don't see any harm in doing this. If the template is quick to type, it might encourage people to use standard date format too. But personally I see no value in further development work towards date autoformatting; the benefits are too minuscule to justify the effort and the additional complexity.--Kotniski (talk) 14:28, 20 August 2008 (UTC)

Bearing in mind, the "don't flog the dead horse" points brought up throughout this set of lengthy "strings", at this juncture, I would approach autodate formatting with some apprehension. FWiW Bzuk (talk) 15:23, 20 August 2008 (UTC).

What makes this really galling is that someone actually did the development work to auto-format the dates without link-syntax. It just needs testing and implementing Rich Farmbrough, 01:37 21 August 2008 (GMT).

I'm highly suspicious that that someone did the work for such a development and didn't bother testing it. My guess is that s/he ran into problems at an early stage. But the community has moved on from that issue: now, it's a matter of who needs to autoformat dates (no one, IMO), and why we should persist with a technical feature that delivers a nebulous advantage. I want to deliver control to editors, not developers, and WYKIWYG is a quick, simple way of achieving this. Tony (talk) 03:02, 21 August 2008 (UTC)

Date linking: User:Tony1

[[::User:Tony1|Tony1]] ([[::User talk:Tony1|talk]] · [[::Special:Contributions/Tony1|contribs]]) is going through articles and using some sort of script to unlink all full dates in the article. He includes WP:MOSNUM in his edit summary, yet as far as I can see, there is nothing to say dates should no longer be linked. In addition, the usual guideline is not to change the format of dates except to make an aticle consistent, and I would have thought linking is covered by this policy too. Any comments or suggestions of how to proceed would be welcome. JRawle (Talk) 09:29, 19 August 2008 (UTC)

The user has replied to me and explained why he feels dates should no longer be linked. It seems reasonable to me. JRawle (Talk) 11:18, 19 August 2008 (UTC)

I take issue with it too. Tony's scripted changes today to Battle of Arras (1917) (a WP:FA) were rather abrupt to say the least. At a minimum, if he's going to do this, I'd expect edit summaries that link directly to an explanation without requiring a degree in forensics. LeadSongDog (talk) 16:16, 19 August 2008 (UTC)

I wasn't aware that an edit summary could link. Are you sure? If so, it's quite possible. Tony (talk) 00:15, 20 August 2008 (UTC)

• “Change” on Wikipedia, is never easy. Thank God for Tony1’s efforts in seeing this issue through to its natural conclusion. Auto-formatted dates should not be linked. And in my opinion, everyone should just forget worrying about what registered editors see and simply hard-code these *formats* straight to the default view that 99.9% of Wikipedia’s readers (non-registered readers) see: the last column in this table, which backhands the vast majority of our readership with the label “What *others* will see”. And as for that bottom option (coded [[2005-05-15]]), we might as well loose that one since 99.9% of Wikipedia’s readership sees only the ugly all-numeric, hand-coded input format. Greg L (talk) 03:18, 20 August 2008 (UTC)

• GregL, we should change what 99.9% of Wikipedia’s readership sees; we should not change the underlying coding of the dates. (sdsds - talk) 03:29, 20 August 2008 (UTC)

Sd, I appreciate your desire to retain a programming functionality, but I think you're way too optimistic in your expectations of technical improvements over at WikiMedia, especially something as major as the introduction of some IP-based system; in any case, geographical IP regions don't precisely map onto date-formatting usage.

Also, I don't think you've addressed the most basic issue: is the month–day/day–month order worth giving more than five seconds of your talent and energy to? Was it ever a problem in the first place? Let's rejoice in the amazing degree of homogeneity in our language, despite its geographical dispersion and the lack of a centralised authority such as the French language endures.

WYKIWYG (What You Key in Is What You Get) is the safest, simplest solution, allowing editors to retain local control and most easily check for inconsistencies and inappropriate choices (surprisingly common); on its side is the convenient fact that it takes less work, scrutiny and induction of newbies, not more. I'm all for templates and toys where their advantages clearly outweigh their disadvantages, but DA is not one of them. Let's allow (high-value) wikilinking to breath a little easier in our text displays. Tony (talk) 04:25, 20 August 2008 (UTC)

Tony, we both know we disagree on this. There is only a small change required to show arbitrarily formated dates to the "99.9%" who don't see them now. (We don't need to attempt to guess at an anon reader's preference. Any consistent format would be fine.)

You think WYKIWYG is "safest", and I'm willing to let you key in (new) dates without the wikilinking you dislike. I am among the many, however, who object to your script-based wholesale rewriting of dates that strips them of the markup that allows them to be auto-formatted. If you were to do this to articles where I had used wikilinked ISO dates, I would feel justified reverting your changes, because there is no consensus about this.

To address your "most basic issue": yes, enabling wikipedia content to be viewed in ways that match per-user or per-article style preferences is worth much more than five seconds of my effort. For example, I would be overjoyed to extend this mechanism to include pounds versus kilograms, as I grow weary of seeing {{convert}} everywhere. Heck, I would like to extend it to allow kilometer or kilometre, at the user's preference! Enwiki should move forward with this kind of functionality, not backwards! (sdsds - talk) 04:45, 20 August 2008 (UTC)

Question for sdsds: How would you prefer to write, in the wiki edit box, the following dates from Gregorian Reform of the Calendar published by the Vatican Observatory in 1983 and edited by Coyne, Hoskin, and Pedersen (p. 219):

On 11 February, 1582 Sirleto sent Antonio Giglio to Mondragone, the villa outside Rome where the Rome preferred to stay whenever he could, and there the Pope signed the bull on 24 February, 1582.

Please assume you are paraphrasing the sentence and so are not obliged to keep the same format as the original. What, if any, remarks would you include about the dates? --Gerry Ashton (talk) 05:39, 20 August 2008 (UTC)

Another question for sdsds: how is it possible to deduce from a reader's IP what preference he has for pounds over kilograms? Kilometre vs. kilometer possibly, to some extent, but is there really any benefit compared with the complications involved? And in any case, the main issue is the mass linking of dates - there seems to be clear consensus against that. If the developers want to spend time working on autoformatting for IP users, then at the same time they can make it independent of the linking. As it is, removing the linking affects autoformatting only for an infinitesimal percentage of readers, while improving the presentation for everyone, so I still see no valid argument against doing it.--Kotniski (talk) 06:36, 20 August 2008 (UTC)

Tony, going back to edit summaries, I agree that it should link to an explanation of why you are unlinking dates, rather than to WP:MOSNUM. The latter doesn't have anything about proposals to unlink dates, which was that threw me in the first place yesterday. If you link to one of the excellent, detailed explanations you have written about the issue, that woud be much more helpful. Some people will still disagree, but others may well be persuaded by your arguments. It would do more to promote what you are doing. At present, editors who have never seen the debates on this talk page will see you've stripped all the links from an article they contribute to, and to them it may seem to be some sort of unilateral action without any discussion.

Edit summaries can certainly be linked. I use linked summaries quite often. See this one from today: [4]. Also, tools such as navigation popups often include a link to themselves in the edit summary. The (fairly recently added) edit summary preview also shows whether the link is going to work properly. JRawle (Talk) 11:29, 20 August 2008 (UTC)

JRawle—very good advice, which I shall heed. Tony (talk) 11:55, 20 August 2008 (UTC)

Responding to User:Gerry Ashton: MOSNUM is a guideline, so the advice that, "Editors should follow it, except where common sense and the occasional exception will improve an article" is likely applicable in the test case you mention. In this case, or in any case where dates might be "special", the specific advice in MOSNUM simply does not apply. MOSNUM should give advice for the "standard" case of a date that involves no particular trickery. (Had the dates not related to this special case of the Gregorian calendar, MOSNUM should advise you to code the sentence as, "Sirleto sent Giglio to Mondragone [[1582-02-11]], and the Pope signed the bull [[1582-02-24]]." This would be rendered: "Sirleto sent Giglio to Mondragone 1582-02-11, and the Pope signed the bull 1582-02-24." Even in this tricky test case any user clicking on a linked 1582 learns about the, "Gregorian Calendar switch" which is mentioned at the top of the page for that year.)

Responding to User:Kotniski: You perhaps have me confused with some other contributor to this discussion. I do not advocate trying to guess at what format might be preferable to a non-logged-in user. Except for logged-in users who have specified a date format preference, I advocate choosing an arbitrary format, and consistently presenting all dates marked for autoformatting to that chosen format. Same for spelling, or any other variation based on user preference: if the user hasn't specified one, enwiki should arbitrarily choose a varition so as to present a consistent article. (sdsds - talk) 21:37, 21 August 2008 (UTC)

date-mess uncovered

The removal of autoformatting in Plug-in hybrid, which is featured article and a good read, uncovered an unholy mess of international and US raw formats that our readers have been viewing for who knows how long. The diff also shows a mistake in the syntax of the autoformating of an ISO date in the refs (2008-6-16).

I've left a note on the article talk page asking editors to decide which format is most appropriate using our guidelines at MOSNUM as a basis, and offering to assist if required. Tony (talk) 11:50, 20 August 2008 (UTC)

Would it not make sense before undertaking a larger effort to strip date linking as to have a bot/script attempt to convert dates to a specific format (would need a manual input to select between "Month Day, Year", and "Day Month Year") such that what are inconsistencies from older FAC articles would be convert automagically, but dropping warnings on the talk page when a date is in a format it cannot parse? I understand requesting manual help for these testing-the-waters cases, but if taken to a mass level, leaving a large number of articles with inconsistent dates that need to be fixed is going to take a lot more effort. --MASEM 13:33, 20 August 2008 (UTC)

Your question has two parts:

• Part 1: Is it possible for automated support to assist with the conversion of all dates in an article to one format?
  • Answer = Yes. The code is fairly simple. False positives and misses are foreseeable so it would be best with human oversight.
• Part 2: Would such automated support be needed before date links are stripped?
  • Answer = No. It is an entirely independent issue.

Regards Lightmouse (talk) 15:34, 20 August 2008 (UTC)

Date formatting: why isn't this done with a template?

I've wondered for years: why is date formatting accomplished entirely by "smart" code rather than by a template? For example, if we had something like {{format date|year=1776|month=July|day=4}} which could also recognize forms like {{format date|ISO=1776-07-04}} and {{format date|year=1776|month=07|day=4}}, then format (as now) based on user preference, but which would provide one single format as the default for everybody not logged in (e.g. "4 July 1776") it would seem infinitely more straightforward than the present magic. It could also deal well with year being optional, and could even be made smart enough to deal with date ranges (e.g. {{format date|year=1776|month=July|start-day=4|end-day=7}} could be shown, depending on preferences, as "4-7 July 1776" or "July 4-7, 1776"). - Jmabel | Talk 17:10, 20 August 2008 (UTC)

• Jmabel, with regard to this: “[the template would] then format (as now) based on user preference, but which would provide one single format as the default for everybody not logged”, that’s the way it currently works—there is one single default for *common folk* (99.9% of Wikipedia’s readership). And one format in particular, [[2005-05-15]], defaults to 2005-05-15 for 99.9% of our readership instead of showing the pretty May 15, 2005 some of us privileged editors like to see.

  I think it’s best if all us editors get over this notion that we are doing any measure of good by creating and using tools that are intended to improve the user experience for us editors (0.1% of Wikipedia’s readership), and which just gives a default view for “everybody else”. We editors should always be looking at precisely what everyone else is seeing.

  Many editors have suggested we simply improve the templates and magic words so they can benefit I.P. users too. But this would require parser functions and wholesale changes to the way Wikipedia’s servers work in the background to gather the requesting readers’ I.P. address, then match those to a database to determine which country they reside in, and spoon-feed content based on their country’s customary practices. That might happen (or maybe not), but it would likely be a year or more before we saw such a radical change. I wouldn’t hold my breath.

  In the mean time, to paraphrase Donald Rumsfeld, we must make do with the editing tools currently available to us. The current tools A) produce Easter-egg links to mindless trivia that is almost always unrelated to the topic, and B) delude editors here into thinking pretty looking date formats are being produced for our readership. It is high time these tools be abandoned. If we’ve got a crappy-ass Mark 14 torpedo that circles around and blows up the person who shot it, you simply avoid using that tool until you are provided something better. Greg L (talk) 19:50, 20 August 2008 (UT

• There already is the template {{Date}} that can take any format and emit it as e.g. "20 August 2008". Should the automatic date formatting be fixed at some point so that IP users will get a consistant formatting (no matter wich), the template could be changed accordingly without much ado. If Tony1 keeps unlinking dates though, which I oppose, then a change like that will mean that all those dates will have to be manually tagged again (can't be an automatic process due to ambiguity). I could live with all dates being wrapped into the aforementioned template though. --Amalthea^Talk 20:31, 20 August 2008 (UTC)

• Amalthea, {{Date}} is still another tool (that thankfully doesn’t link to trivia), that seems intended only to stop editwarring between editors by pacifying them with a format they are pleased to look at. That strikes me as nothing more than an attitude of “I don’t give a dump what 99.9% of our readership sees just as long as I don’t have to look at some other asshole’s method of writing out dates.” I reject that attitude as childish. It also condemns Wikipedia’s current policies for determining when a consensus has been arrived at and how to resolve conflict. To placate editors in the face of these dispute-resolution shortcomings, silly tools for just us were developed.

  We editors should always be looking at precisely what unregistered readers see.

  Further, as far as I can see, hand-coding dates produces better results than the template.

  Why? Check this out: if I use the template to code {{date|Jan 15, 2001}} in order to generate 15 January 2001, the template-generated version currently word-wraps between the “Jan” and the “15” (something I assumed it wouldn’t). My hand-generated dates don’t do this because I use a non-breaking space. Yes, I know that detail could be fixed in the code for the template, but that’s a secondary issue. It is utterly beyond me how anyone around here got it into their heads that we are making Wikipedia a better place by making tools that produce what we registered editors enjoy seeing but just provide a standard default for *others* (read: pretty much everyone); editors should just write out dates. This {{Date}} tool reminds me of putting horse blinders on two fourth-grade boys who can’t get along when they are sitting next to each other in class.

  MOSNUM has plenty or rules for the date formats that would be most suitable for various articles; we editors don’t have to be so damned intolerant that we make tools just so we get to see precisely what we want to see. Greg L (talk) 21:18, 20 August 2008 (UTC)

• You are wrong.
  If you had taken a look at {{date}} you'd know that it *always* emits dates in the format "20 August 2008" at the moment, but is prepared to display it formatted by user preferences or "using some sane and human-friendly default when a user is not-logged-in or when no date-style preference has been set" once MediaWiki has the capabilities to do so.
  I repeat, users that are not logged in or don't have a preference set will at the moment see the same as any other logged in user, and once a better date autoformatting feature is implemented it can be changed so that those 99.9% you mention will still see consistant dates (of whichever format is decided upon, I doubt that there'll be IP based guesses soon).
  Concerning the non breaking space: I agree that this should be fixed. I am also sure that wrapping dates into some kind of automatic formatting (the date template, for lack of a better one) can make sure that dates will *consistently* get those nbsps.
  Again, I'm convinced that the script assisted edits Tony is making right now are among the worst solutions for this problem (and I have yet to find the consensus he mentions) since all he accomplishes is to loose the meta information that makes a fix for it easier in the future, and he leaves the articles in just the inconsistent state he encountered them. If it stays wrapped in a template, a bot can implement a better solution should there ever come one to pass.
  Cheers, Amalthea^Talk 21:41, 20 August 2008 (UTC)

• I did too take a look at {{date}} Amalthea. And I understand exactly what it does. You seem to be missing my point (or refuse to see it). I’ll repeat what I wrote above again: We editors should always be looking at precisely what unregistered readers see. Whatever the Date template shows for non-registered editors, that is all we need to type in hard-coded text. We don’t need tools that are simply designed to placate stubborn editors who insist on viewing dates in a *special* format that regular readers (99.9% of Wikipedia’s readership) can’t see. Where’d you get the idea that tools that provide a special benefit for only us registered editors was a wise thing to do? The very premiss behind making these tools was arrogant and patently absurd for an on-line encyclopedia. Greg L (talk) 04:39, 21 August 2008 (UTC)

• Sorry, I did misunderstand you then, I thought you were still arguing about the inconsistent dates on one page that users will see - I think almost anybody agrees that this inconsistency is very bad.
  I do not see however the harm in giving registered editors the choice to see the date in whichever format they want. There are only 4 different date formats - whatever the default format is for anons, I'm sure that a very high percentage of the registered editors will see the article in just the same way as anons do.
  Furthermore, as sdsds said, a user preference to show metric or imperial units would be even nicer. I agree with him that MediaWiki should go forward with this, not backward.
  Lastly, why don't you propose to get rid of the skinning system as well? All your arguments apply there, too. --Amalthea^Talk 13:00, 21 August 2008 (UTC)

• Good point. Skinning is the root of the entire problem here. The only skinning allowed should have been limited strictly to issues unique to the editing experience, such as looking at the UTC time offset of edits in the history view. The first skinning options and editing tools (templates) that allowed editors to see article content that was different from that which the vast majority of our readership saw was a bad idea.

  The problem with giving editors a “choice” is we are masking editorial problems from the very people who are creating content. If there is a problem with date formatting that is *shocking* to people in the UK, for instance, then we need to develop a consensus amongst editors that a new way for dates must be developed, or decide that the problem isn’t a problem, or decide on an easy-to-follow set of rules to use article-appropriate date formating (my favorite). Simply masking the problem from the people who write articles and participate in establishing MOSNUM guidelines simply perpetuates problems for the vast majority of our readership. And in the case of the [[2005-06-06]] date tool, the problems we’ve been masking are significant. Greg L (talk) 16:32, 21 August 2008 (UTC)

• The Date template only works for dates between 1970 and 2038, because it relies upon programming tools intended for internal use in the Unix operating system. As such, it is only fit for use to time events that occur internally within a computer that was placed in service after 1970 and will be junked before 2038. It is unfit for any other use. --Gerry Ashton (talk) 21:47, 20 August 2008 (UTC)

• That's not good. Obviously I haven't taken a very thorough look at it either. :| I'm not sure if there's a feasible parser-functions way to take in all those date formats, since I wouldn't want to make entering dates difficult by using several template parameters, or relying on editors not to use it for dates outside the supported range.
  Can those unsupported date ranges be somehow identified, to just route them through? --Amalthea^Talk 22:07, 20 August 2008 (UTC)

Why would anyone want a tool that acts differently depending on what date is fed into it? Such a tool is what lawyers like to call an attractive nuisance. --Gerry Ashton (talk) 22:20, 20 August 2008 (UTC)

{{date}} does not handle dates prior to 1970 or 1901 well, hence my attempt at {{Date style}} metatemplate (see there for comparison, but shortly to be tweaked to less wikilinking code of my sandbox User:Davidruben/sandbox3, see Template talk:Cite web#Review of current and sandbox coding - examples for comparison table). A major problem is that one can't in mediawiki directly read whether an editor has set a personal preference (only direct wikilinking an ISO-style date responds to a user's perference, but then if no user preference was set this always then shows just as the ISO style... i.e no best of both worlds of follow a user's preference if set else an editor's preference) David Ruben ^Talk 00:07, 21 August 2008 (UTC)

All of these tools: we just don't need them. Tony (talk) 00:35, 21 August 2008 (UTC)

• What Tony said. Exactly. Editors should write out the damned dates and should choose a format most suitable for the subject matter. If it’s a coin toss, no big deal; there is no “wrong guess” since no one is confused by any date format that has the name of the month spelled out. If any editor thinks that dates should somehow look *better* for just us editors, then don a pair of x-ray glasses. Other than that, please, try to accept the fundamental concept that it is wrong wrong, wrong, to use tools that produce pretty (or prettier) output that only registered editors can see. If some date formatting tool produces a format that you think is good enough for “regular I.P. readers” to see, then fixed text in that same format is good enough for us registered editors to have to look at too. If any editor here disagrees with this premiss, then please produce your “I am really, really *special* license” for inspection.

  Junking these tools will also reduce the link clutter—to mind-numbing trivia no less—that has turned too many Wikipedia articles into seas of blue spaghetti. Way too much effort and keyboard pounding has been devoted to an issue that is just a problem of our own making. Greg L (talk) 04:31, 21 August 2008 (UTC)

I have to agree with the argument that a presentation tool which does not work for anyone except wiki editors is pretty pointless, and worse, potentially misleading to those editors trying to set up a page. I never saw the point of using date formatting and in 3 years have never set it for myself. I don't care how the date comes, though I would naturally write 13 June 1844. I can live with wierd transatlantic eccentricities, but it was my understanding that some editors could not: thus the compromise of introducing date formatting. Which, it has now been pointed out, really achieves nothing for purists in either camp if the effect is simply to pass on the text as written to most readers. Of course, if a template was introduced which defaulted to one style or another for every reader, then the war would no doubt break out again with great force. The fact that date autoformatting doesn't work except ofr die-hards may well have been its greatest attraction. Sandpiper (talk) 08:10, 21 August 2008 (UTC)

Earlier in this "string", Amalthea referred to a "skinning system". What is that? FWiW Bzuk (talk) 13:15, 21 August 2008 (UTC).

See WP:SKIN - you can set it in your preferences, and it changes the stylesheet you get. --Amalthea^Talk 13:23, 21 August 2008 (UTC)

• Which states as follows: “…the presentational style of the pages can be changed, provided you have a Wikipedia account.” Who cares? Probably greater than 99.9% of Wikipedia’s readership doesn’t have an account. We need to be looking at what they’re seeing—not providing a *special* view of pages designed just for us. Greg L (talk) 16:20, 21 August 2008 (UTC)

• Unless there is a plan afoot to go even further than this and completely remove this table from MOSNUM, can we agree that to avoid really junking up articles for the vast majority of readers, that the bottom tool should at least have the double-asterisked caveat shown below?

What you type	What logged-in registered users see (settings on first row)					What others will see*
--	January 15, 2001	15 January 2001	2001 January 15	2001-01-15	No preference	--
[[May 15]]	May 15	15 May	May 15	May 15	May 15	May 15
[[15 May]]	May 15	15 May	15 May	15 May	15 May	15 May
[[May 15]], [[2005]]	May 15, 2005	15 May 2005	2005 May 15	2005-05-15	May 15, 2005	May 15, 2005
[[15 May]] [[2005]]	May 15, 2005	15 May 2005	2005 May 15	2005-05-15	15 May 2005	15 May 2005
[[2005-05-15]]	May 15, 2005	15 May 2005	2005 May 15	2005-05-15	2005-05-15	2005-05-15 **
* Non-registered users and registered users not logged in ** Editors are discouraged from using this format since the vast majority of readers  (non-registered I.P. users) will see a hard-to-read date format.

Is there anyone who thinks that having pretty looking dates for just we editors justifies having MOSNUM tacitly recommend the use of this tool, which A) looks like crap for 99.9% of readers, and B) links to mindless trivia that is almost always entirely unrelated to the article’s content? To save you all the time necessary to wade through previous posts on this page, waiting around for developers to fix this tool so it works equally as well for regular I.P. users would require very substantial changes to the way Wikipedia’s inner workings work and just isn’t a realistic option. So as a practical matter, we need to limit the discussion here to what should be done with the current tools until such time that a better one is provided.

Having said all that, I do see one reasonable alternative: Is there a way to revise the bottom format so it defaults to “15 May 2005” for I.P users? And while we’re at it, we might as well get a non-breaking space between the month and date; the current one word-wraps. If this can’t reasonably be expected to be accomplished in the very near future, let’s get the asterisks in. If not that, then I’m perfectly happy to sit this one out on the sidelines as the realization better propagates throughout the editing community that all these tools, which benefit only we editors (and make links to trivia to boot), were a bad idea from the beginning, and this entire table is finally deleted. Greg L (talk) 19:02, 21 August 2008 (UTC)

• I agree to adding the disclaimer, but would recommend moving the double asterisks to the leftmost "What you type" column (and hope that no one will then actually type those).
  I'm not sure what this will do to e.g. the citation templates which ATM rely on the ISO format for the access date. I could imagine going either way here, remove the wikilinking or still allow ISO dates there (since 99% of our readers won't read the references anyway). --Amalthea^Talk 21:56, 21 August 2008 (UTC)

Here's a thought- why not get rid of the citation templates, since they remain "buggy" and are almost impossible to figure out for the new editor? FWiW, authors routinely "key" in information for editing purposes, it's not an impossible task. It then reverts to WYSIWYG. Bzuk (talk) 22:05, 21 August 2008 (UTC).

• -Amalthea, I was thinking about putting the double-asterisk in the “what you type” column, but I thought people would think you should type the asterisks. Unjustified concern? To address this, how about this way?

What you type	What logged-in registered users see (settings on first row)					What others will see [A]
--	January 15, 2001	15 January 2001	2001 January 15	2001-01-15	No preference	--
[[May 15]]	May 15	15 May	May 15	May 15	May 15	May 15
[[15 May]]	May 15	15 May	15 May	15 May	15 May	15 May
[[May 15]], [[2005]]	May 15, 2005	15 May 2005	2005 May 15	2005-05-15	May 15, 2005	May 15, 2005
[[15 May]] [[2005]]	May 15, 2005	15 May 2005	2005 May 15	2005-05-15	15 May 2005	15 May 2005
[[2005-05-15]] [B]	May 15, 2005	15 May 2005	2005 May 15	2005-05-15	2005-05-15	2005-05-15

Greg L (talk) 23:25, 21 August 2008 (UTC)

I allowed myself to convert your footnoted table above to proper, working footnotes. --Amalthea^Talk 23:47, 21 August 2008 (UTC)

If DA is to be deprecated, I'm unsure whether this table belongs at all in MOSNUM. Tony (talk) 04:45, 22 August 2008 (UTC)

• One step at a time, I think, would be best here Tony. If you toss a frog into hot water, it will hop out. But if you put it into cool water and turn up the heat, it will stay in the water and swim around until it dies. Since editors around here are so emotional about anything being “D-worded,” moving the table off of MOSNUM will help prevent editors from thinking it is a “tool of good standing.” And by keeping the table available somewhere else, we can refer to it for reference as we further debate our options. Greg L (talk) 14:57, 22 August 2008 (UTC)
  • Autoformatting should be documented somewhere as long as it exists; otherwise we will have bug reports by people who don't understand that they accidentally set their preferences. Introductory text of the form: This is what autoformatting does; many editors feel that it should not be used. would solve any problem. Septentrionalis PMAnderson 17:39, 22 August 2008 (UTC)
    • I completely agree with Pmanderson. Greg L (talk) 23:03, 22 August 2008 (UTC)

RFC on September 11, 2001 attacks

Talk:September 11, 2001 attacks#RFC on page title and comma - We need outside opinions on what the appropriate grammar is here. Should the page title and the article start out with "September 11, 2001 attacks" (no comma) or "September 11, 2001, attacks"? A third option is to rename the page to something like "September 11 attacks". We would appreciate comments on the article talk page. Thanks. --Aude (talk) 20:13, 20 August 2008 (UTC)

• With regard to “The September 11, 2001 attacks…”, it’s clear what it means without the comma before the word “attacks”. It seems terribly awkward and improper to me to write “The September 11, 2001, attacks…”. As for what the *formal* rule is, by daughter could answer that (she’s damned good), but she’s not here now. I see no reason to rename the article. Greg L (talk) 20:21, 20 August 2008 (UTC)
  • It's clumsy with or without the comma. I would follow idiom: September 11 attacks, if the various forms of 9-11 aren't dignified enough for us. Septentrionalis PMAnderson 20:25, 20 August 2008 (UTC)

Comma after "attack" is awful. Either "September 11, 2001 attacks" or Anderson's option. Tony (talk) 00:32, 21 August 2008 (UTC)

Unfortunately, the term "9-11" has entered the lexicon of the event and at least some acknowledgment of this date name is necessary, although I would put it in quotation marks once and then use an appropriate written out date convention thereafter, e.g. "September 11, 2001". FWiW Bzuk (talk) 12:10, 21 August 2008 (UTC).

The first words of the 9/11 Commission's report (500+ page PDF) are "THE 9/11 COMMISSION REPORT" so 9/11 must be regarded as an official term for the attack. Although I cannot find a reliable source to confirm my memory, I recall that the air pirates were reported to have chosen the date of the attack because of the similarity of the numeric date to the phone number used in the US to summon emergency help (911). So it is not a matter of having entered the lexicon, it was a deliberate choice of the air pirates from the very beginning. --Gerry Ashton (talk) 18:18, 21 August 2008 (UTC)

Is that the actual origin of the 9-11 use? It's certainly an interesting historical footnote if it's true and should be mentioned somewhere in the article. My point is that the use of the date has been morphed into "9-11" rather than the actual date. FWiW Bzuk (talk) 18:24, 21 August 2008 (UTC).

It's one of several theories. Septentrionalis PMAnderson 18:48, 21 August 2008 (UTC)

Units and readability

An editor has raised an interesting point about the balance between readability and availability of metric units. Marcia Wright said on my talk page (selected quoting):

• I believe that this is much easier to read:
  • Lassen National Forest is a 1.1 million-acre national forest located in northeastern California.
• as opposed to this:
  • Lassen National Forest is a 1,100,000-acre (4,500 km2) national forest located in northeastern California.
  • with the metric units:
• The readability should not be reduced just to have conversions added to the article.

I believe that Marcia is acting in good faith and from her contributions I believe her to be a good Wikipedian. However, she has raised an important issue that comes up from time to time on an international publication. I happen to think that metric units should be provided even if this looks unfamiliar (compared with other specialist or regional publications) or if it degrades readability (by whatever measure we might use). If you look at her contributions, you will see that she helps improve articles about American geography. Please look at those articles and see what you think about this issue of metric units vs readability. Lightmouse (talk) 15:13, 21 August 2008 (UTC)

I agree that in the above example the first version is preferable. But I take this to be an introductory sentence - I presume somewhere later on in the article the area of the forest will be addressed in a separate sentence or paragraph, and there both imperial and metric units should be given.--Kotniski (talk) 15:34, 21 August 2008 (UTC)

Please look at the article and see. Lightmouse (talk) 15:38, 21 August 2008 (UTC)

Following up on the article Lassen National Forest, the version used there, "Lassen National Forest is a 1.1 million-acre (4,500 km2) national forest located in northeastern California." seems to be the best choice. What am I not understanding? is this a case of deprecating metric systems? FWiW Bzuk (talk) 15:55, 21 August 2008 (UTC).

I originally added two metric conversions and they were both removed. Marcia asked for a means of preventing the template being in the article because of page load time. See our conversation. So I suggested that she made manual edits but also started a thread about load time statistics at Template_talk:Convert. I now see that she has added one of the two conversions in a manual form and says that issue is about the readability of '1.1 million <units>' versus '1,100,000 <units>' (which I have some sympathy with). The story is a bit confusing for me. I am not sure where we were, and where we went along the way, but I think that we are almost where we want to be i.e. with both non-metric values having a metric equivalent visible to the reader. Lightmouse (talk) 16:06, 21 August 2008 (UTC)

Americans have to realize that theirs is a rather parochial attitude to the world. There are approximately 2,000,000,000 people in the world who can read English and who might potentially read that article. Of those, about 15% live in the United States. Most of the rest do not know what an acre is, and you can't seriously expect them to go and look it up in the midst of reading it. Even in Britain and its former colonies, most children do not study the Imperial system any more and do not really know what the units mean. Aesthetic differences aside, would it be more readable if I said something was 1.1 million arpents or 1,100,000 arpents in area?RockyMtnGuy (talk) 16:28, 21 August 2008 (UTC)

Which is why to link acre here; but this article is unlikely to read, and extremely unlikely to be written, by non-Americans - so WP:ENGVAR applies. Septentrionalis PMAnderson 16:54, 21 August 2008 (UTC)

I am not sure what you mean by the above comment, but I see that you have edited the article to move the conversion to a footnote. Taking conversions out of running text is another barrier to accessibility for metric readers. Lightmouse (talk) 17:08, 21 August 2008 (UTC)

RockyMtnGuy seems not to understand that this article is, and should be, written in American, so I cited our guidance.

WP:ENGVAR applies to the article, but WP:ENGVAR does *not* say anything about how to write units—WP:UNITS does, and it says that both units units should be given (main unit being US in this case). --Amalthea^Talk 18:09, 21 August 2008 (UTC)

• Parentheses are clumsy writing, a barrier to accessibility to all readers, metric or not. The exact acreage, and a conversion, are in the infobox for anyone who wants them. A link to acre remains in the text, and the footnote to the rough conversion for anyone who cannot recognize that more than a million of any plausible unit is likely to be a big area. Septentrionalis PMAnderson 17:21, 21 August 2008 (UTC)

That looks pretty goofy. What was wrong with "1.1 million acres (4,500 km²)"? That's a pretty standard presentation, isn't it? And why link to "acre"? It's pretty obvious with the conversion that it's a unit of area. --Elliskev 17:29, 21 August 2008 (UTC)

I believe you stretch ENGVAR too far here, Anderson, or is writing such that everyone can understand un-American? 1.1 million-acre national forest vs 1,100,000-acre (4,500 km2) is unfair. Either 1.1-million-acre national forest vs 1.1-million-acre (4,500 km²) or 1,100,000-acre national forest vs 1,100,000-acre (4,500 km²). Let Marcia call it as she will but as a metricated mo I see the readability significantly reduced by the removal of square kilometres. JIMp _talk·cont 17:32, 21 August 2008 (UTC)

I call it significantly (and pointlessly) reduced by the parenthesis; but I will change to million-acre to make clear this is a round figure. WP:MOS (and MOSNUM, especially) should not be used to enforce bad writing; doing so is a disservice to Wikipedia. It would be better to delete them than have them subserve the semi-literate version of political correctness. Septentrionalis PMAnderson 17:46, 21 August 2008 (UTC)

How about this? I added a new unit called "e6acre" (million acres) to Template:Convert. Now we can use it whenever we need to talk about millions of acres. with adj=on it gives 1.1-million-acre (4,500 km²). --Random832 (contribs) 18:00, 21 August 2008 (UTC)

e3acre might also be good ... no, don't panic we're not about to insert {{ucfirst:{{convert|1|e2acre|adj=on}}}} Wood into Winnie the Pooh. JIMp _talk·cont 18:24, 21 August 2008 (UTC)

It really doesn't matter how we get an unreadable parenthesized lump; it's still abominable writing. Septentrionalis PMAnderson 18:40, 21 August 2008 (UTC)

In an introductory paragraph, the goal is to give the reader a general concept of how large the forest is. Neither 1.1 million acres nor 4300 square kilometers really gives me a good concept, but at least with the metric figure, I can mentally note that if it were square, each edge would be between 60 and 70 kilometers. Doing any equivalent mental arithmetic with the customary figure is beyond me. So for me, the 4300 square kilometers figure is more useful. A corresponding customary description would be "a 1700 square mile national forest." --Gerry Ashton (talk) 18:54, 21 August 2008 (UTC)

I agree the 1.1 doesn't serve much purpose; which is why it now reads million-acre forest. "1700 square mile" would be fine too; the whole thing belongs under our clause about four minute mile. Septentrionalis PMAnderson 19:02, 21 August 2008 (UTC)

No, it doesn't matter and yes, parenthesized-lumpless writing is preferable, but it's better than writing something that just doesn't make sense (i.e. imperial/US units to a metric minded reader or metric units to an imperial/US system minded reader). JIMp _talk·cont 18:58, 21 August 2008 (UTC)

I fail to see that four-minute-mile connexion. JIMp _talk·cont 19:04, 21 August 2008 (UTC)

Idiomatic expressions should not be spoiled by reflex conversion. Septentrionalis PMAnderson 19:07, 21 August 2008 (UTC)

Absolutely but a measurement of an actual park out there in the world is not an idiomatic expression ... inspite of any rewording to make it look like one. JIMp _talk·cont 19:16, 21 August 2008 (UTC)

The four-minute mile clause applies to common expressions, not common forms of expression. It would not apply, for example, to "the 0 to 60 time of the Jeep Cherokee was 11 seconds" because the next time the form of expression was found, the vehicle and time would most likely be different. This makes sense, because if it is a common idiom in the English language, even readers who don't use the system of units in the expression as their first choice will learn the meaning of the idiom once and for all, so won't need to do calculate a conversion to familiar units each time the expression is encountered. --Gerry Ashton (talk) 19:19, 21 August 2008 (UTC)

I second what Gerry wrote. JIMp _talk·cont 19:29, 21 August 2008 (UTC)

I agree with Lightmouse's edit. I don't think in terms of square kilometers, but I know that there are people (like Lightmouse) that do. The reverse, adding imperial units, can and should be applied to say a National Park in South Africa that is 450000 hectares—i.e. 450,000 hectares (1,100,000 acres). Sorry, but Marcia Wright is not right on this one. Since I don't think in terms of km2, I just skip over the (4,500 km²) part anyway when reading. Therefore, the readability is not affected for me. —MJCdetroit ^(yak) 20:42, 21 August 2008 (UTC)

Notwithstanding the aesthetics of the four-minute mile, whether it's the statute mile, the metric mile, or the 1500 metres, the fact is that if you put something out on the World Wide Web, everybody on the planet can read it. Case in point - I led a hike in the Canadian Rockies recently. The club I belong to posted it on our Web site. Of the people who contacted me about it, one was from the Netherlands, one was from Israel, two were from France, two were from Japan and only one was from Canada. Since most of them didn't understand feet or miles, we calibrated our altimeters in metres and our GPS units in kilometres. So, if you put something out on the World Wide Web, be aware that anyone on the planet can read it, and not everybody has the same standards as you.RockyMtnGuy (talk) 21:20, 21 August 2008 (UTC)

I am about to invoke the "too many cooks" clause (LOL) as the "improvements" are not there; putting information in a reference note is not a good solution. FWiW Bzuk (talk) 21:28, 21 August 2008 (UTC).

• Lightmouse: I think your first instinct was spot on (but with a modification): That particular article should be written as follows:

“ Lassen National Forest is a 1.1 million-acre (4,300 km²) national forest located in northeastern California. ”

Why? Well…

1. The text you propose is much more fluid and natural to read and is the way one would find such information in a U.S. National Parks’ brochure.
2. The existing text here (“is a million-acre national forest”) seems unduly informal and non-encyclopedic. Since the value is known to high precision (at 1,070,344 acres), going to two significant digits is entirely appropriate for readers interested in the relative sizes of national parks.
3. The article is about a U.S. National Forest and is therefore U.S.-centric. So it should have acres first for that audience.
4. The parenthetical should be provided for the benefit of other English-speaking peoples of the world and in the unit of measure most common for SI-using countries. The style I chose for my example here is, I believe, in accordance with existing guidelines on MOSNUM.
5. It conforms with other MOSNUM guidelines, which calls for the first instances of the primary units of measure to be linked, and for the first instance(s) of the primary measure to be spelled out. The unit of measure in the parenthetical conversion doesn’t need to be linked because it’s magnitude is obvious when juxtaposed next to the primary measure.

P.S. I don’t see the value of the footnote since the exact size is shown in the sidebar infobox.

PPS I did my own conversion and the actual size is 4331.5 square kilometers, so it’s “4,300”, not “4,500”.

Greg L (talk) 21:45, 21 August 2008 (UTC)

I absolutely agree, see comments I made a zillion spaces up. FWiW Bzuk (talk) 21:50, 21 August 2008 (UTC).

I unconditionally disagree. Bad writing is bad writing; and we don't need to convert an intentionally approximate wording when there is a precise statement and conversion on the other side of the page. Septentrionalis PMAnderson 00:07, 22 August 2008 (UTC)

If this is turning into a poll: I agree with Greg L, but can also live with the current version per Gerry Ashton. Square miles gives me (as a metric) a rough estimate that I can grasp, compared to acres which means nothing to me, but I would always prefer the real deal, per MOS:CONVERSIONS. --Amalthea^Talk 00:29, 22 August 2008 (UTC)

So what is wrong with this rewording to avoid the awkward triple item:

Lassen National Forest is a national forest of 1.1 million acres (4,500 km2) in northeastern California

The example wasn't helped by a redundant word. I increasingly see a tendency to resort to multiple hyphenated units, but there's always a more striaghtforward alternative. With conversion, they are better avoided. Another point: how many American readers will be able to conceptualise a million acres? What is wrong with square miles? Tony (talk) 04:41, 22 August 2008 (UTC)

PPS I did my own conversion and the actual size is 4331.5 square kilometers, so it’s “4,300”, not “4,500”. Unfortunately, this problem is due to the double rounding inherent in use of the convert template: the input is given as 1.1 (2 significant figures) and then it converts 1,100,000 acres to km2 (4,451) and then rounds _that_ to 2 significant figures. --Random832 (contribs) 14:23, 22 August 2008 (UTC)

A reason to be careful with {{convert}}. But Marcia was right to begin with: Lassen National Forest is a national forest of 1.1 million acres (4,500 km2) in northeastern California is clumsy and difficult to read; as an opening sentence, it offers metric raaders very little, all of it available elsewhere. If there is a Siberian reserve somewhere notable because it is over a million hectares, a rational American reader will accept that as a summary, and if xe wants an exact area, look further in the article. Septentrionalis PMAnderson 14:32, 22 August 2008 (UTC)

I'm rather surprised by this long discussion on a single example. I thought the issue had been decided once and for all a long time ago. There is a clear guideline to always provide conversions for non-SI units. Readability is not determined by style only, but also by presenting comprehensible information to the reader (and outside the U.S.A. nobody knows what an acre is). −Woodstone (talk) 14:53, 22 August 2008 (UTC)

Please quote the wording which would apply such a mandate to this situation, so that I may dispute it. Any guidance which makes this encyclopedia less readable without making it more clear should be ignored. Septentrionalis PMAnderson 15:47, 22 August 2008 (UTC)

See Wikipedia:Manual of Style (dates_and_numbers)#Unit_conversions. There are a few exceptions, but none apply here. In your opinion it may reduce readably in mine it enhances it. No need to start scolding. −Woodstone (talk) 18:01, 22 August 2008 (UTC)

At the risk of getting pointy, I believe Lassen is in Northern Paiute territory, so perhaps the local unit should be square rabbit-skins per Harry W. Gilmore "Hunting Habits of the Early Nevada Paiutes" American Anthropologist, New Series, Vol. 55, No. 1 (Jan. - Mar., 1953), pp. 148-153

But seriously, if all the fuss is about the parentheses, perhaps we should consider more verbose ways of presenting the conversion. One might suggest something of the form: Lassen National Forest is a US national forest in northeastern California occupying 4331.5 km², locally described as 1,070,344 acres. LeadSongDog (talk) 19:28, 22 August 2008 (UTC)

I haven't read the whole discussion, but conversions should stay in one form or another. I managed to master the languages in all the places I've lived, but never the conversions. My entire household loves them; I can simultaneously read a number on Wiki that I understand, and a number the rest of the family understands. It's one of our strengths; let's not lose it. I can't really relate to complaining about reading around one set of parentheses. SandyGeorgia (Talk) 02:49, 23 August 2008 (UTC)

Proposal to move autoformatting table to its own page

The autoformatting table is a detailed explanation of how a tool works, not style guidance. The convert template is mentioned in MOSNUM and we provide a link to its explanation page. I propose that we do the same with autoformatting i.e. the details are moved to an explanation page that we mention in MOSNUM. Lightmouse (talk) 22:26, 21 August 2008 (UTC)

• I was thinking the exact same thing this morning. I wholeheartedly agree. Greg L (talk) 22:47, 21 August 2008 (UTC)

• I took the liberty of adding the footnote-style note regarding the bottom option in advance of any move. Greg L (talk) 23:31, 21 August 2008 (UTC)

Excellent idea. Tony (talk) 04:34, 22 August 2008 (UTC)

Deprecating Date Autolinking vs. removing Date Autolinking

I think it’s pretty clear from the preceding discussions that a consensus exists for discouraging further application of DAs. Furthermore, when it’s clearly explained why it’s being deprecated, many originally dissenting editors have changed their mind. Considering that this is a major change in the expected appearance of Wikipedia articles, I think this MOS change should be presented in the VP, along with the rationale. Furthermore, I believe a simple statement of the rationale should also appear in the passage in MOSNUM because when people understand the rationale, they’re more disposed to accept it.

That said, there is another related issue that I don’t believe there is consensus on (and reading Masem’s recent posts, it is assumed by some as also a given), and that is the issue of mass removal of existing DAs. As can be noted by the protests of Tony’s trial removals in several articles and a few other editors’ personal crusades, the jury is out on that issue. Implementing such large-scale removals without a consensus of the community as a whole – vice a limited number of interested parties here – can be expected to be no less disruptive than that incurred over the mass deletions of non-free and a majority of fair-use images (which at least had legal reasons behind it, not merely stylistic ones). Face it, a lot of work has gone into inserting them (in accordance with what most editors have understood to be the preferred approach), and ripping it all out at once, with little or no input (much less warning), will lead to a lot of unnecessarily hurt feelings and the loss of a number of discouraged or angry editors. Wikipedians hate being dictated to by a minority of experts.

Accordingly, I’d encourage asking the community’s preference for eliminating them. Mass, perhaps bot-automated, removal is one option. Another is to leave them to the consensus of the editors of each article to decide. (Then they become part of the solution.) I’d prefer the second option to be the default until such time as there is a public consensus for the former. Askari Mark (Talk) 03:16, 23 August 2008 (UTC)

Another option to offer is that of turning off the ability of logged-in users to specify a preference. That is, offer the option of forcing the .01% to see what the 99.99% are forced to see. Then perhaps change will happen on its own! (sdsds - talk) 05:17, 23 August 2008 (UTC)

Bots already remove unwanted linking from date fragments; now that linking whole dates is deprecated too, it seems only logical that the same bots should remove those links as well. It might be a good idea to announce the change loudly to the community, via all the centralized discussion pages, maybe even with one of those notices that appear automatically at the top the screen sometimes (I know they're called something but I don't remember what). That way, when the bot starts removing the links, people will understand that it's the result of a much discussed and widely accepted change, and should be generally less hostile. Improvement of the encyclopedia must in the end take precedence over a few editors' sensitivities.--Kotniski (talk) 08:03, 23 August 2008 (UTC)

Clarification of wording

Our well-established guidelines for which of the two standard date formats should be used in an article, here, say this:

*Articles related to other countries that commonly use one of the two acceptable formats above should use that format.

This is all very well for articles, say, on contemporary Venezuala, which uses US spelling in its English, but international date format; however, I've heard editors say that "hey, this medieval topic shouldn't be in US formatting", or "Beethoven should be changed to international" (it was a couple of months ago, on that very basis).

I think this is an unintended consequence of the guideline that wasn't thought through at the time. I'm not at all comfortable with the implications of the current wording: I believe it should be left up to the editors of each article to decide (usually on the basis of the first editor's choice), as is widely the case until the occasional editor points to this guideline. It matters nought to me that 19th-century Russian composer Tschaikovsky is written in US format.

May I propose an addition to stop this happening? It's longer than I'd hoped for, but I think it does the trick:

*Articles related to other countries that commonly use one of the two acceptable formats above should use that format. (However, this should not be interpreted as requiring per se the conversion of US to international format where a topic, particularly a historical topic, is largely connected with locations in non-anglophone countries that do not or did not use the modern US format.)

My only misgiving is that there may be widespread support for transforming such articles to international format; I've yet to see this expressed, though. Your thoughts, please? Tony (talk) 05:31, 23 August 2008 (UTC)

The above proposed addition is not a clarification, but a change of meaning. It seems natural to me that an article about a person who never set foot in the U.S.A. should not use U.S.A. date format. −Woodstone (talk) 09:52, 23 August 2008 (UTC)

My only misgiving is that there may be widespread support for transforming such articles to international format. I can't see a problem with this. In fact, Wikipedia being an international effort, why not have all dates in international format? --Pete (talk) 10:31, 23 August 2008 (UTC)

The current guidance is:

• 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.
• Articles related to other countries that commonly use one of the two acceptable formats above should use that format.

I find the current guidance difficult to parse. I propose that the bullets be simplified to read:

• Articles on topics with strong ties to Canada can use either format
• Articles on topics with strong ties to a country other than Canada should generally use the more common date format for that country

This does not affect Tony's proposal. Lightmouse (talk) 11:02, 23 August 2008 (UTC)

Lightbot removing links to volume units within infoboxes

Moved from Lightmouse talk page: begin
Please leave lk=on on {{convert}} in infoboxes. -- User:Docu

In all of them? Lightmouse (talk) 22:12, 21 August 2008 (UTC)

I'd very much support this. Leaving links in infoboxes doesn't disrupt prose, but at the same time provides a link for those people who really don't understand the meaning of certain measurements (after all, the site is supposed to educate). Is it possible for Lightbot to detect Infoboxes, or better yet, all templates (aka, anything inside {{ }}), and leave the contents alone? — Huntster (t • @ • c) 19:00, 22 August 2008 (UTC)

Wikipedia:Only make links that are relevant to the context says In general, do not create links to ...Plain English words, including common units of measurement. It goes on to give examples of common units of measurement. The usual reason given for links is to provide access to a conversion factor. There are some people that suggest that even 'second tier' units (i.e. not common but not obscure either) should not be linked when a conversion is provided. I agree with you that a link will help if people do not understand the meaning of *certain* measurements where they are obscure. Can you give me an example edit that you disagree with? Lightmouse (talk) 08:54, 23 August 2008 (UTC)

Personally, I think the units for volumes on {{infobox lake}} should be wikilinked. This reflects the consensus for that infobox. -- User:Docu

Moved from Lightmouse talk page: end

What do people think? Lightmouse (talk) 09:32, 23 August 2008 (UTC)

Removal of metric units

There seems to be some confusion about when it is appropriate to remove metric units. With Lassen National Forest, metric units have been removed and editors have cited three separate reasons for doing so:

• conversion templates are a burden for the reader
• numeric formats are a problem
• Pmanderson's edit with a summary that does not mention removal.

Are metric units really that bad for readers that they should be removed from articles? Lightmouse (talk) 09:50, 23 August 2008 (UTC)