Wikipedia talk:Manual of Style/Dates and numbers

RfC: Linking of dates of birth and death

Proposal: to add the words

These dates should normally be linked.

to the section WP:MOSDAB#Dates of birth and death, and to link the example dates, so the section would read

At the start of an article on an individual, his or her dates of birth and death are provided. These dates should normally be linked. For example: "Charles Darwin (12 February 1809 – 19 April 1882) was a British ..."

• For an individual still living: "Serena Williams (born September 26, 1981) ...", not "... (September 26, 1981 –) ..."
• When only the years are known: "Socrates (470–399 BC) was..."
• When the year of birth is completely unknown, it should be extrapolated from earliest known period of activity: "Offa of Mercia (before 734 – 26 July 796) ..."

...

Rationale There are some - most vocally perhaps Tony - who believe that pretty much no dates should be linked; and this seems to be what Lightbot was trying to achieve, too. But I don't believe that is the view of the majority. On the contrary, I think the balance of opinion, even amongst those who don't want to see pages becoming a "sea of blue", is that it is useful to have at least some date links on a page, to let people establish a broader context for the times in which a person lived, by clicking their way through the date hierarchy especially via pages like List of state leaders in xxxx or xxxx in the United Kingdom, etc. The proposal that at least the date of birth and date of death in a biographical article should be linked has been made independently in at least four different threads: by Scolaire in the section above #Dates are not linked unless; by Carcharoth in the section above #Concrete examples (year links); by Eleassar, relaying a question raised to him in talk, at WT:CONTEXT#Birth dates?; and by myself at User talk:Lightmouse#Date linking request (birth and death years). It therefore seems appropriate to put up this proposal specifically as a formal well-advertised RfC. Jheald (talk) 11:03, 29 September 2008 (UTC)

• Support. The date page hierarchy, and pages rapidly linked from it, provides a useful link to historical context for biographical articles. The biographical articles are stronger for such context; and the birth date and death date are the most obvious choice of dates to link. Jheald (talk) 11:03, 29 September 2008 (UTC)
• Oppose. (1) You're linking an anniversary day and month that is useless to our readers (please demonstrate some that are useful, and not just a magic carpet for discretionary browsers); and many editors will confuse this with the old autoformatting function. (2) Did you mean to "nowiki" the laborious constructions above that are concealed behind the piped linking (([[12 February|12 February]] [[1809]] – [[19 April|19 April]] [[1882]])? I'm sure this will go down very well with editors, who who will not only have to memorise how to do this, but will have to actually do it in every article. (3) You haven't demonstrated why it is worth forcing editors to make a link to a year page (birth/death): while it might be possible in a few rare instances to argue that the year of death page is vaguely useful (e.g., 1963 for the death of JF Kennedy, but even that example demonstrates how the fragmented facts about JFK in that year are better in the JFK article itself, or a daughter article on the assassination). (4) The "year in X" links are fine, except that concealing them behind what looks like a useless year-link is self-defeating, isn't it? Already, at least one WikiProject says not to use them. MOSLINK recommends the use of explicit wording to overcome the concealment. Tony (talk) 11:49, 29 September 2008 (UTC)
  • No, this has nothing to do with autoformatting. I'm proposing that such dates - the year, and the day - in the opening words of a bio article should be linked, end of story; something a number of other editors have also raised. The principal value being for the context that these links, and onward links from such pages, allow readers to click through to and explore.
    I'm not talking about "Year in X" articles, I'm talking about the bare year articles themselves. And I'm not intending to particularly mandate the   characters - they were there already, so I just left them. My proposal is very simple: as a rule, the days and years in those opening words should be linked. I want to see where the balance of the community rests on that question. Jheald (talk) 12:06, 29 September 2008 (UTC)
• Oppose, per many of Tony's comments, and just the fact that these year articles (much less day of the month ones!) don't provide useful historical context, they provide an often enormous list of trivial crap. If a large and well-organized WikiProject were capable of producing actually useful year articles that summarized the truly notable happenings in those years, I could maybe see the linking of years (only) for birth/death/establishment/disestablishment dates (only, for the most part). The problem with this though is that editors will see them linked in the lead sentence and then go around linking them all over the place, and we'd be pretty much back where we started. — SMcCandlish [talk] [cont] ‹(-¿-)› 11:55, 29 September 2008 (UTC)
• Support. I opposed delinking dates in the first place and I still do. -- Necrothesp (talk) 11:57, 29 September 2008 (UTC)
• Oppose, I've never seen much sense in date linking, and links to day-of-the-month articles result in triviality amost by definition. Fut.Perf. ☼ 12:41, 29 September 2008 (UTC)
• CommentSupport. I agree this is a good question to work out. My question is whether we should use what I think you are proposing, the well known and much disliked, "link to the day of year", "link to the year" (which is why people are asking about autoformatting), or if we should be suggesting {{Birth date|yyyy|mm|dd}} and {{Death date and age|yyyy|mm|dd|yyyy|mm|dd}} which provides protection against lightbot and allows for more flexibility in the future. As for those who oppose the "trivia dumping grounds", I suspect that if the links are to specific types of narrowly defined data (such as Births on January 15, 1900 or People who share a birthday on 15 January) most people would be fine with that. dm (talk) 12:43, 29 September 2008 (UTC)
  • As an aside, I changed this to a support. The templates I mention are real templates, and dont need any development. dm (talk) 03:47, 1 October 2008 (UTC)
• ;;partial Strong Support. In that case, the first example should be ""'''Charles Darwin''' {{DL|y=1809|m=February|d-12|mode=eng}} – {{DL|y=1882|m=April|d=19|mode=end}}) was a British ...", with the details of the template worked out later. (And yes, if the question is whether the dates be linked in the lead sentence, my answer is strong support.) Disagree with secret links to 1990 births or 15 January birthdays / January 15 birthdays (if, for no other reason, we'd need staff monitoring which of the latter is linked to.) — Arthur Rubin (talk) 13:22, 29 September 2008 (UTC)
  • Changed to neutral on the day of the year, even if the day of the year article is the one that links back to the person, and the year article does not, because of inadequate notability. It should, however, be pointed out, that [[15 January|15 January]] [[1990]] would block autoformatting, and the only consensus we have is that autoformatting is bad, not the linking. — Arthur Rubin (talk) 01:58, 18 October 2008 (UTC)
• Support with comment There are many who do click and want to click on a date link to look at a reference in context, and whether that is trivial, banal or whatever is not my business, nor mine to judge. I can understand that someone may wish to click on a link to find out the context of a date of birth to the world around them at the time. Do I do it? No. Should it be allowable? Yes. For instance a child born during a battle in the local area, or being named Victoria, and that being the date of the coronation of Queen Victoria, or some other event that may have an effect on that person's environment. This information can be quite relevant. So the issue then becomes managing it, and making it useful. Is there 'overlinking' on dates, most definitely, and the information should be most specific, however, the request is specifically for Dates of Life. With regard to the comments about triviality ... for goodness sake, the difference between trivia and excellent knowledge is solely your own virtual framework and environment. If some people thrive on trivia, good luck to them, WP is here for all types. Not asking for extreme, let us find the median position. -- billinghurst (talk) 13:54, 29 September 2008 (UTC)
• Oppose. Linking some, but not all full dates in the article will be confusing. I don't see birth and death date-linking to be valuable at all. Most biographies do have categories for year of birth and death that would get your average browser to the year page anyway. Karanacs (talk) 14:21, 29 September 2008 (UTC)
  • Query to Billinghurst: Your assertion that many people click on and want to click on a date link seems unlikely—do you have sources for this? Tony (talk) 14:42, 29 September 2008 (UTC)
    • Thanks Tony. wikt:many Anecdotally from reading, especially the commentary when it was on User_talk:Lightmouse; some (light) discussions with genealogists, who are a little date focused. I too would love to see evidentiary information about date links and whether they are followed or not. If someone has the right wand to produce that data, it would be lovely. To Karanacs the proposal is just Dates of Life, not all dates. — Preceding unsigned comment added by Billinghurst (talk • contribs)
    • For tony to make such a statement that it is unlikely only proves that he is not paying attention to the comments being made against delinking of dates. I have stated on several occasions (as have others) that I do click on dates (sometime only to see if the article is associated to the date). As for evidence I recommend that someone does a query on the toolserver for all the date articles and see if the hits reduce over the next few months as more and more articles have the dates delinked. I believe we will find a marked reduction in the traffic to those date articles do to their delinking.--Kumioko (talk) 16:27, 29 September 2008 (UTC)
• Oppose. It adds complexity and I just don't see the value. Haukur (talk) 14:59, 29 September 2008 (UTC)
• Support I never agreed with Tony's 2 dimensional view that date linking is bad. Wikipedia is a 3 dimensional database of articles and is not bound by the 2 dimensional rules of a paper article. If we have an article in wikipedia that is linkable to an article then we should link to it (whether ir directly relates or not). That doesn't mean that it should be linked 4 or 5 times but it should be linked and the birth and death dates to me are reasonable. If we go along with this delinking of dates argument that tony presents then next we will be delinking the city and state of birth, military ranks, allegiances and any other link that is not directly related to an articles content. I think that this date argument sets a very ugly precedent. Additionally, given the volume of arguments for and against this venture it should be obvious to everyone (regardless of how they feel about whether dates should or should not be linked) that this does not meet consensus, regardless of how the vote previously came out.--Kumioko (talk) 15:06, 29 September 2008 (UTC)
• Strong but partial support I believe linking the year to the bare year articles for births and deaths in bio provides useful context information. I'm actually in favour of linking years (decades etc.) where ever the historical context is significant to the subject of the article, even if the subject itself is not significant to the period of time linked. However, I am not as convinced of the value of linking the month and day, especially since those links would not seem to add much context without the year. PaleAqua (talk) 15:53, 29 September 2008 (UTC)
  • Two simple points to PaleAqua: (1) Where was the consensus to link these items in the first place? (2) No one is suggesting a slippery slope to no wikilinking; rather, I sense that the motivation is the direct opposite: the encouragement of a stronger wikilinking system through the avoidance of extremely low-value dilutions. Tony (talk) 16:45, 29 September 2008 (UTC)
    • 1) Rhetorical statements are unhelpful. Where is so much of the information and documentation of templates, convention, etc. Wikis evolve, we are talking about a controlled evolution. 2) No, you are correct, no slippery slope suggested, it was Dates of Life only. Low value to you, statements to the contrary by others that dates of linking are not of low value seem to be ignored or derided as of low value. :-( billinghurst (talk) 17:12, 29 September 2008 (UTC)
• Support. It certainly does no harm. Also usefull for lovers of trivia. Let readers decide what they want to read. G-Man ^? 19:55, 29 September 2008 (UTC)
• Oppose. One of the dates in the example, 26 July 796, would be displayed to those who have selected the "2001-01-15T16:12:34" date format preference as 0796-07-26. The unique format in the preference menu clearly defines this date as an ISO 8601 date, even though that term does not appear on the menu. Also, the discussion leading to the implementation of date autoformatting makes it clear this format was intended to be ISO 8601. ISO 8601 requires dates to be in the Gregorian calendar, and requires mutual consent before information exchange partners exchange any date before the year 1583. Since the date 26 July 796 is in the Julian calendar, both requirements are violated. --Gerry Ashton (talk) 20:15, 29 September 2008 (UTC)
  • Comment: this discussion is not intended to be about autoformatting. This is about hard-linking of the dates, rendered as written, which is how 99% of readers will see them. If there are bugs in autoformatting, then there are bugs in autoformatting. User beware. But we shouldn't let the tail wag the dog. The question is, regardless of autoformatting, should these dates be linked? Jheald (talk) 20:50, 29 September 2008 (UTC)
    • You're wrong. Every date linking discussion is always about date autoformatting until the date autoformatting cancer is excised and incinerated. --Gerry Ashton (talk) 21:38, 29 September 2008 (UTC)
• Oppose Just because an article exists on Wikipedia that can be linked to, doesn’t mean it should be linked to. Links should be topical and germane to the article and should properly anticipate what the readership will likely want to further explore. Linking of years (1982), isn’t germane most of the time and should be limited to intrinsically historical articles like French Revolution—in which case, the linked dates would be older, like 1794. What the bot is doing that I find really valuable is the de-linking of dates (October 21). If someone was born on that date in 1982, no one gives a damn if “On this date in] 1600 - Tokugawa Ieyasu defeats the leaders of rival Japanese clans in the Battle of Sekigahara, which marks the beginning of the Tokugawa shogunate, who in effect rule Japan until the mid-nineteenth century.” This isn’t not proper technical writing practices. Greg L (talk) 20:25, 29 September 2008 (UTC)
  • I just want to clarify that just because you don't "give a damn" doesn't mean knowone does. If knowone cared then there would be no need to have a On this day section in the main page.--Kumioko (talk) 20:39, 29 September 2008 (UTC)
    • Indeed. Maybe some people are interested in who else shared the same birthday, or that an English rugby union star was born on the feast-day of the patron saint of McDonalds. If WP has these pages, I think it's inappropriate to presume that because WP:YOUDONTLIKEIT, nobody else should be allowed to find them. Jheald (talk) 20:50, 29 September 2008 (UTC)
      • To Jheald: So you cite WP:YOUDONTLIKEIT. That’s sort of a “if it’s blue, it must be true” argument; if there was a WP:I REALLY REALLY LIKE IT AND IF AN ARTICLE EXITS ON WIKIPEDIA, IT SHOULD BE LINKED TO essay, I might “prove” my point. To Kumioko: I have no problem with the “On this day…” on the main page because all readers know what they will be taken to if they click on a link; they aren’t Easter eggs. And to both of you: This isn’t an issue of right or wrong; it’s a grey area centered around the issue of not desensitizing readers to our blue links through excessive linking. These are links to trivia. Too few readers, after they’ve stepped on these date land mines, want to bother with them any more. Greg L (talk) 21:24, 29 September 2008 (UTC)
        • I agree that date and year links can be land mines and unhelpful to readers. Let's make that clear first. However, I seriously doubt that links that are clearly birth and death years will mislead readers in your "land mine" sense. Take this example: "Charles Darwin (12 February 1809 – 19 April 1882) was a British ..." In my view, people may wonder what the links are, but when they click on them will realise "ah, an article on the year, that makes sense". They will then know this when they see it on future articles, and either click through as desired, or ignore them. What they won't do, in my opinion, is click on the link and think "oh, I was expecting an article on this person's birth" or "oh, I was expecting an article on this person's death". i.e. when clearly linked in a specified and limited context (birth and death years), year links are not Easter egg "land mines", and they are not excessive linking (two links per biographical article). Carcharoth (talk) 04:17, 30 September 2008 (UTC)
          • I have observed that articles tend to be more heavily linked in the lead section; birth and death dates also tend to be the first to appear after the subject's name. Linking to these date articles would strongly contribute to the strong sea of blue in the opening paragraphs. While death dates may be consequential in certain cases, the only possible exception birth dates being generally a non-event is Jesus Christ, and nobody knows JC's exact birth date or year anyway, so I think this is a red herring of a debate. 219.78.19.154 (talk) 15:47, 2 October 2008 (UTC)
            • Well, Alfonso XIII's birth was probably an event... Anyway, your overlinking argument is a good one. If we are to link some dates in a biographical article, then it would make sense to link birth and death dates, but doing it in the lead is not very good. If we say "do it only in an infobox", plus get rid of the autoformatting, then I like it better. -- Jao (talk) 15:57, 2 October 2008 (UTC)
              • Actually, Darwin's birth did involve a minor event; it was the same day as Abraham Lincoln's. I should prefer to have this trivium availabe behind a link to restarting the proverbially WP:LAME edit war about whether it should be in the lead...
                More seriously, the year of birth does provide context, and would provice more if the year articles were better. On medieval articles, it is often of some interest on what saint's day a given person is born; and so on. Septentrionalis PMAnderson 02:59, 3 October 2008 (UTC)
• Support - While I believe that most dates should not be linked, I believe that, in biographical articles, dates of birth and death would serve as helpful links. We link to the biographical articles of persons born on a particular date on that date's article, so why not link back to the date from the biography? – PeeJay 20:45, 29 September 2008 (UTC)
• Oppose - I don't see the value of linking the dates of birth and death. The previous objections to all date linking still seem to apply. Day-of-the-month linking is still trivial even when the date is someone's birth date EdJohnston (talk) 21:17, 29 September 2008 (UTC)
• Oppose - extremely low value links. If someone is interested in the "context" of who else was born on September 12, they can type those few characters into the search box themselves. These are trivial connections that clutter articles needlessly. Ground Zero | t 21:42, 29 September 2008 (UTC)

*Oppose - I have yet to see any argument that comes close to convincing me that these date links provide any sort of relevant context. Yes, they provide context, but the context is so general that it seems useless to me. And yes, I have heard the argument that "just because it seems useless to you, doesn't mean it's useless to everyone." This is a valid argument, but only to a point. Linking every word in every sentence to Wiktionary would probably be more useful than this, in my view. And I don't think that one would get any massive rash of support, either.--Aervanath lives in the Orphanage 22:31, 29 September 2008 (UTC) (Changed !vote: see below)

• Oppose That would just makes everything more complex. Besides, I have yet to read a convincing argument on why date-of-birth and date-of-death links are necessary to aid the reader's understanding of the article's subject matter. Dabomb87 (talk) 02:20, 30 September 2008 (UTC)
• Support linking of birth and death years once at the appropriate place in an article (with the second option being a formal written support in the manual of style for using the birth and death year categories). Oppose linking of dates as these are, in my opinion, trivial links. This was my position in an earlier thread quoted above, though I may not have made it clear enough. I obviously disagree with those who think birth and death year links are trivial in biographical articles - it is my opinion that birth and death years are integral metadata information for biographical articles. Currently, such information is found either as: (a) plain text in the lead sentence, with some articles still having the dates linked; (b) birth and death date categories; (c) entries in the infobox; (d) entries in the Wikipedia:Persondata metadata information. Until the Manual of Style specifically mandates that the information for birth and death years needs to be in a form that can be analysed by computers (ie. metadata - and yes, linking is a form of metadata when used correctly), then delinking birth and death years without checking for the existence of the other metadata is a destructive process. I support reduction of overlinking, and avoiding a sea of blue links, but also support the retention of some form of clickable links to take the reader from biographical articles to our chronology categories and articles. Carcharoth (talk) 03:54, 30 September 2008 (UTC)
• I'd only support this as a reversion to the policy of all date linking, in other words linking dates of birth and death are no more or less valuable than any other date links. Either the standard should be to link all or to link none. - fchd (talk) 05:37, 30 September 2008 (UTC)
  • How so? Can you provide an example? I think the article with wikilink dates, eg. "He was promoted to Captain on 1 March xxxx ..." shows that THE date has only has relevance within the article itself, not to the world events at the time.
    At the moment, the issue with much of the discussion is the value judgments rather than relevance or usefulness. Many say it is of low-value where it means it is of low value to them. Whereas many of those supporting, say they find it useful, and they find it is of relevance for their research. I understand my biases, I would like the nay sayers to consider that it this is about relevance and perspective, not their values. --billinghurst (talk) 08:54, 30 September 2008 (UTC)
• Oppose Firstly, we already have consensus that wikilinking of dates is deprecated, so having this as part of the guideline would be a seriously retrograde step, and make a mockery of it, IMHO. Secondly, I would would be somewhat horrified at extensive wikilinking of birth and death dates: the vast majority of biographies I have come across have had these dates linked, and I just feel that these links add nothing to any of the articles. What I am talking about includes EIIR, where the only date I would probably retain is the date of coronation; I might also consider linking the dates of death of Mao Zedong and John F. Kennedy and other leaders who died in office, or other world figures who died at the height of their influence - for example John Lennon. However, we already have articles on the Coronation of the British monarch, Assassination of John F. Kennedy, and Death of John Lennon, which renders the linking unnecessary in the examples given, also proving Tony's point. I would say that even Albert Einstein's birth and death dates are but biographical facts which add little significance to the world if linked to date and year articles. If somebody really wants to look up 18 April 1955 for a context surrounding Einstein's death, they can just as easily type it in the search box or the address bar. It seems to be rather bureaucratic to oblige editors to add wikilinks to these whilst removing all the other wikilinked dates, when there is so much to do here on WP. Ohconfucius (talk) 10:29, 30 September 2008 (UTC)
  • Consensus was only reached after being repeatedly opposed. Tony simply kept resumbitting it until it reached consensus. I have been editing for a couple years on WP and I have never seen any change that has been so hotly contested as this. Your right though in that consensus was reached, now it is up to all of us to refine the details of the decision so that it best supports the project overall. I can live with the decision that dates should not be linked (although I don't agree with it per se) but I do think that certain key dates such as birth and death should be allowed.--Kumioko (talk) 18:27, 30 September 2008 (UTC)
• Support I use birth and death links all the time , as well as links in other key dates to get an historical context to what I am reading. Wikipedia year articles give a continuous timeline of what else was going on in the world at the time an event happend. They provide useful context and background and allow the reader to get immersed into a particular historic point in time. They are an invaluable resource unique to Wikipedia. Removal is a retrograde step. Lumos3 (talk) 12:43, 30 September 2008 (UTC)
• Weak support. Clearly some readers do find these useful, and the wide support for doing it can be seen in the fact that it has been so widely done. (If it had been introduced by bot, of course, this would not follow, but I see no sign that it has been.) We encourage multiple ways of linking articles together; categories and nav templates and links; this is merely another. I would much more firmly support weaker wording; but it is already established that normally means most people do, but you don't have to even for FA and GA, which should be weak enough. Septentrionalis PMAnderson 14:58, 30 September 2008 (UTC)
• Support, in the cases described by Carcharoth above. A less obvious way I have found links useful is to use them to see what is linked to a given article & the birth/death dates are one important way this works. Further, until this latest push to delink all dates, no one ever raised the issue that linking birth/death dates was unnecessary. I believe it deserves an exception -- & the spirit of ignore all rules more than justifies us to make an exception to any rule when the exception improves the encyclopedia. -- llywrch (talk) 18:50, 30 September 2008 (UTC)
• Qualified oppose. I don't have a strong opinion on the autoformatting question. Personally, I've always thought that our readers were smart enough to correctly read a date whether it was presented as 19 Jan 2008, Jan 19, 2008 or 2008 Jan 19. But I know that others disagree and I don't feel strongly enough about it to argue.
  On the more important question of whether the links are useful as links, I think they should pretty much all be removed. Linking a birth or death day to a page about that day of the month is invariably trivia. While many books publish such trivia, I do not consider that to be a proper function for an encyclopedia. There is nothing encyclopedic about the subject of the biography that the reader can learn by following the link to a page of other trivia that happened on all the other 19 Jans in time.
  The argument for linking years is better but still not strong enough in my opinion. The general argument for it (repeated by several people above) is that it provides historical context and can provide a path to the events which influenced the subject of the biography. I consider this a weak argument because the degree to which a newborn can be influenced by events outside his/her immediate family is trivially low. Child-development specialists will tell you that influences in the first 5-8 years are almost entirely domestic or, at best, highly local. The appropriate link for developmental context would be to the appropriate decate article covering the ages somewhere between 10 and 30. Likewise, a link to a death year tells almost nothing about the person's life except in the rare case where the death itself was a cause for notability.
  My opinion is also influenced by the observation that the "year" pages are massively overlinked. The odds of finding anything useful either on the page itself or by following "what links here" is miniscule. I've never yet followed one of those links and learned anything useful. Rossami (talk) 19:25, 30 September 2008 (UTC)
  • There are at least two uses for these links on birth/death dates. The first is an example of data management -- to maintain the Categories "X births" & "Y deaths". Not everyone who creates or improves an article remembers to include biography articles in these kinds of categories. The second is an example of user friendliness -- it helps end users to determine who was born or died on specific days. There are a lot of people out there who want to know who was born -- or died -- on a given day, & these links help them to research this information. While the Persondata information could offer the same information, so far Persondata is manually created & not yet present in all biographical articles. -- llywrch (talk) 03:13, 1 October 2008 (UTC)
• Support - It's clear that not everyone finds these links useful, but it's equally clear that some do find them useful to a degree, myself included. Jheald's proposal seems like a fair compromise. I'm confident that linking a date or two in the lead won't turn the rest of the article into an indecipherable sea of blue. --Bongwarrior (talk) 19:38, 30 September 2008 (UTC)
  • Bongwarrior: OK, “some” find the [year] links useful. Is that the test you think should be used here: (“some”)? Or do you think it is more than just some, and that the body of readers who would actually want to read through lists of trivia in “year” article are sufficiently numerous to merit yet more blue links in our articles? Greg L (talk) 22:14, 30 September 2008 (UTC)
    • My impression from the above is that those who find the year links useful are coming more from the metadata side of things, rather than the trivia side of things. It would also be nice to have some acknowledgement that birth and death years are less trivial (though some people do clearly see them as still trivial) than a random mention of a year in a random article. And also that linking birth and death years does not contribute to a "sea of blue (links)", but is actually limited to a specific place (at the start of the article) and to two specific links. To expand on the metadata side of things, I'd be happy if a sustained effort were made to bring biographical articles into compliance with some standard style, ensuring that all the articles had Wikipedia:Persondata (currently woefully limited in its application - to respond to Kaldari's point below), that all biographical articles had birth and death year categories (or the 'unknown' equivalents) and the "biography of living people" tag (where applicable) and that all biographical articles had {{DEFAULTSORT}} correctly applied (to aid the generation of a master-index, as well as categorisation). If half as much effort went into that as into whether to link birth of death dates or not, then some progress might be being made. As it is, biographical articles account for around 1 in 5 of Wikipedia's articles (and, I suspect, a significant fraction of newly created articles), but only a small fraction use Persondata, thousands and thousands of biographical articles are not sorted correctly in the index categories, and many lack birth and death year categories. Many biographical articles also lack the {{WPBiography}} tag on their talk pages. This is one reason why I feel as strongly as I do about not just removing birth and death year links until a proper audit of the biographical articles has been carried out (you can, if you like, think of it as the "date audit" clashing with plans for a similar "biographical audit" and the "date audit" removing metadata links that might have been parsed by the "biographical audit"). To take that one step further, I wonder if the contributions log of Lightbot can be analysed to reveal how many birth and death years were delinked on biographical articles where no birth and death year categories were present? I presume such an analysis would be possible? Carcharoth (talk) 23:27, 30 September 2008 (UTC) I asked Lightmouse here if he can help.
      • I suspect that the actual readers click on links much less than we think they do. There's no evidence for their popularity. The concept of wikilinking is great, but needs to be rationed carefully. No studies have been conducted on readers' attitudes or behaviour in relation to them (for example whether readers tend to read through as much of an article as they're ever going to and then consider hitting a link, or whether they divert on the spot), but common sense tells me that the utility is fragile. Tony (talk) 02:30, 1 October 2008 (UTC)
        • You are probably right. I would say it would depend on a combination of factors: (1) Whether the reader understands the term or knows about the object/event/person linked {information/definition); (2) Whether the article contains sufficient context to explain things and avoid the need for a reader to click away to another article (insufficient article context); (3) Whether the reader is bored by the article they are reading and whether any particular link looks more interesting (diversionary browsing); (4) Whether the reader (after reading the whole article) wants to read up further on a particular topic (discretionary browsing). It depends on the reader to a large extent. What we, as editors, can do, is ensure articles have sufficient context to reduce the need to link, keep articles interesting, keep metadata separate from linking, and try to ensure high-quality linking (linking to good articles and to the correct articles) and to avoid overlinking. If there was ever a push for levels of linking, then one good metric would be "if a fact in article A is mentioned in article B and vice-versa, then that is a primary link", with other links being "background" or "definition" links. Trouble is, there is such a spectrum of reasons for linking, that levels of linking just allows for edit warring. If some software thing like "there is a reciprocal link" could be enabled to turn a link a different colour, that might work, but then too many different colours makes things silly as well. Maybe a preference to only have reciprocal links display? Carcharoth (talk) 03:00, 1 October 2008 (UTC)
          • Wow. Tony, are you saying we should unlink everytihng, not just dates or a few countries, but everything? Links aren't popular? We need to ration them? This certainly explains some of your underlying motivations. dm (talk) 03:47, 1 October 2008 (UTC)
            • It’s not too complex Dmadeo. Links should be judiciously used. They should be highly topical and germane to the subject matter. They should invite exploration and learning for the intended audience. Linking to electron is perfectly fine for the Atom article but would be boring and desensitizing to readers reading up on Planck units; the majority of the visitors reading that article already know what an electron is. The litmus test shouldn’t be whether or not some readers will find it interesting, but whether a good number of the target readership would find it interesting enough to click on. For too long, too many links have been added to Wikipedia’s articles because an article existed and could be linked to. But with 6,909,101 articles on en.Wikipedia, hundreds of them nothing but date-related trivia, plus even more on Wiktionary, the number of articles to link to is now astronomical and our articles have become excessively linked, effectively turning them into giant, boring, blue turds. Tony is right. We don’t need links to mind-numbing list of randomly-generated trivia nor to common countries. Nor to Manhole cover in the street out in front of Greg L’s house (it’s at a latitude of 47° 39′ 9.1″ for those who would actually be interested in that). It’s not that nobody is interested in clicking on all these links; it’s just that not enough readers are interested in clicking on them. IMO, the reaction to often strive for in readers when we provide links should be “Oh, WOW. I didn’t know they’d have an article on that too!”. Greg L (talk) 04:31, 1 October 2008 (UTC)
              • Precisely. As an aside, you should subst that 'number of articles' template, otherwise in a year's time it will show the number of articles at the time someone reads the archives, not when you wrote this - what do you mean, "no-one reads the archives"? :-) Though there could be a useful distinction, I think, between levels of information on an article and what to link to. Not everyone reading the Planck units article will know what an electron is - that is why you could link it once at the first appearance, and then not link it again (which is normal practice anyway). Consider the reader who wants to click "electron" but can't. They will either edit the article and add a link, or they will look "electron" up by searching for it. But they will be thinking as they do so "why didn't they give me a link to click on?!". But even relevant links are uninteresting to some. The first link on Planck units is units of measurement. I have no interest in clicking on that, but because it is relevant, it stays. So relevance is probably more important than whether a link is interesting. As for links to common countries, there are exceptions to every rule. If you have a list of countries, sometimes it makes sense to link all of them, rather than just some of them. Your "oh wow" point is one viewpoint (and something I agree with). The other is the semantic web - see WP:BUILD. Going too far one way or the other (overlinking and underlinking) could be very damaging. How would you propose to avoid underlinking? Carcharoth (talk) 05:02, 1 October 2008 (UTC)
                • I get the overlink problem, and it seems like a theoretical problem, but not really one in practice. I think it's a lot better to deal with a particular problem article with a simple MOS guideline and involved editors actually editing the articles. Trying to prescribe exactly how to do this in the MOS devolves into lists of what's acceptable and what's not (ie: unlink the United States, but not Australia). I've seen others describe this as overinstruction or instruction creep and I'm starting to feel that there's a small number of vocal people who really like the idea. I find it offputting. dm (talk) 08:19, 1 October 2008 (UTC)
          • This is exactly what I was afraid would happen, in Tony's opinion above wikipedia should be nothing more than a publically updated encyclopedia britanica with a few links sprinkled in the article for certain key events. Tony, THIS IS NOT A 2 DIMENSIONAL DATABASE, stop trying to force your narrow views on everyone else. I agree that many articles are overlinked and I understand what you are saying, but having the links is useful and they generate trafic to other articles perpetuating the cycle of publically updated information. If we start stripping off links then one of the primary selling points of wikipedia is lost and we might as well buy the paper set when the salesman comes to the door.--Kumioko (talk) 15:12, 1 October 2008 (UTC)
• I have to respond when my views are being misrepresented. I'm sure people aren't deliberately making things up, so I wish they'd check their facts first. (1) I see little value to the readers, and much unnecessary blue in prominent positions, in the linking of common country names, especially English-speaking countries. Just why every single popular culture article should have a link to "British", "UK", "American", "United States", "Australian", "Australia"—I've counted seven to one country in a single article—is quite beyond me. This includes such little-known entities as "India", "China", "Russia", and some European countries. If it's a world map our readers require, they should be made well aware of its existence on the main page, since these country articles swamp the linking reader with huge amounts of information, most of it unrelated to an article topic. (2) It's easy to accuse me, in an exaggerated and frankly quite unfair way, of wanting to strip away all or most links; but in reality, I'm pro-wikilink; I believe people who complain about the notion of a more selective approach to linking are, without their realising it, working against the wikilinking system by diluting the valuable links to such an extent that they are ignored by most readers. It's a great way to kill of a great system. I'm trying to make it more effective. Tony (talk) 15:40, 1 October 2008 (UTC)
  • …there! Another link to mindless trivia. Why? I link, therefore I am. Greg L (talk) 20:47, 1 October 2008 (UTC)
  • P.S. I agree completely when Tony wrote “I believe people who complain about the notion of a more selective approach to linking are, without their realising it, working against the wikilinking system by diluting the valuable links to such an extent that they are ignored by most readers.”  Well said. Greg L (talk) 23:16, 1 October 2008 (UTC)
    • It may surprise you, but I agree with what Tony said as well. We just draw the line at different points. People will always have different ideas about what to link and what not to link. If you want to successfully persuade more people to reduce overlinking, it might be worth expanding WP:CONTEXT to explain things in more detail. I also think part of the problem is that editors often think "do we have an article on this?", and then try a wikilink to find out (using preview). When it turns out to be blue, they check it (hopefully) and then leave the link there because they are pleased that we have an article on whatever. The pleasure at seeing a wikilink work is such that it can be very hard to consciously remove it. By the way, thanks for the essay (I'm sure I've seen a similar essay somewhere before). It makes some interesting points, even if I think putting vomit in the "see also" section is a bit over the top and faintly insulting, as is linking to insanity, but it's your essay. I would add some footnotes to the essay, giving examples of "fascinating" trivia from the October 16 article (I didn't read all of it, but I did skim it), but that might not be appreciated. Seriously, have you ever thought of putting articles like October 16 up for deletion? You sound like you would be happy if they were all deleted. Finally, thanks for the photo of a sewer manhole cover. I've placed this photo in the sanitary sewer article - might as well use the picture to improve an article as well (did you know some people actually collect pictures of manhole covers? See here. There is also some interesting history behind some manhole covers. But then if you are recoiling in horror at the thought of this, then I guess you wouldn't appreciate things like Station Jim either. Carcharoth (talk) 23:56, 1 October 2008 (UTC)
      • Carcharoth: When you write “Seriously, have you ever thought of putting articles like October 16 up for deletion? You sound like you would be happy if they were all deleted.”. Perhaps I might come across that way but, no, I wouldn’t want them deleted. Just de-link them.

        There are just too few people who are reading up on, for instance, Hugh Beaumont (actor), who are really going to read more than the first two entries after they click on a date link. I’d bet that 99.9% of the time, the typical reaction is “Hmmm… that’s what these links do” and then they click their browser’s ‘back’ button. Even with my challenge in the essay, it will be interesting if anyone can ante up and actually read only two of those trivia articles.

        By better anticipating what readers to a given article will be interested in further exploring, we increase the value of the remaining links. If someone is in a mood for long lists of historical trivia, it’s easy enough to type them into the search field.

        And I agree 110% with you when you write about the litmus test many editors use in deciding whether to link or not: if it can be linked to, then link to it. Greg L (talk) 00:35, 2 October 2008 (UTC)

        • Thanks for the reply. I must admit that when I wikilink an article without wikilinks (sometimes a badly written one - the lesson there is that it is better to rewrite the article before wikilinking), I have tended to add links to find out if we have articles on certain things, and only then winnowed the links down to those that are most relevant (and sometimes not even that). I will, in future, be trying consciously to increase the quality and 'impact factor' of any wikilinking I do. I still think that wikilinking tries to do too much - acting as (among other things): a dictionary/glossary; a 'related topics' section; and a further reading section. Carcharoth (talk) 02:08, 2 October 2008 (UTC)
• Support, I think the above comments provide a variety of compelling reasons why editors might want to link dates. What I would actually prefer is for editors to be given explicit discretion in whether to link these dates on any given article. Within the context of the rest of the MOS I think the proposed language is closer to that ideal than the existing text. Christopher Parham (talk) 02:01, 2 October 2008 (UTC)
  • I have five articles in mind already for an insertion of a link to Greg's essay on the sewer cover outside his house. Seriously. Link as much as you can, wherever there's a tiny opening to do so; after all, in today's world, everything can be related to everything else by one, two or three steps. it won't hurt the valuable links.Tony (talk) 02:27, 2 October 2008 (UTC)
    • The October 16 article has nigh on 12,000 incoming links. The same article has some 260 lines/events listed. The 364 links to other date articles created by {{months}} hardly dents the total. There is a serious imbalance here. 'October 16' is only one of 366 such articles with a very similar problematic. I am not saying that all articles should be back-linked from the date page, or that the majority are related to biographical d-o-b or d-o-d, but I would contend it is one valid perspective on the rather pandemic overlinking to date articles. Ohconfucius (talk) 04:22, 2 October 2008 (UTC)
      • If one looks only at links from article space and further exlude the 1199 lists, 1403 titles of the sort "2008 in medicine", 366 days, and 12 months, the count drops to 7425. Still high, but less outrageous. Looking closer at, say, XACML we see it is only linked by the date on a cited reference. I see no reason for linking citation data that is already well-structured, as in this date= field of a cite tag. On the wild assumption that only 2/3 of those are date= or accessdate= instances, that gets the number into a reasonable range.LeadSongDog (talk) 20:21, 3 October 2008 (UTC)
        • I suspect there may be some truth in that assertion, but unless and until all those citation templates are de-linked, we have no way of knowing. Ohconfucius (talk) 08:59, 9 October 2008 (UTC)
• Oppose. For the same reasons as other opposition. Lightmouse (talk) 13:36, 2 October 2008 (UTC)
• Come on… At least a couple of you “Support” editors ought to be taking me up on my challenge. If you can actually read four whole date and year articles, you can be the first recipient of your very own Sewer Cover Barnstar. Are there no takers? Greg L (talk) 03:03, 3 October 2008 (UTC)
  • I looked at your four date articles. I'm sure it's not going to convince you, but they didnt seem that bad. Someone had gone through and organized them enough to make them interesting. They arent going to be everyone's cup of tea, but I'm not sure why you're so offended by them either. I suppose suggesting you just don't look at them won't help either. dm (talk) 23:43, 3 October 2008 (UTC)
    • Dmadeo, when we, as editors, are deciding on whether or not to link a word or topic in an article we are writing, I would suggest setting the bar a bit higher than, “that didn’t seem so bad.” I might even be so bold as to suggest that we set the bar a bit higher so that in many cases, the reader’s reaction to seeing a blue link would be “Way cool… I didn’t expect they’d have an article on that too!” Greg L (talk) 00:14, 4 October 2008 (UTC)
      • With respect, I'd suggest setting the bar at whatever level makes you feel like contributing to articles. That level will be different for me and for anyone else, but thats fine. I encourage you to link however many words you'd like, as long as you dont mind when I do as well. dm (talk) 01:31, 4 October 2008 (UTC)
• Oppose. As illogical, fussy and confusing as when we decided after prolonged discussion not to autoformat, just a short while ago. 86.44.28.60 (talk) 18:36, 3 October 2008 (UTC)
• Oppose as confusing, since the policy is now *not* to link dates without particularly compelling reasons. "saving some curious readers the trouble of typing a year/date into the 'seach' gizmo" just doesn't seem sufficiently compelling. Sssoul (talk) 17:22, 5 October 2008 (UTC)
• Support linking years at least once. It is a powerful way to update and expand the year pages to use the 'what links here' button and see what pages refer to a particular year. Jcwf (talk) 00:23, 6 October 2008 (UTC)
• I don't know if this has been mentioned above, but there is a very relevant CFD discussion at Wikipedia:Categories_for_discussion/Log/2008_September_30#Category:Deaths_by_age. Some people, who seem to be in the majority, want to create a series of categories, automatically generated, of Category:Deaths at age 28, Category:Deaths at age 29, and so on. Whether they need the links being discussed here I don't know. Johnbod (talk) 00:59, 6 October 2008 (UTC)
  • I don't see the relevance. — Arthur Rubin (talk) 03:35, 6 October 2008 (UTC)
• Oppose Not needed, per very many above. Johnbod (talk) 00:59, 6 October 2008 (UTC)
• Support (changed from oppose): I've changed my stance here because, while I frankly still can't see how linking of dates is useful, it is clear to me that there is a significant minority of editors who do find it useful. If it's useful enough for even a few editors, then it is something which we should be linking.--Aervanath lives in the Orphanage 04:26, 7 October 2008 (UTC)

• Comment strange recursive reason for changing an opinion. You now support because you have seen a "significant minority" of other editors support? Do your previous convictions not amount to anything? Ohconfucius (talk) 11:18, 25 October 2008 (UTC)

• Support the proposal. I find it useful. Deb (talk) 13:55, 8 October 2008 (UTC)
• Oppose I have been very pleased to see a plethora recently of edit summaries reading "date audit per MOS:NUM" and I strongly feel that no wikilinks should be used for ornamental purposes which is what these links are. The next step would be linking full stops. (<- wikilinked) People are used to these links but they should go. People will get used to not having them, and if in one instance out of approximately 163 times reading the number 2008 they actually wanted to check out that page, they will not be annoyed at having to type it into the search box. __meco (talk) 17:32, 12 October 2008 (UTC)

• Oppose I think it's a horrible idea and will encourage even more pointless wikilinking. I agree with Meco and most of the others who oppose this proposal (sorry, no new reason I can think of for opposing, it's more or less all been said.). Doug Weller (talk) 09:55, 25 October 2008 (UTC)

• Oppose any naked linking of dates and years serves no purpose. So having it at the start of articles is a particular bad idea. --HJensen, talk 21:34, 26 October 2008 (UTC)

• Oppose linking of birth and death dates is trivia and anyone interested can easily type in the dates. Dates are incredibley overlinked as it is. RainbowOfLight ^Talk 09:07, 27 October 2008 (UTC)

Infobox templates

• BTW: up to this point it's 14 support and 13 oppose (if I counted right and ignoring any weak/partial distinctions). Sounds to me like there's no consensus either for or against this particular point. But it does point out that there is a large contigent of people who do want limited date linking, especially for something such as birthdates. As far as I know, lightbot is not unlinking the {{Birth date|yyyy|mm|dd}} and {{Death date and age|yyyy|mm|dd|yyyy|mm|dd}} templates, so perhaps we can say "In biographical articles, limited use of {{Birth date|yyyy|mm|dd}} and {{Death date and age|yyyy|mm|dd|yyyy|mm|dd}} may be helpful" dm (talk) 08:27, 1 October 2008 (UTC)
  • Strong comment: What this tells me most clearly of all is that we have lots of !votes from incoming parties totally unaware of the rest of debate (over three years worth) and thus largely-to-totally unaware of the negative aspects of date autoformatting. As just one example among many, I doubt that more than a handful of them have considered the fact that around 40% of surveyed articles had inconsistent date formats in them. This is largely because editors assume that the autoformatting just "handles it", and forget that 99.99% of Wikipedia's users are IP address readers, not editors, with no date preferences to set, who are all seeing "3 July 1982" in one sentence and "August 7, 1983" in the next – all because autoformatting ensures that most editors themselves simply don't notice the difference. This is happening in nearly half of our articles. That alone is enough to end this debate right now. — SMcCandlish [talk] [cont] ‹(-¿-)› 05:00, 4 October 2008 (UTC)
  • Do these templates render the dates in bright blue and have all of the disadvantages of the date autformatting system? Tony (talk) 08:33, 1 October 2008 (UTC)
    • Don't you mean the *advantagees* of date autoformatting? - fchd (talk) 08:40, 1 October 2008 (UTC)
      • Um, no, he meant disadvantages. — SMcCandlish [talk] [cont] ‹(-¿-)› 05:00, 4 October 2008 (UTC)
  • One or other of the two "birth date" templates MUST be used in infoboxes, if the birth-date is to be included in the emitted hCard microformat. Whether or not they link those dates does not affect this; and can be set according to whatever is the final community consensus. One or other of the two "death date" templates will be needed, when the hCard spec is updated to include "death date".Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 10:15, 1 October 2008 (UTC)
    • I don't have the expertise to understand this. What I can tell you is that it's great that many of the infobox templates have recently been modified so they don't augoformat the dates. Tony (talk) 10:48, 1 October 2008 (UTC)
      • In short: the templates are needed for technical purposes (related to metadata). It doesn't matter (for those purposes) whether they link the dates, or not. But people shouldn't be discouraged from using them, because of formatting, as not doing so will break one of the functions of the infoboxes in which they're used. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 11:12, 1 October 2008 (UTC)
        • For the record, these templates are currently not emitting links (since 1 September). Jheald (talk) 11:29, 1 October 2008 (UTC)
          • The birth and death date template age calculation may be wrong for a person who was born under the Julian calendar and died under the Gregorian calendar. They also provide no way to indicate what calendar was used for the dates. --Gerry Ashton (talk) 14:42, 1 October 2008 (UTC)
            • Those are valid concerns (and are being discussed elsewhere, I believe) but are unconnected to the issue of linking; also, such cases seem to be vastly in the minority. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 15:50, 1 October 2008 (UTC)
              • Not being a template programmer, I don't know if the concern can be fixed. I am reluctant to recommend a template that cannot fulfil its intended purpose, and might not be repairable. --Gerry Ashton (talk) 16:07, 1 October 2008 (UTC)
                • Forgive me for jumping in here, but can you explain something in simple terms to me? What is the purpose of the metadata, and the parsing thereof? I have seen countless mentions on this talk page that if dates were linked, such as birth and death, the collection of metadata would be made easier (am I right here - even if this can be achieved through plain text). This maybe the case, and several editors above wish it to be so, but I don't understand why. Maybe this issue isn't relavent here, but could somebody humour me. Dates should/would/could/may (whatever) be linked to allow for the easy collection of metadata. But why? (I'm not criticising metadata, or those who use it - I just don't understand it's purpose.) In anycase, for birth/death dates, is that not what {{persondata}} is for?–MDCollins (talk) 21:47, 1 October 2008 (UTC)
                  • It should be possible to generate a list of every biographical article on Wikipedia, along with the biographical data (where known). To do that, you generally need mature and comprehensive metadata coverage. Unfortunately, the maintenance of metadata on Wikipedia (en-Wikipedia at any rate) lags severely behind the rate of article creation (persondata, as you say, is one of the places where metadata should be placed, but as there are other places as well, such as the hcard format Andy mentioned above, and since persondata is used in only a small fraction of articles, there are problems). Wikilinks are sometimes analysed as a form of metadata, and certainly a mature and well-developed system of date markup would allow for applications. Geographical co-ordinates are given in a standard way - maybe dates should be as well. It is possible to go too far with this, though, since Wikipedia is primarily an encyclopedia, not a database (yes, I know the underlying software uses database tables, but I'm talking about the content here). It's a question of getting the balance right. I'm perfectly happy for dates and years to be mostly delinked (with a few exceptions), but the metadata concerns also need to be addressed. Carcharoth (talk) 23:34, 1 October 2008 (UTC)
    • No one was talking about removing those templates, anyway. — SMcCandlish [talk] [cont] ‹(-¿-)› 05:00, 4 October 2008 (UTC)

a compromise/interim proposal?

this whole date-link thing is like a Passover situation: since the linking of dates for autoformatting is now discouraged, it's pretty clear that bots and scripts that can assist with the huge task of unlinking that kind of date-link are really needed (otherwise it'll take years, and meanwhile editors who imitate what they see will keep adding them). the question is how to designate certain linked dates as *intentional* so that the bots, scripts and/or editors unlinking dates manually will leave those alone. one proposal (discussed here) was to designate such "intentional" date-links by putting them in the "see also" section with nondate words in them to let the bots and scripts know they shouldn't unlink them, for example:

[[1943|Other notable events of 1943]]

[[18 December|Notable events on 18 December throughout history]]

[[1978 in music]]

some editors feel that solution is "too much trouble", but it's nowhere near as much trouble as leaving the massive job of unlinking now-deprecated links to be done by hand just because a bot might undo a link someone cherishes - and even if people are doing the unlinking job manually, they still won't know just by looking at them that this link here is cherished by someone, but that link there is free to go.

the editors who want birth/death dates linked at the start of biographies are proposing that that positioning should designate those dates as "cherished/untouchable". the trouble is that bots can't be taught to recognize position in an article; for a bot to understand that it should leave a link untouched, the link needs to include a non-date word. it's not easy to think of a non-date word to insert in birth/death-date links without making the sentences awkward - especially considering that the formatting breaks whole dates up into two parts: [[calendar date]] [[year]].

but: would the people who want to keep birth/death dates linked in the first lines of biographies be satisfied if the years remained linked - at least until we think up a better solution - and the calendar dates were unlinked? i understand that some editors feel the birth/death year links are potentially of interest for the context they might provide, but the calendar-date links don't provide context - they provide what amounts to historical trivia, which (i posit) *is* very adequately relegated to the "see also" section where the few readers who want to know what mishmash of events went on on that date throughout history can easily find it, marked explicity as "[[18 December|Other events on 18 December throughout history]].

if that would satisfy the people who cherish these birth/death-date links in bios, then the gallant bots that are waiting to assist with the necessary task of unlinking meaningless/now-deprecated date links could at least proceed with the unlinking of calendar dates. in good Passover fashion they would leave alone any calendar date link that includes a non-date word, like "[[18 December|Other events on 18 December throughout history]]. so if some calendar dates that someone cherishes do get unlinked, that can be repaired by adding a non-date word or phrase to the link to shield it from the next "pass" of the bot and moving it to the "see also" section so that sentence flow is not encumbered.

and meanwhile we could think some more about how to designate linked years for "Passover purposes".

i hope someone sees what tree i'm trying to bark up here. Sssoul (talk) 12:52, 13 October 2008 (UTC)

The major problem with this idea is that users should never have to alter their behavior to accommodate the bots; in fact the situation is always the reverse, in that bot behavior should be modified to accommodate the editors. If a bot cannot be written in a way that does not require human editors to modify the way they are editing to make the bot's task possible then the bot should not operate. This also raises the issue of existing articles - requiring users to go back and "mark" existing dates in order to prevent the bot from removing links is placing far too large a burden on the editors. In short : bots are created to make editors' tasks easier, not to force them to do more work. As such I find this proposed compromise - while a valiant attempt at finding a middle grounds - to be unacceptable. The central question that must be answered (and has not yet been significantly explored) is whether or not a retroactive removal of now deprecated links is acceptable to the community. Shereth 17:25, 13 October 2008 (UTC)

thanks first off for recognizing that i'm trying to be constructive. maybe i haven't made idea clear enough, though: i don't mean that any editors would *have* to mark any dates in any special way; everyone would be free just to leave the calendar-date links as they are and let them be unlinked. perhaps that wouldn't be a problem for anyone, since the argument that some date links provide historical context doesn't apply to calendar-date links. it seems plain that the vast majority of calendar-date links were created purely for autoformatting, and since linking for autoformatting is now discouraged/depracated, i don't understand (at all) what the resistance to removing those links is based on.

but meanwhile in case someone really feels it's valuable, encyclopedic, etc, to offer readers a link to [[18 December|Other notable events of 18 December throughout history]], they *could* create something like that if they wanted to, knowing it wouldn't get unlinked by a bot. (maybe by another human, but that's a different question!)

also, policy/guideline changes very often *do* entail people changing the ways we edit. in this case the policy deprecating links for autoformatting is the source of the changes. the bot is just meant to assist with making them. Sssoul (talk) 17:59, 13 October 2008 (UTC)

If I might just co-opt something you said, which I believe to be the very root of the problem : "The policy deprecating links for autoformatting is the source of the change." This whole mess gets fixed when one very simple question is answered, that question being, "Does the deprecation of certain types of year/date links necessitate the subsequent (and retroactive) removal thereof?" If the answer is yes, then Lightbot should resume, and the above idea can be implemented to allow editors to retain some links with some changes. If the answer is no, then the bot should not be resumed and the fate of now deprecated links left to the editors of the articles in which they are found. That question has yet to receive any community-consensus based answer, and I believe answering it would fundamentally solve the problem. Shereth 18:27, 13 October 2008 (UTC)

... i'm really struggling to fathom how the policy deprecating linking for autoformatting could possibly be interpreted as implying "but let's keep all the existing date links that have no purpose except for autoformatting" - but if that indeed needs clarification, how do you suggest seeking clarification - another RfC? mediation? or ...? Sssoul (talk) 19:07, 13 October 2008 (UTC)

ps: having slept on this ... Shereth, when you wrote "the major problem with this idea is ...", it seems to me that what you actually meant is "my [ie your!] main objection to this idea is ..." your view is as important as anyone else's, of course, but it seems to me that it would be fair to hear some other people's reactions to the idea before labelling one aspect of it a "major problem". Sssoul (talk) 08:53, 14 October 2008 (UTC)

• There is a fundamental flaw with Shereth's argument and objection. Humans are endowed with intelligence, while bots are not. In order for things to be automated/automatable, things have to be done according to a certain logic. Economists realised that long time ago, and Taylor invented the concept of the production line; the vacuum cleaner was invented so we no longer clean floors the same way as before, the same applies to almost every labour-saving device you can think of. To say that humans should go about and be humans in exactly the same way as before is, with all due respects, bollocks. Ohconfucius (talk) 09:23, 14 October 2008 (UTC)

• Re Ohconfucius: You have completely missed my point. See the response by Lightmouse below which catches my sentiment exactly. Software should simplify the tasks undertaken by humans and not complicate it. I do not mean to imply that we shouldn't adapt the way we do things to accommodate new technological advances. If that's what you got out of my statement then you completely misunderstood what I was saying.
• Re Sssoul: Fundamentally I'm not opposed to removing date links. Fundamentally I really don't give a crap one way or the other. What's got me so animated here is that as an administrator I see (and deal with) complaints from sundry editors who have got their knickers in a bunch because of Lightbot removing some links and they do not see any consensus to remove links. I'm somewhat tired of having to deal with/respond to the situation. The reason I am so adamant about getting consensus is so that next time I see someone pitch a fit on a noticeboard about Lightbot removing date links, I can just point to the consensus and be done with it, rather than having this debate re-ignited for the umpteenth time. Shereth 13:51, 14 October 2008 (UTC)

thank you Shereth - so how do you propose seeking the consensus you consider necessary to clarify whether the deprecation of date-linking for autoformatting includes the premise that existing date-links that serve no purpose anymore should be undone? Sssoul (talk) 15:35, 14 October 2008 (UTC)

An RFC, coupled with a clear and unambiguous statement (such as "Does the deprecation of bare year links mean they should be systematically removed via bot") would establish a sufficiently strong consensus on the matter. Shereth 15:41, 14 October 2008 (UTC)

well ... why just "bare year links"? the now-depracated autoformatting also involved calendar-date links. maybe both questions had better be asked at once, to avoid someone saying yet *another* RfC is needed. Sssoul (talk) 15:59, 14 October 2008 (UTC)

The wording was just an example suggestion, naturally it would make sense to cover all of the bases. It might also make sense to keep the individual points as discrete questions to keep folks from getting confused about what they are or are not supporting, but it definitely would make sense to cover all of the issues in a single RFC and get it done with. Shereth 16:05, 14 October 2008 (UTC)

thanks for clarifying. would it be worth making the "by bot" question a distinct point from the other two? as in (and i hasten to add that i don't mean this is exactly how they should be worded):

• 1] does the depracation of autoformatting mean existing calendar-date links that served no purpose but autoformatting need to be systematically undone?
• 2] does the depracation of bare-year links mean the ones that exist need to be systematically undone?
• 3] if yes to either: is it desirable to enable a bot to do the systematic unlinking, or should it be done only manually? (i feel like it would be fair to point out right away that there are ways to "earmark" both kinds of date-link so that a bot would not undo them, but that that part *would* need to be done manually - a link to a brief description of the suggested ways to "earmark" the links could be included.) Sssoul (talk) 17:04, 14 October 2008 (UTC)

I like this approach. Shereth 17:27, 14 October 2008 (UTC)

You make a fair point, Shereth. Software should eliminate, reduce or simplify human tasks. If only such reasoning had prevailed when auto date formatting was proposed ("its easy, all you have to do is link some but not all dates and in this exact way"). Now we have to find a way of clearing up mess attributed to autoformatting. We can do it with or without software assistance. Lightmouse (talk) 17:35, 13 October 2008 (UTC)

I won't disagree with you. I just have to repeat my insistence that community-wide consensus regarding the fate of said deprecated links be cemented prior to taking any sitewide actions. Shereth 18:27, 13 October 2008 (UTC)

Sssoul proposed bot delinking of calendar dates such as [[18 December]] and described some other features of his proposal. You are making yourself very clear that you think it requires the expressed opinion of many people. That is clear. You are one of the many people and your opinion is valid. Would you personally accept Sssoul's proposal? Lightmouse (talk) 18:51, 13 October 2008 (UTC)

I do not oppose the de-linking of calendar dates, no. This does not mean I personally accept Sssoul's proposal in its entirety Personally I do not see the need for such a compromise if only a demonstrable consensus could be reached one way or another. For all I care you can turn Lightbot loose on calendar dates, as I believe the primary objection that keeps coming up is regarding years, not the calendar dates. Shereth 14:01, 14 October 2008 (UTC)

I do object. Linking calendar dates should be even rarer than linking years, but there are occasions where it adds value. The link from Pope Sylvester I to December 31, St. Sylvester's Day, should stay. Septentrionalis PMAnderson 14:15, 14 October 2008 (UTC)

Pmanderson, do i understand that you see no possibility of designating that particular calendar date as valuable/meaningful in that article by inserting a nondate word in the link - for example [[31 December|Notable events on 31 December throughout history]]? Sssoul (talk) 15:35, 14 October 2008 (UTC)

Thanks. That's clear, as usual. I agree that the ideal solution would not be a compromise. I also agree that the heat of the debate is about years. I hope we can both agree that it has been about 'solitary years' (blah blah [[1974]] blah blah) rather than the year that was linked to enable autoformatted of a full date (blah blah [[18 December]] [[1974]] blah blah). Thus I would summarise Sssoul's proposal to turn into a bot specification such as:

• leave solitary years alone and delink all other date components/compounds unless they contain a non-date term.

Many articles must be exempted as a whole e.g. date related articles themselves. Lightmouse (talk) 14:51, 14 October 2008 (UTC)

for the record, i personally would also prefer a non-compromise solution - i'm just seeking to help editors who are saying they want to keep particular dates linked as exceptions to the current policy deprecating date-linking. what i'm trying to do is point out how they can "protect" the links they want to keep, if they want to. and i focussed this latest suggestion on calendar-date links because i think there's less distress about/resistance to unlinking those. and i still hope a satisfactory resolution can be found for dealing with year links. Sssoul (talk) 15:35, 14 October 2008 (UTC)

It's already "marked as valuable/meaningful"; it appears three times in the infobox: once as his death, once as the end of his papacy, and once as his feast; also in the sentence: In the West, the liturgical feast of Saint Sylvester is on 31 December, the day of his burial in the Catacomb of Priscilla. At least the last two should be linked; they should not be convoluted to satisfy some piece of MOScruft. Both mean, and should say and link to, 31 December. When MOS ordains bad writing, as it does all too often, it is malfunctioning. Septentrionalis PMAnderson 17:27, 14 October 2008 (UTC)

thanks Pmanderson - i understand what you're saying. i just want to note in passing that no one is suggesting eliminating calendar dates from any articles - only asking whether all how many of them need to be linked to lists of miscellaneous events that happened on the same date throughout history. i understand that you feel this one *does* need to remain linked to such a list for the article to be understood, but that stating explicity in the link what it's a link to would not be okay with you in your opinion be detrimental to the article. Sssoul (talk) 17:40, 14 October 2008 (UTC)

I don't support linking all of them - and have said so at least twice; the effect of this proposal, however, is to link none of them, and substitute (for some of them) see also links which many readers will not know exist. Septentrionalis PMAnderson 18:05, 14 October 2008 (UTC)

okay, i've amended my statement above. i don't understand why "many readers will not know" that there are links in the "see also" section, but please note that the bot will leave intact any date-link that has a non-date word in it, wherever it appears; such links could appear in the body of the text, for example:

• For fuller understanding, see the list of miscellaneous events on 31 December throughout history.

i feel that making date-links that you consider important explicit that way (and eliminating the masses of meaningless date-links) will have the advantage of clarifying for readers that your link really *is* worth following. Sssoul (talk) 06:40, 15 October 2008 (UTC)

• I would still tend to believe that, even in the case of Sylvester, '31 December' is a low value link. It is infinitely more informative to link to the appropriate Name day article, because that's where the true meaning is. Ohconfucius (talk) 07:57, 15 October 2008 (UTC)

i agree, Ohconfucius, but unless i'm misunderstanding Pmanderson, he/she takes a different view. i'm just pointing out that the proposal *does* accommodate date links that an editor truly feels are essential to understanding an article. Sssoul (talk) 08:09, 15 October 2008 (UTC)

• Pmanderson's just pointing out the exceptions which should prove the rule. There will always be exceptions, and we just need to agree on a way of treating them. His opposition to delinking by bot appears to be a bit Luddite to me. Is there an agenda? Ohconfucius (talk) 08:39, 15 October 2008 (UTC)
  • Then I suggest Ohconfucious look up Luddism: opposition to machinery on other grounds than whether it will work better. (Sometimes those grounds are also valid; harmony between editors is a good, and one which bot reversions tend to corrode.) But in this case, the bot will work worse: bots should not be used where there are exceptions, because they will not notice them, and (if reversed) they will come back and edit war for them. (He might also look up exception proves the rule; that saying has two senses: the legal one is irrelevant here, and the other is destructive testing. Enough exceptions blow up the rule.) Septentrionalis PMAnderson 17:42, 15 October 2008 (UTC)
    • Incorrect. The original legal definition is the exact relevant one. If the rule is to remove all square brackets which surround 'mmdd', 'ddmm' or yy, then it is evident that all others are deliberate and are to be kept per the intention of one or more editor. It is a rule which humans and bots alike would have few problems in policing. Ohconfucius (talk) 08:51, 17 October 2008 (UTC)

"bots should not be used where there are exceptions, because they will not notice them" - again, the point of the proposal is that bots and human editors will notice exceptional date-links that are "earmarked" as exceptional. without earmarking them somehow, neither bots nor human editors have any way to recognize that the links [[31 December]] or [[1978]] should remain linked in one particular article. the proposal is pointing out a way to earmark the links that some editor values highly and wants to keep. if a link isn't worth the trouble of earmarking it, one might well ask whether it's really a high-value link. Sssoul (talk) 07:31, 17 October 2008 (UTC)

Quite correct. (2 messages up). The bot should not be run until a convention for "earmarking" dates is established and published, preferably for at least one month. You're claiming your argument is to the contrary? — Arthur Rubin (talk) 21:54, 17 October 2008 (UTC)

• If you accept that deliberately linked dates must be formulated differently to just placing square brackets around calendar dates and calendar years (and you do, don't you?), then it is the very next logical extension that we can resumed delinking all those which are [[mmdd]], [[ddmm]], [[mmmdd]], [[ddmmm]] or [[yyyy]], delinking by bot or by script regardless of where they are placed. I just fail to understand what further objection there could possibly be to restarting the delinking by bot? Ohconfucius (talk) 23:34, 17 October 2008 (UTC)

(outdent) The root question one should be asking is whether or not the benefit of having superfluous/deprecated links removed is worth the cost of having editors manage the exceptions to the rule, a question I sincerely hope to see answered by a broader audience. Shereth 17:53, 15 October 2008 (UTC)

I think most of us (but probably not all) would be perfectly okay with mass catch-too-much delinkings if we were somehow certain that the relinkings we made to the article after this bot-edit would not be reverted again by bots. Sssoul's suggestion is one way of ensuring that, but I'm sure there are others. -- Jao (talk) 18:04, 15 October 2008 (UTC)

• I agree with the questions/statements by both the above. I see no real issue to un-doing (whether by bot or by script) the date links which currently exist. Those date links which are 'deliberate' need to be re-made in another way which is obvious to bots other editors alike. Ohconfucius (talk) 08:46, 17 October 2008 (UTC)

Anderson and others, if it means so much to you to link a certain anniversary day or year (and I haven't yet seen one that is useful, frankly), then simply make an explicit piped link in the "See also" section. It is as simple as that, and everyone is happy. Tony (talk) 08:38, 17 October 2008 (UTC)

• The 'why should we put in extra effort to put "special" date links in? argument' is exactly as Tony says. If it's worth putting in, it should be worth the effort. And if it's worth the effort, there should be no complaints about needing to put the work in ;-) Ohconfucius (talk) 08:46, 17 October 2008 (UTC)

putting "highly-valued date links" in the "see also" section should perhaps be regarded as a "strongly encouraged option", since some people have expressed anxieties that a reader might somehow miss that section. the main point (as i understand it) is that making "highly-valued date links" explicit by including a non-date word in them is what will designate them (for both human editors and bots) as "this link is highly valued by some editor, not merely a remnant of autoformatting or overlinking". and that "earmarking" will serve its purpose no matter where in the article the link appears. if some editors really strongly prefer to add For more historical context see the list of [[1943|notable events of 1943]] or For whatever reason, see the list of [[12 May|assorted events that have occurred on 12 May throughout history]] to a paragraph instead of to the "see also" section, it's no skin off my nose. bots will leave those links linked thanks to the non-date words they include; human editors might debate whether or not the links are really useful and/or what the best position for them is, but they'll know someone feels they are valuable links. Sssoul (talk) 09:04, 17 October 2008 (UTC)

The (weak) majority that the year of birth and death in biographical articles should be linked seems as strong as the expressed argument (not consensus) that all years should be unlinked. Any bot needs to take into account appropriate exceptions. — Arthur Rubin (talk) 21:59, 17 October 2008 (UTC)

the bot - like human editors - will recognize exceptions that are designated as exceptions. the proposal is to designate "exceptional date links" by adding a non-date word to them. most people in the RfC above were talking about keeping the birth/death-years linked; all they need to do is add something like For more historical context see the list of [[1943|notable events of 1943]]. calendar-date links don't appear to be very high-value to most of the people in the RfC above, but if they are of value to someone, they too can be designated as exceptional. it's not that hard. Sssoul (talk) 23:11, 17 October 2008 (UTC)

That is clearly unreasonable. I would consider "(born July 4, [[1776|1776 ]])" the maximal acceptable tagging to be required of editors if the RfC fails to reach a consensus for exclusion. (The status quo is inclusion.) — Arthur Rubin (talk) 23:29, 17 October 2008 (UTC)

I suppose another alternative I would consider acceptable would to be to include:

• {{for|other events occuring in the year of birth|1943}}
• {{for|other events occuring in the year of death|2008}}

before the lead. You might consider that less acceptable than the status quo, as would I, but I would consider that an acceptable alternative. — Arthur Rubin (talk) 23:33, 17 October 2008 (UTC)

• There are two problems: (1) these examples are piped in a way that conceals from the readers their destination. (2) Positioning this type of link before the lead is far, far too prominent for what is almost certainly a pathway to a sea of irrelevant material. I strongly disagree with this suggestion on both grounds. "See also", when spelt out, is both unintrusive and more likely to be clicked on by readers than a concealed solitary year link in the running prose. I have no idea why you find objection in this solution. Tony (talk) 01:12, 18 October 2008 (UTC)
  • The [[1943|events of 1943]] approach violates the guideline for the #See also section (so it needs to be discussed in the talk page of that guideline), and conceals the name of the article. {{for}} in the #See also section seems appropriate for some cases, but the status quo that birth and death years are linked requires a consensus to overturn, as no consensus has been established that year links are always inappropriate. {{for}} in the lead seems an appropriate option, if you insist that the bot should be allowed to run amuck without consensus. {{for}} at the start of any section would be allowed by my proposed modification to the proposal here, and I see no reason why the link to the birth year should be moved out of the lead. I'd accept , as an alternative, the infobox templates emitting the year link, although you seem opposed to that, possibly because you think the year link is misleading, even though it's the actual name of article. If that is your reasoning, I can't understand why you think it's misleading, unless you want to propose moving all the year articles (and handling the templates which link to them). (I've got some idea how many templates are in question, considering the 1900s to 1900–1909 moves.) — Arthur Rubin (talk) 01:43, 18 October 2008 (UTC)

seeking common ground again: it seems that Arthur Rubin's concerns are with year links, not calendar-date links. it appears that so far only two people in this discussion object to unlinking calendar-date links that no one has "earmarked" as exceptional/high-value links. common ground is good.

as for year links (which were not meant to be the focus of this "compromise/interim proposal" - but so be it), they are currently misleading: autoformatting plus overlinking mean year links currently appear to be meaningless. a major part of the point is that when a year link *does* have meaning/value, "earmarking" it will make it explicit what the meaning/value is, which will increase the likelihood that readers might actually make use of the link.

if the main question is where exactly in an article the explicit/earmarked year links should appear, maybe suggesting recommended options would be sufficient: the "see also" section is one possibility - the info-box is another - adding a footnote would be another - a sentence added to the paragraph where the date appears would be another. Sssoul (talk) 12:42, 18 October 2008 (UTC)

Question for Shereth: You say "What's got me so animated here is that as an administrator I see (and deal with) complaints from sundry editors who have got their knickers in a bunch because of Lightbot removing some links and they do not see any consensus to remove links. I'm somewhat tired of having to deal with/respond to the situation." You later talk of editors who "pitch a fit" about the issue at a noticeboard. I have not yet asked, but need to now, whether this is still the case. How many editors have "pitched a fit" in your experience (it's strong language, so we're not talking of just queries and requests for where the practice is mandated, such as I've seen at Lighmouse's talk page). I'd be pleased if you placed evidence before us so we can judge the extent of the problem in numbers, intensity and timeframe. I note that new practices and policies, especially those that change long-established practices, are indeed the subject of emotional reactions by editors who may have spent considerable time and energy in inserting square brackets. But that is not what should concern us here, since editors have now been spared that manual labour in their creation of new text, and are greatly assisted by automated and semi-automated means WRT existing text. Lightmouse and others, including myself, have had considerable success in engaging with editors who query, or even complain of, the removal of the links in question.

Please be more explicit in laying out the evidence so that we can discuss it in informed terms. Tony (talk) 15:16, 18 October 2008 (UTC)

Oppose the linking of days of the year and years. All the proponents of these links should first earn their Sewer Cover Barnstar by honestly and truly accepting the challenge before coming here to run variations of this theme up the flagpole to test the winds. Absolutely no one actually *enjoys* reading these lists of trivia; the nearest I’ve seen an editor get to earning their barnstar was half the full challenge. And the opinion of that editor after that exercise was this: “That wasn’t so bad.” Well… that reaction comes up quite short of a ringing endorsement for linking to these God-awful articles. Step right up, you advocates of date linking; be the first to actually be able to stomach reading four entire trivia articles that the links take readers to. Then come back here and report to the others if your experience was…

1. Worse than having a stick poked into your eye.
2. Worse than getting on a bus and having to sit for five minutes next to a bum who smells like butt crack
3. A thoroughly boring experience and you don’t really expect any reader to actually read more than 2% of what’s there before hitting the “back” arrow on their browser.

After one of you has actually earned the barnstar, then we can talk about why you really think linking to these articles is such a thoroughly marvy idea. Greg L (talk) 01:37, 20 October 2008 (UTC)

What is going on here w/r/t linked dates?

Came here from the autoformatting subsection, where I left some comments, only to find that the apparent "consensus" behind altering the date/year linking policy is only Tony's cherry-picked talk page (in short, it strawmans his opposition; downplays the opposition clear on the subsection page and here; and blames date linking for errors caused by autoformatting, which are far more efficiently solved by removing preference autoformatting if it's a legitimate problem.) Has the only vote so far been about British v American date formatting?

What gives? Is there really no consensus? And if so, why is the policy changing and why are bots being developed to auto"correct" existing pages? -LlywelynII (talk) 22:50, 2 October 2008 (UTC)

I second this, and have argued with Lightbot's owner about this before. Linking dates allows readers to use "what links here" on dates to find out what occurred on that date, and lets readers quickly see concurrent events worldwide for a given article's scope. This bot shouldn't be running until there's consensus. If the changes aren't noticed right away, it can be a real pain to undo it's efforts. -- Kendrick7^talk 18:32, 5 October 2008 (UTC)

Autoformatting of dates is a pile of crap. It has been extensively debated for months. If Tony has a talk page that you don't think is convincing, that is a straw-man argument. Whether Tony's talk page is convinding or not, the concensus to not autoformat dates exists. Just read the talk page archives for this guideline. --Gerry Ashton (talk) 18:46, 5 October 2008 (UTC)

I do not agree with not linking the dates either and I would like to point out that the consensus was reached only after the 3rd or forth time of being no consensus.--Kumioko (talk) 19:28, 5 October 2008 (UTC)

About using "What links here" to find out what happened on a date: With autoformatting, "What links here" will pick up every reference published or accessed on that date, making it impossible to use autoformatting to find articles related to specific dates. —Remember the dot ^(talk) 19:45, 5 October 2008 (UTC)

That really only applies to dates from the 20th century on, maybe the 19th. There's no reason for this bot to be running around delinking dates from the fourth century, etc. -- Kendrick7^talk 22:02, 5 October 2008 (UTC)

"What links here" on 5 October 427 can only show you links to "5 October" or "427", it can't show you links to "5 October 427" specifically. All the centuries are mixed together, making it very difficult to use linking to find events that happened on a specific date. —Remember the dot ^(talk) 22:17, 5 October 2008 (UTC)

You can get pages mentioning 5 October 427 from the intersection of sets of "what links here". Gimmetrow 03:46, 6 October 2008 (UTC)

From 5 October 419 through 7 November 427 .... — Arthur Rubin (talk) 03:58, 6 October 2008 (UTC)

Your point? You could get every article linking to any date in any range by getting every article lining to every date in the range. Gimmetrow 04:11, 6 October 2008 (UTC)

Arthur's point is that if you take the intersection of what links to October 5 and what links to 427, you'll include an article containing that text: "From 5 October 419 through 7 November 427", although it has absolutely nothing to do with 5 October 427. -- Jao (talk) 18:06, 6 October 2008 (UTC)

The intersection may include other articles, but it will include every article you want. Gimmetrow 18:20, 6 October 2008 (UTC)

Yeah, I know that, Remember. Still: a reader might want to know what else was occurring in 427 so they can get a wider historical context to the article they are reading, and linking it let's them do that in one click. Why is that such a terrible thing? -- Kendrick7^talk 20:04, 6 October 2008 (UTC)

I have no problem with depreciating autoformatting. The question is w/r/t linked dates, particularly years. That seems very much not approved by consensus. Tony's arguments regarding "high value links" are rather silly. People may only click a few links upon visiting a page, but they don't click any of them by accident. If they click through the date, it's because they want context. More often, no one will click the dates, but it's useful information for those improving or examining year pages.

It boils down to reducing Wiki's information and functionality for aesthetics; personally, I'm against that. -LlywelynII (talk) 13:45, 6 October 2008 (UTC)

The old version of this manual used to encourage everyone to wikilink every single date. Your attempt to edit the manual can easily be interpreted as "go ahead and go back to the old policy of editing every single date, if that is your preference". This approach has been clearly rejected and your edit should not be allowed to stand. --Gerry Ashton (talk) 22:21, 5 October 2008 (UTC)

Plus the old policy was to only wikilink all full dates, which was done for the sake of autoformatting. For sole years, or month-year, the policy has always been to wikilink only when called for by WP:CONTEXT; thus, there has never been any consensus to wikilink all dates for the sake of linking. I'm not saying consensus can't change, just pointing out in what direction it would have to change, as many seem to be unaware of that. -- Jao (talk) 22:26, 5 October 2008 (UTC)

There's still no consensus that all year links should be removed, so Lightbot should be decertified as a bot, and those who unlink all dates using AWB or other automated systems, without checking each link for applicability, should be decertified for use of automated tools (after a warning). — Arthur Rubin (talk) 03:40, 6 October 2008 (UTC)

I second that removal of years has not been shown to be approved by concensus and decertifying Lightbot is an excellent idea, although I don't know where to go about saying so. Feel free to link to my support from the appropriate page. -LlywelynII (talk) 13:47, 6 October 2008 (UTC)

Agree that Lightbot should be decertified and anyone de-linking dates en masse should stop. --UC_Bill (talk) 14:55, 6 October 2008 (UTC)

And another agree here. Under the current discussion, Lightbot is well out of order. - fchd (talk) 16:00, 6 October 2008 (UTC)

• First, this argument that year links are necessary for the editors of year pages to orient themselves—to find leads to appropriate information to include on these pages, is utterly bogus. Has anyone heard of the search box? If you need to rely on WP itself rather than outside sources for your stimulus, just type in a year. Second, can someone point to the consensus for linking years in the first place? Tony (talk) 16:10, 6 October 2008 (UTC)

• There is no requirement to find consensus to allow for the linking of years and dates. Consensus would be required to enforce either always linking them (which no one is suggesting) or never linking them (which is what Lightbot is enforcing). There is no demonstrable consensus that these links must be removed, and given the concerns that continue to be raised it's continued use to remove all linked years is disruptive. Shereth 16:15, 6 October 2008 (UTC)
• Granted I may be paranoid that my keyboard will break but I don't like the idea that a subset of articles should be reachable only via the search box. — CharlotteWebb 16:16, 6 October 2008 (UTC)
• I realize that you think that these links should not exist. However, I think it is clear that there is no consensus for these mass edits. Please see Wikipedia:Requests_for_arbitration/Episodes_and_characters_2#Fait_accompli for why you should not being making these sorts of edits on a large scale without forming a consensus first. It feels like you're just trying to wear all of the opposition down by refusing to acknolwedge it and simply persist in making the edits until it's the status quo. -Chunky Rice (talk) 16:54, 6 October 2008 (UTC)

• This, I think, is what has me worried about this entire approach, and the Fait accompli from that ARbCom is exactly right. I've been watching the debate, and I completely understand and agree with the point of delinking dates and all that. However, this last step, completely depreciating date linking, was brought to the community (across many boards, appropriately), but only a 7 day period elapsed with maybe.. 20-odd editors responding during that time, and suddenly it was "consensus". I am not saying the consensus isn't there for this change, but clearly there needs to be more discussion of the issue. The matter should have been brought up via an RFC or a watchlist-details notice or some other means to invite a much larger discussion; this might have prompted different solutions (maybe the MediaWiki devs would have been kicked into gear to give us a usable autoformatting solution, but there have been other practical solutions such as templates as well after this change was made that seemed to have support) The end result would have likely been the same, but personally a result I would be more comfortable with it once a much larger discussion was made given the wide impact date linking has on WP. --MASEM 17:13, 6 October 2008 (UTC)

• Yes. Watchlist notification was discussed on 10 September but nothing was done. RFC was suggested on 13 September but it all seemed to get confused (Greg opposed RFC for a reason that seems to be of a personal nature, although he himself always said he wanted a larger input, did I get that right...?) and nothing was done there. No idea why, really. Of course, Tony's arguments have been visible in quite a few places, not only on MOSNUM, but still most people must have missed it (which would have been the case after an RFC or VP announcement as well, I'm sure; watchlist notification would reach more people). -- Jao (talk) 18:33, 6 October 2008 (UTC)

• I can't fault Tony on trying to spam (in a good way) as much as possible to get the word out, but the spamming was never really to a point of requesting input in a typical RFC fashion; I know when it was posted to WP:VG, it was more confusion on the point as opposed to any discussion. --MASEM 18:45, 6 October 2008 (UTC)

Well, Tony, if you only want year articles to be accessible via the search box, and not through links, the natural conclusion is that that should be the default for all links. - fchd (talk) 17:54, 6 October 2008 (UTC)

There is a difference between links that pass WP:CONTEXT and links that don't. We write "A demo version was released via download on May 1, 1999" and "the game received positive reviews from gaming websites" (examples from today's featured article), and nobody complains that the reader who suddenly feels an urge to read more about downloads, reviews or websites has to type those words in the search box. Why would a reader be more likely to wish to visit May 1 or 1999 than any of those three? -- Jao (talk) 18:33, 6 October 2008 (UTC)

For me it comes down to whether the year is "within living memory" or not. 1999 is within living memory, whereas an article dealing with events 70+ years ago is not, because a reader is increasingly unlikely already know the historical context of the article's subject or to have learned it from elders over the course of their lifetimes, and increasingly likely to need to know to really understand the article the further back in time we go. -- Kendrick7^talk

I'd like to point out a fundamental difference between linking years and linking other terms. After a short time, everyone gets to know that we have articles on most years. This is not so with other terms, such as dummy load. So linking some terms serves to alert readers that an article is available on a topic, when it isn't obvious this is the case.

I advocate using infoboxes or templates for significant dates, and I don't mind if the years are linked within those infoboxes or templates. I also don't mind having the first instance of a year linked in an article, if the year is significant. Obviously years that are present in the reference list are seldom significant.

As for the degree of scrutiny needed before using a semiautomatic tool to delink the dates in an article, I believe a person should skim the article and get a sense of the state of the dates in the article. The use of a semiautomatic tool is justified when there are a number of inconsistent date formats in the article, or when nearly all the dates are linked. The use of a semiautomatic tool is not justified when the dates are in a consistent format and only a few dates are linked. --Gerry Ashton (talk) 22:02, 6 October 2008 (UTC)

User:Lightbot paused

I have paused User:Lightbot (at least if it behaves according to the instructions) per the above concerns. I would like to see some sort of consensus here that the task of de-linking dates has any kind of consensus prior to resuming the bot's work. Thanks, Shereth 16:07, 6 October 2008 (UTC)

• I can’t profess to be unbiased on this issue. I think linking to material that is unrelated to an article is unwise. I’ve written an essay on the matter (here) and expanded on that essay here on my talk page.

  Evidence for a consensus is unclear at this point. The above poll and discussion showed opinion was about evenly split (17 to 15) to no longer link the dates of births. I also believe there has been a developing consensus lately that the linking of calendar days (like March 12) is worse than linking years. Linked years is more of a grey area since there are more circumstances (like history-related articles) where the judicious linking of years is thought by many to be appropriate.

  There also seems to be an intertwining of issues. By de-linking calendar days/years, the bot was also removing autoformating. Autoformatting, which produced *prettier* results only for A) registered editors, who B) set their user preferences, was deemed as unwise by a consensus and has been deprecated.

  The complexity now, is that linking of dates is part of autoformatting and this won’t change until the developers disable the autoformatting function of the links. As a necessary consequence of delinking, Lightbot was replacing them with fixed-text dates in a specific format (Euro/International, or US). This aspect alone brings out passions and opinion is all over the map on how to choose date formatting in articles. A guideline that would key the date format to what is most appropriate to the subject matter failed and the current guideline is weighted towards defaulting to what the first major editor used. So formatting of dates after Lightbot has visited is intertwined with the issue of delinking dates.

  I would propose that we all get onto the same page as to whether there is any meaningful difference between linking of dates and autoformatting of dates (for simple years, like 1987, there isn’t), and try to progress forward from there. Trying to arrive at a consensus is made more complex by the fact that many editors arrive here late to the discussion after articles have been affected by Lightbot; we have to start from square-one with these editors. Greg L (talk) 19:41, 6 October 2008 (UTC)

  • We can wikilink dates without invoking autoformatting: [[March%2012]], [[2008]] will give March 12, 2008. Mind you, I strongly suggest a template form for this instead of hand-writing it. --MASEM 20:26, 6 October 2008 (UTC)

Greg, you wrote:

• the bot was also removing autoformating and Lightbot was replacing them with fixed-text dates in a specific format

Lets be clear about one thing. Lightbot does not delink autoformatted dates. Many people would be delighted if it did, but it does not. Lightmouse (talk) 21:37, 6 October 2008 (UTC)

• • I have to say, it would be great if you would write up a description of what Lightbot does on it's user page because right now, there's no way to tell, as far as I can see. -Chunky Rice (talk) 21:45, 6 October 2008 (UTC)

• Indeed. Let’s get it from the horse’s mouth. I’ve struck the contested text. What are the true facts here Lightmouse? What is your bot doing that has editors’ nickers so in a bunch? I’ve clicked on some of your activity as assisted by some AWB software and the result was the deletion of double brackets around dates. Of course, I completely agreed with what you were doing there and think it improves Wikipedia. And I think you properly read the general consensus when you made your move with AWB. But now I’m confused. Are there two kinds of computer-assisted activities going on here? Note further that by taking away the brackets, the dates get locked into their raw way they were coded. For editors who were looking at the world through their damned date preference setting, many would think AWB was changing the date format. The effect of AWB is confusing to some and this is aggravated by the thoroughly moronic action of autoformatting, which gives only some editors a special, rose-colored view of editorial content that no regular user sees.

  There is no point revisiting the issue of what date format to use in articles; that was thoroughly hashed through, starting here in Archive 110, via two run-off-style polls. It hasn’t even been a month since then, so it is unlikely the mood has changed.

  So task at hand is to push for a clear consensus on the circumstances under which it is appropriate to employ links to calendar days and years. Not too many editors disagree with the premiss that links should be sufficiently topical and germane to any given subject to merit being linked to; the issue is where to draw the line and how to memorialize the nuances in an easy-to-follow, clear guideline. Any bot activity should narrowly limit itself to whatever that guideline calls for. Greg L (talk) 21:56, 6 October 2008 (UTC)

• Whatever comes of this (or any related) discussion, I'd also like to see the question of whether or not bot or script-assisted removal of wikilinks to dates/years is appropriate finally put to rest. Shereth 22:02, 6 October 2008 (UTC)
• (EC)Thanks to an idea in your sig, we can nowrap the "faked" date to prevent it moving about. (see User:Masem/datetest for an example). Again, this needs to be simplified via a template, but its doable. Just that the template needs to know what format to pump out. --MASEM 22:05, 6 October 2008 (UTC)

I am always a little surprised when I come across the assertion that Lightbot delinks autoformatted dates. It delinks any date except autoformattable dates. That is it. A solitary year is not autoformattable. I personally like the phrase 'date fragments' but some people didn't like that. The issue was extensively discussed in the bot approval. The bot user page (User:Lightbot) provides links to its three separate approvals, look at the bullet points in the one called 'Lightbot 3'. I wrote it in bullet point form in an attempt to make it clearer. If you are still uncertain about what a date that isn't autoformattable means, come back to me. Lightmouse (talk) 22:19, 6 October 2008 (UTC)

• Maybe we are talking about two different bots. What about this example, Lightmouse? Let’s agree on the simple facts here. Greg L (talk) 22:24, 6 October 2008 (UTC)

← And we come to the crux of the issue, Masem. There are nice ways to link fixed-format dates that circumvent autoformatting and gives all visitors to Wikipedia the same date format. On Sept. 16, we settled on the date format editors should use in articles. So now the issue to settle is the circumstances under which dates should be linked. Here’s my take:

If someone is reading up on the famous architect Frank Gehry, providing a link to beautiful architecture, like Falling Water, is a good idea. But…

We devalue links and bore most readers of that article by providing a 1929 link to an article that says “March 3 - Revolt attempt of Generals José Gonzalo Escobar and Jesús María Aguirre fails in Mexico.”

• The issue is not whether or not these lists have any socially redeeming value whatsoever; it is whether or not they are sufficiently topical and germane to any given subject to merit being linked to; that’s all.

  The nearest thing to a completely random list that has been successful is the Guinness Book of World Records. But, given the nature of what’s in that book, and the fact that is is organized into classifications (natural disasters, human feats, etc.), it can actually be read rather linearly with some measure of enjoyment. Wikipedia’s random lists of who-knows-what come up quite short of “compelling reading.” I don’t buy into the implicit argument that ‘since nearly everything is in date articles, they are suitable links to put into any article.’ To rebut that attitude, I submit How to Bore People in Five Simple Steps.

  Links to years in truly historical contexts are appropriate: in an article on the Great Depression, judicious use of links like 1929 make sense and do a good job of exploiting the promise of hyperlinking, as first envisioned by Paul Otlet in his 1934 book, Traité de documentation (Treatise on Documentation) as interestingly covered here on YouTube.

  But for general-purpose uses like birth years? I don’t think so; if visitors are reading a Wikipedia’s article on, for instance, Frank Gehry, they are most likely there because they are interested in famous architects and beautiful architecture. Accordingly, we add value to the Frank Gehry article and encourage learning and exploration by providing a link to Falling Water, not by linking to 1929 (the year Mr. Gehry was born). But if there was an article on Notable architectural events of 1974 (the year of his first major design), then by all means, let’s provide a year link to that article.

  As for specific calendar days, like like March 12, so few readers would be interested in wading through any of these lists, we would only diminish the value of links and desensitize readers to them were we to link to them.

  I also think Wikipedia’s Fairness In Advertising policy ought to be better applied. For specific calendar days (which ought to be quite rare) links would work as follows:

Pearl Harbor was attacked December 7, 1941 (list of random events throughout history on Dec. 7).

There’d be far fewer of date links being clicked on after that. In all seriousness, I suggest that year links be aliased so they better disclose to the reader what they will be taken to. I suggest as follows:

The Great Depression followed “Black Thursday” which occurred on October 24, 1929 (other notable events of 1929).

Greg L (talk) 22:24, 6 October 2008 (UTC)

Hmmm, a date that will live in infamy, but apparently not be linked. Maybe we should just say that Pearl Harbor, by amazing coincidence, occurred on Pearl Harbor Day. I think birthyears should be linked, because the world a person is born into tells you a lot about their life, and the year articles exactly provide that. -- Kendrick7^talk 02:50, 7 October 2008 (UTC)

• I can see what it was approved to do - which appears to be whatever it wants. What I want to know is what it actually does. Please write a short summary. -Chunky Rice (talk) 22:34, 6 October 2008 (UTC)

To answer Greg's question... you are not referring to Lightbot. You are referring to Lightmouse. Lightbot can run when I am asleep, Lightmouse can't. The Lightmouse contributions often involve a script and my fingers pressing 'Save page'. I find it difficult to answer the request by Chunky Rice because it does a lot, there are hundreds of lines of code. Think of the list of all things that might be called a 'date', then think of a list of all things that might be called a 'valid autoformatted date', then subtract the latter list from the former list and you will have a list of all the things it might delink. For example, in its last edit, it removed one link to '1961' and one link to '1968'. You can see from its recent contributions that it is mostly solitary years because that is what most non-autoformattable dates are. Lightmouse (talk) 22:48, 6 October 2008 (UTC)

• Maybe the simplest way to convey what Lightbot does is to provide four links here that illustrate its typical activity with dates. Greg L (talk) 22:51, 6 October 2008 (UTC)

I am confused by that request. There are over 380,000 examples. You can pick any one of them just by going to the contributions. Why are we doing this? Lightmouse (talk) 22:58, 6 October 2008 (UTC)

• The problem I have is that the approval includes "other edits," which aren't specified, to edit dates/numbers etc. as "part of general MOS guidance", which is also very vague. I don't have a good idea what this bot does. Just a sentence saying, "This bot unlinks non-autoformatted dates." would be helpful. Right now, there's no way to tell what it's doing. -Chunky Rice (talk) 23:00, 6 October 2008 (UTC)

Lightbot unlinks non-autoformatted dates. Lightmouse (talk) 23:01, 6 October 2008 (UTC)

And that's the entirety of what it does? -Chunky Rice (talk) 23:02, 6 October 2008 (UTC)

It is also approved to edit units of measure in a variety of forms. Note that approval might not translate into activity. Lightmouse (talk) 23:09, 6 October 2008 (UTC)

If you want to know what it is approved to do, see:

• Wikipedia:Bots/Requests for approval/Lightbot 3 and the two other approvals (links on that page)

If you want to know what it actually does, see:

• Special:Contributions/Lightbot

it is currently focussed on delinking solitary years because people believed such links as inferior to autoformatting links. There seems to have been a flip flop in that belief. Lightmouse (talk) 23:15, 6 October 2008 (UTC)

I think there is not so much a flip-flop in belief, as a change in which belief is being discussed. Towards the beginning of the discussion it was brought up that many editors saw that full dates and calendar dates were linked to enable autoformatting, so they just linked every year in incorrect imitation of what they saw. It think that's true, that's what really happened. Now the discussion has shifted to "what about the years that were linked deliberately?" Of course, Lightbot can't tell the difference. Perhaps if Lightbot could search an article for unlinked years, and not operate on any article that contains an unlinked year, that would reduce the problem. After all, if some years are unlinked, that would imply that the ones that are linked were deliberate. --Gerry Ashton (talk) 23:28, 6 October 2008 (UTC)

Lightbot can't do that, it can only work with a few characters and sometimes a whole line. Even if it could, that would mean that the four useless links to '2003' in The Escape Engine would not be unlinked because the year '2002' is not linked. Or the useless links to '2009' in Upcoming Telenovelas could not be unlinked because there is an unlinked year '2008' (that article is definitely overlinked because it also contains linked solitary months). Lightmouse (talk) 23:52, 6 October 2008 (UTC)

Sorry, but that sounds like an even stronger case for Lightbot to be dropped. While I generally support delinking of individual years as I believe the links have limited value (and I've delinked a number manually when making other edits to articles), it is clear from the above debate that a significant proportion of the editors here do find value in them. Also, when and if Lightbot is re-started, and all it is doing is delinking standalone year links, perhaps a more informative edit summary than "(Date links per wp:mosnum/Other)" might be in order? - fchd (talk) 05:27, 7 October 2008 (UTC)

... in case a newcomer's perspective is of any value: as Gerry Ashton wrote: "many editors saw that full dates and calendar dates were linked to enable autoformatting, so they just linked every year in incorrect imitation of what they saw." having gone through that exact phase not very long ago, i really welcomed the new policy deprecating date-linking because of its beautiful clarity. as you can see from my edit history i was unlinking/reformatting dates manually for a while, then tried a script for a day or two; and i'm deeply dismayed to learn that the policy is still so controversial. but reading the arguments being presented here ... it seems people agree that the autoformatting needs to be either abandoned or changed to template form; it seems people agree that not every date should be linked; it seems people agree that some dates (mainly years) do deserve to be linked. the trick is to formulate a rule that's clear (including to newcomers).

it's simplistic but: what about putting links to the date pages that people consider important/valuable in "see also" sections, rather than making them "in-line" links? Greg L's suggestion that such links should be identified as (for example) {([[1929|other notable events of 1929]])}} would work very nicely in the "see also" sections, as would "1978 in music"-type links. and it seems like it would be clear enough (even to people who haven't read the policies) that not every date mentioned in an article needs to be listed there. Sssoul (talk) 09:40, 7 October 2008 (UTC)

If anyone hasn't noticed, Lightbot has re-started again. - fchd (talk) 11:03, 7 October 2008 (UTC)

Apologies, that appears to be Lightmouse the user rather than Ligthbot the bot. Either way, the end effect is about the same. - fchd (talk) 11:24, 7 October 2008 (UTC)

Sssoul's idea sounds like an excellent move. There would be nothing more disconcerting to readers than to see some years bright blue and some black. Consistency in the main will be preserved, and the few occasions on which year pages might be deemed vaguely relevant to a topic may be convered in the "See also" section. Tony (talk) 12:01, 7 October 2008 (UTC)

Putting [[1929|other notable events of 1929]] in 'See also' rather than in the main body sounds good to me. Lightmouse (talk) 14:35, 7 October 2008 (UTC)

Excellent suggestion, Sssoul! That sounds like a beautifully phrased compromise. I wholeheartedly agree.--Aervanath lives in the Orphanage 16:11, 7 October 2008 (UTC)

Of course the article year X is going to link to notable events in year X, it's WP:COMMON sense. Readers looking for temporal context shouldn't have to scroll all the way down to the see also section; that's a terribly WP:BURO-cratic solution. We wouldn't do something like that for articles providing geographical context (i.e. a link to Azerbaijan), we just use an inline link and everyone is happy. -- Kendrick7^talk 16:18, 7 October 2008 (UTC)

Mostly going to have to agree with the above. The compromise sounds nice at first, but I don't really see it as a solution. "See also" is the appropriate place for related topics that can't be linked in the main body of the text; inline links are always superior, if for no reason other than the fact that readers interested in context should not be expected to scroll to the bottom of the article. If a link to a year (or date) is appropriate to the context of the article, it is appropriate as an inline link. Shereth 16:23, 7 October 2008 (UTC)

In-line links and citations are always better. Perhaps we should reflect on why and when an editor should link to a date, rather than how. A bullet-point list of criteria in the style guide should suffice; and perhaps linking should be the exception rather than norm. Millstream3 (talk) 16:26, 7 October 2008 (UTC)

I still like "within living memory" as a good rule of thumb, which would make most everyone happy, at least from the examples most people are providing against linking, which involve years from 1990 onwards. Links to years even octogenarians can't remember anything about which provide temporal context to the article are OK, links to years less than 70 sols ago are generally to be avoided. If I'm writing an article that involves the year 1058, I insist that this year should be linked, and I'm not going spend the rest of my life reverting LightBot and script-kids every few days. -- Kendrick7^talk 17:25, 7 October 2008 (UTC)

plainly there's a whole range of different views, but a "within living memory rule of thumb" is way too arbitrary to address the main problem i see with the date-linking - which is that unclear/inconsistent policies give too many people the mistaken impression that *all* dates should be linked. and the fact that it can be difficult to decide which geographical place names to link isn't (to me) an argument in favour of leaving excessive masses of dates linked for no reason - which is the current situation.

everyone in this discussion so far seems to agree that currently there *are* too many date links, mainly due to the now-deprecated (?) autoformatting, and to editors who think that since some dates are linked then *every* date should be linked. the bots/scripts were developed to assist in undoing some of that excess. i understand the objection to the bots/scripts - in the course of undoing masses of useless/ill-conceived date links, they've also undone some date links that someone felt were useful. so the point is to find some way to eliminate the excess date links and the confusing principles that mislead people into excessive date linking without doing away with date links that some people consider valuable.

some people who want to keep certain specific date links feel that scrolling down to the "see also" section is too much trouble. but leaving some dates linked creates an ongoing need to undo overzealous date linking - which is *also* too much trouble. so what other compromises do people propose for a clear and consistent policy on date linking? Sssoul (talk) 18:06, 7 October 2008 (UTC)

ps: Kendrick7 wrote: "Readers looking for temporal context shouldn't have to scroll all the way down to the see also section", and Shereth wrote: "readers interested in context should not be expected to scroll to the bottom of the article." i don't think i understand why not - if someone is interested in the temporal context, skipping to the bottom doesn't seem particularly difficult.

but if that's really too much to ask of interested readers, maybe a template could be created to add a box of "links to dates mentioned in the article" to the "contents" box on articles where there are editors who feel strongly about making it ultra-simple for readers to jump to date pages. Sssoul (talk) 20:13, 7 October 2008 (UTC)

I don't think a rule of thumb for articles relating to 70 year old+ events is any more arbitrary than the argument that we can't link to the year 472 because too many articles link to the year 2005. In my opinion, you're alternatives fail WP:CREEP; we can put that in the rules, but no one is ever going to go to this much trouble. -- Kendrick7^talk 20:38, 7 October 2008 (UTC)

um ... i haven't raised any arguments related to the number of articles that link to 2005. the "70+" idea seems to me pretty arbitrary in its choice of "cut-off date"; but the main problem i see with it is that it will *look* arbitrary - for example in biographies of people whose lives/careers "straddle" the cut-off date. policies that look arbitrary won't be very helpful in alleviating confusion over what dates to link.

as for WP:CREEP, i don't think my proposals would require elaborate instructions. "don't link dates in articles; links to important dates can be added to the 'see also' section" seems pretty straightforward. (yes, a template attached to the "contents" box would call for a few more instructions - that's one reason i prefer the "see also" proposal.)

"no one is ever going to go to this much trouble" ... well, everything is "too much trouble" if no one feels strongly enough about it. i thought the whole point was that some editors feel strongly about making it ultra-easy for interested readers to link to some year pages. if that's not the case, let's go back to the "see also" idea.

anyway to reiterate: the proposals so far seem to be:

• link all years prior to 1939 and unlink all other dates - is that right? (i don't know anything about bots/scripts so someone will have to chime in about whether a date-unlinking bot/script could be taught to do that. i feel this policy wouldn't do much to alleviate the confusion about what dates should/shouldn't be linked, but ... the confused will always be with us, i guess.)
• unlink all dates in articles, and put date links someone considers important in a separate section - either the "see also" section or a box that could be appended to the "contents" box on articles where someone wants it. (i hope date-unlinking bots/scripts could be taught to leave sections like that alone. maybe this is "too much trouble", or maybe it sounds promising.)
• unlink all dates. (bots/scripts exist that can assist with this, but some people protest that certain valuable date links are being or may be unlinked.)

any other ideas? Sssoul (talk) 21:38, 7 October 2008 (UTC)

I think the best bet is for editors to link dates when they believe that the date provides valuable context, and not link them when they do not. I don't expect editors to have a problem exercising this type of editorial judgment. Christopher Parham (talk) 21:57, 7 October 2008 (UTC)

... but the current situation is that there are masses of date links that are *not* based on editorial judgement - they're based on the now-defunct autoformatting policy and on misunderstandings of it and/or of other policies. the masses of ill-conceived links need to be eliminated; the question is how to designate date links that someone feels are genuinely valuable for understanding the article so that those don't get eliminated along with the useless/ill-conceived date links. Sssoul (talk) 07:51, 8 October 2008 (UTC)

There are also masses of dates that have been unlinked not based on editorial judgment, by this bot - the best way to ensure that date linking reflects the judgment of editors is to leave the decision for editors to make an a case-by-case basis. Christopher Parham (talk) 02:32, 9 October 2008 (UTC)

What we were to only allow year dates to link to "YYYY in field" pages, each of those having a separate table for other "YYYY in field" pages? That is, say I've got an article on a politician (and only a politician); then date links from that page would link to "1999 in politics" (and possibly "1999 in United States politics" if the field is considered too large). If the topic was a crossover, the editors would have to select the best appropriate links, so a politician that may have been a professional athlete before would have both "in politics" and "in sports" year links. In other words, this is sort of a category structure (which it what sounds like people want but keeping it inline). Now, and I would say this is critical, this works under the assumption that we normal avoid surprise links (eg linking to "YYYY in field" but only displaying "YYYY" with no additional context), but if we made this universal across pages, this would no longer be a surprise.

The unfortunate drawback is that this cannot be bot assisted, at least easily. A bot might be able to determine the page's primary field by looking for the first WikiProject on the talk page, but this is going to fail on crossover articles, and there's potential for hit and miss. Individual editors would be needed to standardize this approach Wiki-wide. --MASEM 22:04, 7 October 2008 (UTC)

well ... does every year that some editor feels is important to link to have associated "YYYY in field" pages? i kinda doubt it. Sssoul (talk) 07:51, 8 October 2008 (UTC)

A technical response to Sssoul's questions about bot capabilities:

• a bot can delink all dates after a 'threshold date' such as 1939
• a bot can delink all dates except those that contain a non-date word such as [[1929|other notable events of 1929]]. But it can't distinguish between [[1929]] in one section and [[1929]] in another because a bot doesn't know about sections.
• a bot can delink all dates (we already knew this)

My other idea: full date linking (autoformatting) is the disease, overlinking of partial dates is merely a nasty symptom that has got out of control and keeps coming back. We could try for consensus for bots to treat the disease rather than the symptom. I am sure many of the pro-delinking people would support that. Lightmouse (talk) 22:08, 7 October 2008 (UTC)

For the record, I didn't suggest linking to years mentioned prior to 70 years ago would be mandatory! While I generally agree with Christopher above, if LightBot could be taught the difference between 1939 and 1939 BCE/1939 BC (well, those articles don't exist yet, but you get the idea), I would have no objection to it making a one time pass to de-link all years and decades after 1939. I would guess that would cover 90% of all year links, given Wikipedia's tendency towards WP:RECENTism. -- Kendrick7^talk 22:17, 7 October 2008 (UTC)

Yes, it is easy for a bot to distinguish between solitary years such as [[1939]], [[1939 BCE]] and [[1939 BC]]. I notice that there is increasing acceptance that full autoformatted dates should also be delinked. That could be done at the same time. Lightmouse (talk) 22:44, 7 October 2008 (UTC)

From the RfC above, there is no consensus at all to unlink dates of birth and dates of death at the top of bio articles, whether full or not. And in the absence of a clear RfC that can be linked to, I'd suggest there's not much evidence of consensus to delink any other dates either. Jheald (talk) 23:09, 7 October 2008 (UTC)

While I think dates should be unlinked, I have to agree with Jheald: there's no consensus for a mass unlinking of anything quite yet.--Aervanath lives in the Orphanage 04:25, 8 October 2008 (UTC)

I'd agree that dates are overlinked (unless we can get the MediaWiki dev's to incorporate geo-presence for formal date formatting), but I'm most unhappy with the current mass unlinking. I'd suggest it stop for now, except by strictly manual methods.

I'm still intrigued by LightMouse's comment on LightBot's method: "I find it difficult to answer ... because it does a lot, there are hundreds of lines of code". Interesting that, bot approval is just a matter of confusing up the code 'til no-one can understand it? Changes in guidelines are immediately enforced with spaghetti code? Trust me, it really does work, honest. Hmmm. Franamax (talk) 08:40, 8 October 2008 (UTC)

Response to Aervanath and Franamax: so it's a case of WP:IDONTLIKEIT, is it? By that I mean, you just appear to be unsettled by the kind of prompt adaptation of which wikis were built for. The longer the cancer of overlinking and the dysfunctional date autoformatting is left, the harder it is to fix. Every new editor comes to WP and copies the practices they see. It is not practical to make such an important change in slow motion. Were you thinking of a decade-long program? Tony (talk) 10:17, 8 October 2008 (UTC)

Well, no, I try not to frame my arguments in blue. I do generally accept the overlinking rationale (pending geolocation auto-preference, wherein date-links would make perfect sense). What I'm not comfortable with is the pace and scale, in particular when I see bot-op and script-assisted edits. I worry about what gets left in the dust behind the vehicle. In particular, I'm not clear on when exactly date-linking is appropriate. Did we arrive at a consensus somewhere that it shall never ever occur? Franamax (talk) 10:32, 8 October 2008 (UTC)

a few people have expressed this greater confidence in date-unlinking that's done manually - which puzzles me some. as long as links are not designated as "this is a link someone thought about and wants to keep", doing the unlinking manually just means it takes longer than doing it with the help of a well-designed script or bot. i don't see the point of slowing down a process if there's agreement that it needs to be carried out. if someone doesn't support the process then i don't suppose they want it carried out slowly *or* rapidly.

moving well-founded links to the "see also" section and "piping" them when necessary - for example [[1965|Other notable events in 1965]] - would be a way of designating them as well-founded, considered, intentional, etc. maybe there are other workable ways to designate them, but that's one suggestion on the table at the moment. Sssoul (talk) 12:08, 8 October 2008 (UTC)

Comment. I'm glad Lightbot was paused. While I agree that irrelevant years should be de-linked per WP:CONTEXT, it appears that some people believe this means "articles about years should be orphaned". And I don't think a bot can be able to understand whether a link is relevant or not. I did remove the link in http://en.wikipedia.org/w/index.php?title=Quintic_equation&diff=244401457&oldid=239497393, as the fact that the theorem was published in 1824, rather than in 1624 or in 1924, is totally irrelevant to the point being made (that there is no formula for general quintic equations over the rationals in terms of radicals); on the other hand, linking the year when somebody was born, or a historic event happened, or a book was published, in the article about the person/event/book itself, provides the historical context in which the person lived, etc. -- Army1987 (t — c) 16:59, 10 October 2008 (UTC)

Avoiding linebreaks in dates

If dates aren't linked, allowing a line break between month and day looks bad. But having to add &nbsp;s to every date would be a pain, and hard to read while editing. If autoformatting was able to convert a linked date in accordance with a user's preference, could it be made to produce an unlinked, hard-spaced date? If not, what about creating a set of templates to do that? So {{1 January}} would produce "1&nbsp;January" by default, or "January&nbsp;1" for users with that preference. And the other way around for {{January 1}}.
—WWoods (talk) 19:32, 15 October 2008 (UTC)

Developers do not put improvements into production on a timetable that allows interaction with the editing community. That makes consensus-building with developers just about impossible. Without developer assistance, there is no access to preferences, and even if there were, few readers have preferences set. So WWoods proposal amounts to always using the format "15&nbsp;October 2008" and there is no consensus to always use that format. --Gerry Ashton (talk) 20:29, 15 October 2008 (UTC)

'Avyear' template

We have a template {{avyear}} in the aviation project (as in Template:Avyear) which links to years in aviation, it is used only in the infobox next to an aircraft's first flight date. First flights are listed in the relevant aviation year section, I strongly believe that this linking is totally in context and was very disappointed to see it delinked by Lightbot, there are now 100+ 'year in aviation' articles that are not easily accessible. Notwithstanding previous discussion about 'years in aviation, music', easter eggs etc I think it is (was?) a fantastic feature. I was demotivated after seeing many hours work undone and stopped editing for a week or two (not a protest, just very cheesed off). Sure, don't link to individual days or years if they are not in context but please leave an easy way of linking to very relevant 'year' articles. Nimbus (talk) 23:56, 20 October 2008 (UTC)

It is (or ought to be) a general principle in hypertext that the text of links should make clear where the link takes. If someone sees a printed version of that page, would them be able to figure out what article the 1962 refers to? Remember, not everyone has the ability of hovering on links to know where they go in advance. -- Army1987 (t — c) 09:02, 22 October 2008 (UTC)

Surely none of the links on a printed version of an article will work?! It seems discussion on this date linking subject is not over yet, watching with interest. In the meantime I will neither link or unlink 'aviation years' to stay safe. Nimbus (talk) 15:34, 22 October 2008 (UTC)

In printed encyclopedias, including some predating the WWW, I've seen marks in an article suggesting to read some other article, such as underlining its title or preceding it with a picture of a finger. Now imagine one has a complete printed copy of Wikipedia. If someone sees "atomic nuclei", they can realize that the editor was suggesting to read the Atomic nucleus article, despite the irregular plural; writing "atomic nuclei (see Atomic nucleus)" would be overkill. But which would be clearer, "1962" or "1962 (see 1962 in aviation)"? -- Army1987 (t — c) 11:34, 23 October 2008 (UTC)

Nimbus, there is no reason for Lightbot to be delinking {{avyear}}. First, there is no consensus for it, just a strong desire for it among some of those who equally strongly favor swift eradication of all other extant date-related linking (such removal also not having any broad consensus for it). Secondly, MOSNUM has no authority to eliminate historical date linking, which it is effectively doing by such delinking, since it lies outside the stylistic purview of MOSNUM. Askari Mark (Talk) 20:39, 25 October 2008 (UTC)

Unwikilinking full dates?

I have to say, I'm seriously thoroughly ticked off over this whole recent move to unwikilink full dates. Not so much over the fact that it was done, but moreso over the fact that I can't see that it was debated or discussed very well, or openly. Sure, I see several discussions at various pages, but I still can't see any single debate of "yes, do it" versus "no, don't". I mostly see several isolated discussions by very few editors that somehow someone seems to think represents the consensus of everyone. Seriously, what happened is not "consensus" -- it looks more to me like the cabal dictating its way on everyone.

What I really dislike about not having full dates wikilinked isn't so much the autoformatting issue, although I think we've opened up a HUGE can of worms on that one since, without autoformatting, there's inevitably going to be more edit wars between the British/European date format versus the American one. But what I really don't like about non-linked full dates is the fact that, in text, the date is not highlighted in any way, and it doesn't stand out like it did with the link there. I think the wikilinked full dates added some emphasis on the date in the article and, with particular attention to articles about historical topics, this was especially helpful. Taking these wikilinked dates away is doing a serious disservice to our readers.

I'll probably just get overruled by the cabal for this, if not outright banned for being an ass ;-), but I would strongly urge the community to reverse this decision. Dr. Cash (talk) 22:23, 21 October 2008 (UTC)

I didn't know about this change either, but I firmly agree with it. It always annoyed me to see those blue dates for very little reason that I could discern. I assume that casual visitors to WP were equally annoyed — or at least confused. I say, let it ride for a while and see if the sky falls on us. Sincerely, your friend, GeorgeLouis (talk) 22:31, 21 October 2008 (UTC)

The discussions about this issue were ongoing for YEARS. If it was decided by the Cabal, they certainly took their time. Kaldari (talk) 22:47, 21 October 2008 (UTC)

If someone is reading up on the famous architect Frank Gehry, providing a link to beautiful architecture, like Falling Water, is a good idea. But…

We devalue links and bore most readers of that article by providing a 1929 link to an article that says “March 3 - Revolt attempt of Generals José Gonzalo Escobar and Jesús María Aguirre fails in Mexico.”

• Derek: Yes, what Kaldari said. With what must be tens of thousands, if not hundreds of thousands of Wikipedians (registered and not), there is no way that I know of to communicate to them all and keep them apprised of what is being contemplated here. For what it is worth, there are many editors who frequent this venue and many more still who occasionally come here to chip in with their views on maters. As a result, there has been—and continues to be—a wide variety of views expressed here. Note too, that this move was very thoroughly discussed (albeit in your absence). And whereas this issue certainly hadn’t been discussed out on the curb in front of your house, I can assure you that it was discussed “openly”: here on Talk:MOSNUM. I can also assure you that there have been a number of editors here acting as an ambassador of sorts who have been advocating your position.

  I believe too, that the recent move is very wise. Even though these blue links for days of the year and years are available to link to, doesn’t mean it is a good idea to actually link to them and turn yet another bit of body text to blue; far too few readers are interested in actually reading more than four or five line items in these random lists of pure trivia the first time they click on such links. From thereafter, they rarely click on them again. Links within Wikipedia articles should always be topical and germane. Properly chosen links anticipate what the readership of any given article would likely be interested in further reading. If you disagree on this fundamental point (whether these trivia articles are of sufficient interest to our readership to merit linking to them), I encourage you to accept the Sewer cover barnstar challenge. If you can actually earn such a barnstar, then I think you will have a stronger case for advocating the linking of these dates. Greg L (talk) 22:53, 21 October 2008 (UTC)

• As the change, if fully implemented, would effect the majority of articles on Wikipedia (although many would be improved), it should have been discussed on the Village Pump (policy) before implementation was done. It wasn't. It's now (in the past month or two, long after the guidelines change) being discussed in relevant WikiProjects, namely Wikipedia:WikiProject Years and Wikipedia:WikiProject Days of the year. — Arthur Rubin (talk) 01:37, 22 October 2008 (UTC)

• Some aspects of this change were discussed at Wikipedia:Village pump (technical)/Archive 46#Autoformatting standard violation . --Gerry Ashton (talk) 01:45, 22 October 2008 (UTC)

• True, but that's the technical aspects. The consensus was that autoformatting is deprecated, not date linking, although there is now probably a consensus that dates should almost never be linked (although that wasn't pointed to by the VP), but there's no similiar consensus for years. — Arthur Rubin (talk) 01:55, 22 October 2008 (UTC)

• The policy has been for years that full dates are linked always, only because of autoformatting. Without autoformatting, the policy on full date links is the same as the policy on other date links and linking in general, that they are linked when especially relevant to the context of the article. —Centrx→talk • 03:58, 22 October 2008 (UTC)

Derek Cashman, I've decided not to ban you "for being an ass". :-) Instead, I simply want to point out the dilution principle (one of four or five significant disadvantages that were never properly considered when we were fooled into adopting this autoformatting thing back in 2003). Every additional bright-bluing of text comes at a cost, which is that it vies for the reader's attention with links in the vicinity. That cost needs to be balanced against the benefit of the blue. There's no benefit at all to readers, since they're not logged in. Tony (talk) 02:06, 22 October 2008 (UTC)

Arthur, the disparagement of chronological links has been at CONTEXT for some time; it certainly pre-dates the current flurry of queries an the popular recommendation to highlight selected links in the "See also" section of an article. Tony (talk) 02:06, 22 October 2008 (UTC)

There's a difference between "don't link unless there's a reason" (presently here, and at WP:CONTEXT), "don't link unless there's a specific reason" (formerly here, added without consensus), and "don't link unless there's an overwhelming reason" (what you seem to be saying), and "don't link unless the article falls into a specific type in which date links are appropriate"" (what Lightbot is enforcing). — Arthur Rubin (talk) 02:14, 22 October 2008 (UTC)

• Are you asserting that "reason" by itself, unqualified, doesn't mean "useful" or "arguable" or "supportable" reason? The onus is still on an editor who wants to link a solitary year to demonstrate why it is useful in the context. This is hard to do, as suggested above when my repeated requests of certain complainants to provide examples of the solitary year-links or anniversary links at issue are met with silence. Methinks this is just a case of WP:IDONTLIKEIT, without having though through and properly weighed up the evidence. Tony (talk) 03:01, 22 October 2008 (UTC)

• As Tony suggested some number of threads above, an outstanding method (IMO) to provide historical trivia to readers is to provide a well-aliased link in the See also section. If there is an article on the War of 1812, then a link that is aliased to read “Notable events of 1812”, is a wonderful, least-intrusive way to let readers know of the availability of these lists without cluttering up the body text with more blue. Implicit here, is that these sort of aliased links in the See also section would be best limited to intrinsically historical articles. There’s no point providing a link in the See also section of Angela Lansbury so trivia on her birthdate (October 16) can be researched. If you actually read the Oct. 16  article, you can wade through *fascinating* bits like this…

“ 456 - Magister militum Ricimer defeats Emperor Avitus at Piacenza and becomes master of the western Roman Empire. ”

Really, if we’re going to be providing links in an article on a famous female actresses like Angela Lansbury, then keep the links germane. A rather germane link would be “Robo-biscuit actresses of the 21st century.” Greg L (talk) 03:58, 22 October 2008 (UTC)

• The interesting new thing about this section is that Derek wants to have dates stand out in the articles. I haven't seen that argument before; of course, for a person who wants this, it is a compelling one. But then, someone might want to have all personal names bolded, all negative numbers in red, etc., and I simply don't think these things are appropriate for article prose. At any rate, our links should be used for linking. Using them for something else (autoformatting) was a bad idea from the start and now that we can avoid it, let's. -- Jao (talk) 05:28, 22 October 2008 (UTC)

• thanks Jao - i agree that links should be used for linking, not for "highlighting" purposes. Sssoul (talk) 06:47, 22 October 2008 (UTC)

• To Tony above; for the years of birth and death in articles about a person, we've established a supportable reason. You seem to be saying it's not a reason, rather than it's not a good reason. — Arthur Rubin (talk) 14:28, 22 October 2008 (UTC)

Arthur: I know that you're keen to see biographical articles start with a splash of bright blue, but nowhere do I see evidence of this "supportable reason" you talk of. Are you sure this is not just wishful thinking? Tony (talk) 14:51, 22 October 2008 (UTC)

• We’re covering old ground and discussing unwise new concepts, like using color for emphasis. It’s not hard (unless you add the factor of “change on Wikipedia”, which is rarely easy). In‑text blue links should be strictly limited to links that are particularly relevant to the article and will be of likely interest to that readership. This isn’t rocket science here. Links to random trivia are just that: links to lists of random trivia. Honestly, I wouldn’t even link dates in our Trivia article. Too few readers ever reads more than five or six entries in these articles after their first encounter with them. As for using text color for emphasis, that’s easy. Unwise, IMO, but easy. Greg L (talk) 16:19, 22 October 2008 (UTC)

The heart of the matter: no strong demonstration of consensus

First off, I completely agree about delinking (I'd rather see a scheme done in a way we can recover autoformatting w/o linking, but that's not critical). However, when I read this page, the thing that stands out is that while the effort to remove date linking due to its horrendous implementation has been a process that has taken a couple years, it has been a slow, glacial change. However, this last change - and the most drastic, going from "data linking is depreciated" to "date linking is not to be used except in limited cases", which removes the concept completely - seems to have been pushed through by a handful of editors with minimal time for discussion at a broader level. I know Tony put messages on several top Wikiprojects to explain why date linking was being removed, but this is not the same as a centralized discussion or an RFC to determine, even with the strong body of reasons to get rid of it, if date linking should be abolished. The fact that editors are coming here and not finding such a consensus even with what are in the archives, is why there is still a point of contention - it feels like the discussion was all done behind closed doors and suddenly put into place with no chance for further rebuttal.

To me, the matter can be settled if someone would create an RFC to get it out there (specifically, the matter of removing date links completely), announcing it to the Village Pumps, WP:CENT, and if we can convince those that maintain is, watchlist-details. Run it for two weeks, and then say it's complete, archiving the page. Presumably, it will have support to remove dates, at which point, those naysayers will hopefully stop complaining now that consensus has shown to be there. This is an issue that effects every user of WP (including anons, even if for the better), so passing it as a consensus with as small an input that it seemed to have just feels wrong. But there might be significant resistance to that, and if that's the case, we then need to talk about it more. Right now, it is impossible to judge that from what we have presently, and that's my primary concern. --MASEM 16:33, 22 October 2008 (UTC)

• “The fact that editors are coming here and not finding such a consensus even with what are in the archives, is why there is still a point of contention - it feels like the discussion was all done behind closed doors and suddenly put into place with no chance for further rebuttal.” That will always be a problem on Wikipedia. There is a wide variety of volunteer, contributing editors. Most are unregistered and have no idea of the existence of special forums like this nor any interest in participating even if they did know. Even many registered editors are similarly inclined. It’s not at all surprising that editors wouldn’t have any knowledge of this decision to deprecate the linking and autoformatting of dates until they see their favorite articles affected by a bot.

  I don’t think the solution is yet more debate; debate would go on forever if we widened this and started all over fresh. And the conclusion would be the same: autoformatting, which can only be seen by registered editors and doesn’t benefit ordinary I.P. readers, would be deprecated. Linking of dates would also be deprecated since such links are rarely topical and germane.

  The end result of all that exercise would be an education for those who come in late to the discussion. So maybe what we need, is just that: a vehicle for education. Maybe we could have an essay explaining in detail, the rationale behind the decision. The trouble with that is, there are opponents of deprecation who would contest the “facts” in such an explanatory essay and, in the end, it wouldn’t be possible to write anything that was actually explanatory; it would just be a neutered re-hash of the nature of the debate.

  We’re just living the results of what is a pure democracy where there are no elected representatives. We rely instead on de-facto, volunteer leaders like Tony and others who care enough to stay engaged long enough to be familiar with all the details and persuade others to their point of view. It’s hard work and Tony has shown remarkable patience in the face of numerous instances of editors coming here complaining about how the linking of dates is such a splendid thing and no one told them about the change. In the end, we do have a general consensus. But a general consensus is not 100% of editors in full agreement. And it never was. Wikipedia would grind to a halt if every single editor who had a knee-jerk reaction to the deprecation of his or her linked dates could come here and make a politically-correct-sounding argument over how he wasn’t included in the decision so the decision is null and void.

  Now that I think about it, the best solution is to have two essays. One, written perhaps by Tony, could explain the reasoning behind the decision. If someone disagrees, they can write their own essay. It’s time to move on. The amount of debate that has gone on here over the years over just dates is nearly unimaginable. Greg L (talk) 17:02, 22 October 2008 (UTC)

Maybe Signpost can run an article about it? Also, I still think that the policy itself should be clarified. I don't think the casual reader would understand that it is saying "don't blue-link dates unless you have a specific reason why you want people to click on that link." I think that many editors continue to link dates for two reasons: 1) they like the way the blue-link LOOKS - that it gives prominence to the date. 2) newbies see other articles with dates blue-linked and think that this is the normal way to wikify articles. It takes some getting used to for those of us who have created many articles and carefully linked the dates to wrap our heads around such a big policy change, so I agree that publicity would be very helpful about the change, together with a clear explanation of WP:OVERLINK and the other reasons for the change. Happy editing! -- Ssilvers (talk) 17:07, 22 October 2008 (UTC)

If anyone wants to measure community acceptance, it is easy. Just go to Featured Articles, Good Articles etc and see for yourself. Those articles are, by definition the best that Wikipedia can offer. You can measure the ratio (actual date links)/(potential date links). I bet actual date links represent a tiny fraction of a percent of potential date links. Lightmouse (talk) 17:18, 22 October 2008 (UTC)

That's not really a valid method of gauging acceptance since you, Lightbot, Tony, and others seem to have already gone through and delinked dates in many featured articles. — Bellhalla (talk) 17:46, 22 October 2008 (UTC)

The way to solve that would simply be to examine the verisions at the time those articles passed FARLeadSongDog (talk) 22:25, 22 October 2008 (UTC)

Edits by bots should not count when trying to determine whether there is consensus. Tennis expert (talk) 10:12, 23 October 2008 (UTC)

• Good summary of the status quo and the nature of the problem, Ssilvers. I agree. Greg L (talk) 19:10, 22 October 2008 (UTC)

Lightmouse's suggestion assumes that FA and GA are good measures of community consensus and that most FA and GA reviews have considered this matter. Neither seems plausible. Septentrionalis PMAnderson 03:16, 23 October 2008 (UTC)

Given how recurrent the complaints about date unlinking are, it is obvious that there is no widespread consensus. I suggest running a wider RFC or poll, announced via the watchlist, community portal, etc. to try to get more people than the MOS regulars to voice their opinion. --Itub (talk) 08:52, 23 October 2008 (UTC)

• I'm not all that sure it's "obvious". What you might imply (and are you?) is that the consensus has changed. It took us two years to get this far, so I think it ain't all that likely to have changed in such a short time. The discussion here seems to be centred on some aspects of linking, but I don't see a wholescale revolt against the underlying principle of deprecating links. Anyway, how many people and how many discussions will it take to constitute a valid consensus? I believe what counts is that this was essentially a style matter, and it has reached a consensus here. In addition, the WP community is in a constant state of flux. We have to nail this thing at one point in time, and it appears to have been done. The horse is well and truly dead. Ohconfucius (talk) 09:28, 23 October 2008 (UTC)

Exactly right, Septentrionalis. Tennis expert (talk) 10:12, 23 October 2008 (UTC)

Indeed Ohconfucius. The "complaints" that Itub squeals about are coming from a tiny, loud minority of WPians who've probably been around for too long and are hard-wired against formatting improvements, because "they just don't like it". I'm sorry that they missed out on the long long debate here during 2007 and 2008, widely promulgated, at least towards the latter part.

What is telling is that when asked for examples of why the linking cancer should be allowed to persist, they are silent or offer (occasionally, if you're lucky) rather lame evidence.

What is also missing from their noisy complaining is the elephant in the living room: where was the original consensus for the linking frenzy? Please locate it for us (methinks it never existed).

It would be a breath of fresh air if detractors stopped in their tracks and thought of our readers: I hear of no readers, nor of WPians at large, lining up to bag the efforts of those who are bringing about these long-overdue improvements. Tony (talk) 09:49, 23 October 2008 (UTC)

• It's a filibuster. The conservative (used advisedly, I might add) opposition has observed a chink, and now want to exhume the putrid equine remains. Ohconfucius (talk) 10:58, 23 October 2008 (UTC)

If there ever were a "linking frenzy", that frenzy alone is evidence of consensus to have the links. You seem to be stuck on your personal dislike of date linking, in contravention of WP:IDONTLIKEIT, which you so conveniently and repetitiously cite to support your position. How ironic. Tennis expert (talk) 10:12, 23 October 2008 (UTC)

• Doesn't need to be a real frenzy, AFAICT. The guideline's been in existence for several years, and I know there are editors who refused to promote articles to WP:GA or WP:FA without rigid adherence to the linking (in the same way as there are those who insist on the use of citation templates). IT may also tend to support to [my] view that most editors are merely law-abiding citizens who have been obediently inserting links because a guideline says so, for the benefit of DA. Or perhaps someone wrote a bot to automatically turn dates blue for the sake of the red-green colour blind? Ohconfucius (talk) 10:58, 23 October 2008 (UTC)

"The tiny and loud minority": what a perfect description of the self-selecting MOS legislature! (Of which admittedly I'm part of by posting here.) I agree that the consensus in this page changed. But the consensus out there never did, as evidenced by the reaction that often occurs whenever someone uses a bot to massively unlink dates in articles. FWIW, I don't give a damn about date links and don't use them myself. But I am interested in other issues related with units and numbers and it is annoying to have this page overflowing with WP:LAME discussion about date links for so long. --Itub (talk) 12:15, 23 October 2008 (UTC)

• Wikipedia is not a democracy; it operates by consensus. It is not a form of government; the chief difference is that there are times when governments must act, and doing nothing is worse than either decision. In democratic governments, the majority has a right to rule (limited by other rights); here, decisions can be put off until the eve of our publication date. Therefore a filibuster is no evil; it is a sign that some temporary and local majority (all our polls are temporary and local) prefers their own way to the compromises required to get a minority to join the consensus. Septentrionalis PMAnderson 13:39, 23 October 2008 (UTC)

• I would say that although I remain deeply unhappy about the inconsistent application of international vs American date formats in that 'no man's land of the non-British-non-American article. However, most of us are prepared to live with the decision arrived at by consensus, for the sake of harmony. Note that that debate only took place over a few weeks, if not days. How can we be sure it was reached correctly, and over a correct length of debate?? Sure, it may be OK to filibuster a bit, but when the debate has continued for two years, it is already one helluva filibuster. The battle lines are entrenched, and there is not much to suggest that minds will be changed any more. Ohconfucius (talk) 14:22, 23 October 2008 (UTC)

• No, that's not fair. The debate went on until a (possibly temporary) consensus among the people monitoring this page at the time was arrived at, for about a week. Once the changes started being made according to that "consesnus", then the complaints followed, and have been following since. I think the "consensus" needs to be revisited. — Arthur Rubin (talk) 15:02, 23 October 2008 (UTC)

• This is precisely the issue. Let's all be fair here; the goings on at WT:MOSNUM are not exactly well broadcast throughout the community as a whole. To your average Wikipedia editor, the methods by which these guidelines were written up and decided upon are a fairly esoteric process - it is no surprise that the discontents often refer to the regulars here as "MOS-wonks" or something less flattering. Many editors aren't even aware of the existence of the Manual of Style as a whole, let alone the obscure subsection of a section dealing with autoformatted dates. Since it's "just a guideline" it isn't that important, but when attempts are made to enforce the guideline it gets people's attention. That's why RfC exists. No one would reasonably expect that an RfC would be filed for small minute matters, but when attempting to enforce something on a Wikipedia-wide level it requires Wikipedia-wide consensus. Various avenues have been suggested here (using RfC, signpost, some other form of catching the community's attention) but for some reason have yet to be undertaken. Having been made aware of this issue for some time, I don't doubt that this back-and-forth is going to continue ad infinitum until there is some coherent consensus that can be pointed out. Shereth 15:10, 23 October 2008 (UTC)

Arthur, you say in your edit summary "Quite. Agree with title"; do you agree with the opening sentence underneath that title, naturally by the same editor? "First off, I completely agree about delinking". Tony (talk) 15:12, 23 October 2008 (UTC)

Shereth, it's all very easy—specious, actually—to keep demanding more and more consensus just because you don't agree. Perhaps your strategy is to be dissatisfied until there's a complete referendum of all WPians who've ever set foot on the project; maybe even all visitors too? It's just a spin-way of trying to discredit a well-established consensus. Purusing this line, you could batter down any consensus that has ever been established on WP, and freeze policy and practice to your liking. People won't be fooled by this. Tony (talk) 15:17, 23 October 2008 (UTC)

You are misrepresenting my position, Tony. I have stated that I do not oppose the deprecation and subsequent removal of these links. It makes no difference to me one way or another what the outcome of this debate is. My primary complaint is that as an administrator I have dealt with (far too often, I may add) complaints that a bot or a user is removing wikilinks from an article with no discussion and that no consensus can be found for this kind of action. You are employing a straw-man argument here by talking about a referendum of all editors and visitors - this is patently absurd. My position here is simple : people want to see consensus on this matter. Why is it so hard a thing to ask, that the proponents of enforcing the guideline seek broad consensus to do so? Shereth 15:31, 23 October 2008 (UTC)

• I seem to recall Tony asking you for examples of complaints you have had to field. There are no visible signs on your talk page, so would you please be so kind as to point us in the right direction? Ohconfucius (talk) 08:47, 24 October 2008 (UTC)

• If Tony (or anyone) asked for some examples, I missed the request and apologize. This and this and this for some AN/I discussions relating to Lightbot's delinking of dates. There are older examples but I don't really have the time to dredge through the archives. Shereth 13:53, 24 October 2008 (UTC)

• Twice I've asked. Thank you for providing these links, Shereth. I wonder why you didn't pass them straight on to Lightmouse himself. One thing that distinguishes his bot-management is the polite, sensitive and responsive way in which he handles complaints (I've learnt something from it). The benefit of going straight to the horse's mouth in this case is that they provide feedback to Lightmouse, and he is in the best position to respond to them positively (see his talk-page including archives for much evidence). There was no need for you to become stressed at having to deal with them, and since you're on record more than once as saying you have nothing against the delinking of dates and date fragments, I see no problem. Tony (talk) 15:12, 24 October 2008 (UTC)

• Again, I apologize for having missed your previous requests; it was not an intentional oversight. To my knowledge, Lightmouse was informed of the discussions each time they have come up, and a quick overview of his talk page archives seems to confirm that. A quick perusal of those archives also serves to highlight just how often there are concerns regarding date delinking! To his credit, Lightmouse has been cooperative in terms of halting operation of the bot during this period of discussion. However, there has been some level of dissatisfaction with merely being pointed at MOSNUM (or MOS:SYL) as justification for the removal, because the guideline does not mandate a mass removal of links, just their deprecation. I understand that in the view of some, this is merely a matter of semantics; however it is a point of contention among a significant number of users and it really is prudent that a firm consensus is established one way or another that does not rely upon interpretation of semantics. That is really all I am asking for, a centralized discussion regarding how (if at all) to enforce the deprecation of certain types of date links. Shereth 15:29, 24 October 2008 (UTC)

That means we'll have to hold a consensus-gathering palahva for every single point in the style guides. It will take years. The community, as I've said, is firmly behind the deprecation, and it is notable how few people do complain aside from the loud, tiny minority here. Can you outline a major plan for consensus gathering for every point of the style guides? This is a new turn of events. Tony (talk) 06:56, 25 October 2008 (UTC)

• The examples given by Shereth are not on point, because they are largely about operation of Lightbot or AWB in ways that (1) obviously give wrong results, or (2) delink solitary years. This section is about whether there is consensus to delink (1) full dates or (2) a day-month combination (in either order). --Gerry Ashton (talk) 14:18, 24 October 2008 (UTC)

Tony1, you have a poor understanding of what "consensus" means. There isn't any sense in which there can be more consensus. It's not that kind of metric. In order for there to be a consensus, the points raised by all significant factions need to be taken into account and resolved to their satisfaction. That hasn't taken place with regards to the loss of useful metadata; therefore there is no consensus. It's that simple. If the participants to this discussion can figure out a way to preserve the metadata — either by halting delinking entirely (which I'd prefer) or by coming up with some alternative markup for dates that doesn't trigger autoformatting or produce wikilinks (which you'd presumably prefer, and I'd grudgingly accept) — then the delinking can proceed with a valid claim of consensus. You've already demonstrated your bad faith in this discussion by refusing to put your contentious activities (mass delinking of dates) on hold while we discuss this, so I don't expect you to give this point the consideration it deserves, but perhaps others will be more reasonable. And perhaps I'm wrong and simply caught you during a grumpy period of time, and you'll be more amenable to these points now. Either way, a solution that satisfies all the parties involved really isn't that hard to implement, if we just give it an honest effort. --UC_Bill (talk) 15:42, 23 October 2008 (UTC)

is the RfC regarding the proposal to link birth/death dates in the first lines of biographies still in progress or abandoned or ... ? did the people who wanted to keep those dates linked even read the "interim/compromise proposal" that was offered in hopes of satisfying both sides? only three people (one of whom didn't participate in the RfC) expressed opposition to that compromise proposal - does that mean the 32 other people in the RfC accept the compromise, or that they aren't really interested in reaching a consensus? hard to tell ... which also makes it hard to tell what purpose another RfC would serve ... Sssoul (talk) 17:43, 23 October 2008 (UTC)

• To UC Bill. It is a total absurdity where you write “…[in a consensus decision,] the points raised by all significant factions need to be taken into account and resolved to their satisfaction.” That’s pure über hogwash. A consensus on Wikipedia is not 100% of editors in full agreement and it never has been. If a consensus required that all editors have made peace with a decision, hardly a damn thing would get done around here. We’d still be using the IEC prefixes for computer memory (“256 mebibytes of memory”) if the standard was set as high as you like. That’s because, in the example of the IEC prefixes, there are still hold outs who will never ever agree to the consensus view. Do you think that because there are editors who will never be satisfied with the consensus view that any asinine practice should be grandfathered in? No change for the better can occur if there are holdouts? Preposterous.

  It should come as no surprise that once dates start being de-linked, certain editors who just love these blue links to mindless trivia that no one actually reads will come here with a “WTF” reaction and demand that the decision be reversed. By your standard, we would then have to reopen debate and work on getting these editors up to speed on why the deprecation of autoformatting of dates and de-linking was a wise decision.

  What we need here to put this to bed is an essay explaining the decision that we can point editors to. Then, the next time someone comes here saying, “But, I wasn’t included in this decision” so it wasn’t a good decision, and there was no proper consensus (‘cause I wasn’t included), and who died and made you boss(?), and I think links to mindless lists of random trivia is good and valuable and germane to the subject matter and is sound technical writing practice (although I never actually managed to *read* one of these abominations from start to finish myself)”, well, we can just point them to the essay explaining why the decision was made.

  Yes, there was a consensus decision. Indeed, 100% of the editors were not in agreement. But it was a consensus decision and we don’t have to re-open the debate each and every time someone new arrives here saying they don’t like it. Greg L (talk) 20:28, 23 October 2008 (UTC)

If "someone new" is raising a valid point (such as the loss of metadata) then yes, you do need to reopen discussion, and you should also demonstrate good faith by holding off on contentious behavior while discussing the matter. Also, I never mentioned anything about 100% of the participants needing to agree. I wrote that all significant factions need to have their points taken into account. What constitutes significance is a matter of opinion.

Perhaps I should have taken the same kind of hard-line, uncompromising approach that you have. Perhaps I should have started edit-warring with you and Tony (and whoever else) by reverting your date delinking. But I didn't, because I happen to think that doing so would just be making a bad situation worse.

I happen to like date links, I like date autoformatting, and in fact I'm in favor of what you'd call "overlinking" in general. I disagree that having more links is somehow detrimental. But I don't push those views, here or elsewhere, because I realize they're in the minority. It's also a simple matter to add more links if the majority opinion ever changes. My biggest complaint with your mass delinking of dates is that there's no easy way to undo it, should policy ever change (back) to favoring date linking, or should some other approach to auto-formatting be developed. You're destroying metadata, and that's simply unacceptable. There's a big difference between a date in an article body and a date in a quote, and without that metadata, there's no way to differentiate between them. It's also more difficult to identify dates when they're not marked up. There are a variety of reasons that metadata is a good thing, and by getting rid of it you're throwing the baby out with the bathwater. I know you hate date autoformatting and linked dates, and that's fine. I happen to disagree with you, but I'm willing to put my opinion aside as long as your actions are (in theory) easily reversed.

So let's just come up with a replacement markup for dates, and modify the script you're using so that it replaces wikilinked dates with dates that are marked up some other way. Then I'll be willing to call it a consensus, and will be glad to help explain the situation to others, and possibly even help out with the markup change. --UC_Bill (talk) 21:43, 23 October 2008 (UTC)

• UC Bill, if you’re going to backtrack on what you really mean, then just man-up and say so. But please don’t imply that I misread what you wrote, which was as follows:

Tony1, you have a poor understanding of what "consensus" means. There isn't any sense in which there can be more consensus. It's not that kind of metric. In order for there to be a consensus, the points raised by all significant factions need to be taken into account and resolved to their satisfaction.

Did you or did you not write that? So, just ‘pardon me all over the place’ for somehow thinking you were suggesting that a consensus isn’t a “shades of grey” issue where there can be varying degrees of it, and how a consensus requires that all points have to be resolved to all party’s satisfaction.

As to how you “happen to like date links,” you’d have a hell of a lot more credibility that you actually are making an informed stand on this issue if you accepted the Sewer cover challenge and earned your own Sewer Cover Barnstar. Then I’d at least have a better “theory of mind” regarding you and how you actually believe these trivia articles are “good reading.” Greg L (talk) 22:07, 23 October 2008 (UTC)

P.S. As to your love of autoformatting: this too, is an example of why we need an essay explaining the reasoning. But in a nutshell, autoformatting of dates can only be seen by A) registered edtiors, who B) have taken the time to adjust their user preferences. The results can’t be seen by regular I.P. users (the vast majority of our readership). In the best of cases, the *special* view the registered editors get masks inappropriate date formats (such as a default-European date in a clearly-American-related article). In the worst case, a date format like [[2005-07-06]] (which makes *pretty* dates like July 6, 2005 or 6 July 2005 for a very small fraction of our readership), produces nothing more than a crappy looking 2005-07-06 for everyone else. The decision to make autoformatting that benefited only some registered editors was brain-damaged from conception.

As for making autoformatting work for everyone, we’ve gone all over this before (and I’ll repeat it for you now): To make date autoformatting work for everyone, would require major tinkering under the hood of the servers by the developers to make Wikipedia’s servers look to the requesting readers’ I.P. addresses, figure out what country they live in, look to a database of what date format is customary in their country, and spoon-feed custom content to them. If you think you can get the developers to do this, more power to you. I’m not holding my breath for their response. Nor do I think all that effort is in any way necessary. IMO, all we need is Euro-style dates for most articles, and American-style dates if it’s an article closely associated with the US. End of story. The consensus view here however, was to throw in a healthy dose of “first major contributor” into the guideline. I didn’t agree with that consensus view, but I accept it. I wish you’d do the same regarding the deprecation of autoformatting and linking of dates. Greg L (talk) 22:31, 23 October 2008 (UTC)

Hey Greg, two things:

1. Notice the word "significant" in what Bill wrote (and ironically in what you quoted) — that word is.. you know.. significant here.
2. There's no need to resort to IP address lookups when the browsers themselves advertise what locale settings they want. For instance, my browser sends the following information (among other things) with each and every single request:

Accept-Language: en-US,en

Adding a check for that to the servers is a pretty simple change. So that's not really an issue at all. The issue is getting everyone to agree to what features they actually want, not with the technical part.

--Sapphic (talk) 00:40, 24 October 2008 (UTC)

• I saw the “significant” word. Lawering. It doesn’t significantly change the point of his message (“how a consensus is a black & white thing with no shades of grey and how pretty much everyone has to have complete buy-in and be satisfied”) does it?

  Regarding this news about the browser sending the info on what country someone hails from, I didn’t know that. I think it would be great if the developers produced parser functions for template writers like Lightmouse and Randmom832 to use. Frankly, I can think of better tasks to tackle with such parser functions than futz around with date formats. I read right over Euro and US-style date formats; neither bothers me in the least. I think there is way too much whining about the subject, honestly.

  What is more disruptive to my train of thought when I’m reading is when I see dialect-based spelling differences that don’t look right. My brain gives me a (!) interrupt as I read. For instance, if I encounter “you realise that the spell check on Safari just now put a dotted red line under ‘realise’ as I type this sentence in America with my browser”, I take notice. I’d love it if there was a template that used special new parser functions that looked to where the reader lived so one could—if they wished to—type {{dialect|US|UK|realize|realise}}. Then we could really spoon feed custom content to our readers. I could write The British people during WWII were noted for their {{dialect|US|UK|envious disdain of the overpaid, oversexed American soldiers|warmth and hospitality to the yanks}}.

  In advance, congratulations on your attempt at a Sewer Cover barnstar. Greg L (talk) 02:36, 24 October 2008 (UTC)

• I'm not claiming a 'Sewer Cover barnstar', but would like to mention that having looked at some of the date articles, I would describe the experience as no more than "hmmm... interesting". Of course, the events which happened on a given date are almost never related, but are the coincidences of history; some of the events which happened in a given year are related, but the relevant articles usually deal with linking the issues in a more than satisfactory manner. Isn't it ironic that we have a policy which discourages trivia sections, yet we have hundreds of articles of 100% pure trivia? I guess some people get bored reading serious articles here and look out for diversion! Ohconfucius (talk) 09:45, 24 October 2008 (UTC)

I can't speak for Bill as to the point of his message, but I agree that a consensus is a black & white thing, which I think was his point. Consensus isn't a matter of degrees, unless you want to count the breadth of a consensus (i.e. the number of people participating) as such a measure. A consensus is a position that is acceptable (though perhaps not preferable) to a group of people. ALL the people. The subjective part is deciding who should be in the group. Stubborn loners are typically excluded. Small minorities with valid complaints should not be excluded, no matter how small they are. A consensus is a goal on Wikipedia, not a requirement. It's an ideal. If a minority group has a valid point, that point must be addressed in any decision that could properly be labeled a consensus decision. Bill's point (which I agree with) is that delinking dates eliminates metadata. A valid fix is simple: replace the wikilink brackets with some other markup. Something as simple as <date> would work, since it has no meaning in either HTML or in Wiki coding. The reason that this satisfies the goal of consensus (at least on this point) is that the people opposed to delinking can be told "well, if majority opinion changes, the dates can trivially be re-linked by a bot." That's not the case if the dates have all markup stripped away, because then there'd be no automated way of restoring the links. Plus, adding markup leaves open the possibility of fixing the autoformatting feature in the future, and having it use the new markup (or having a bot change the markup as needed.) Removing the links entirely simply destroys information and makes everything harder. --Sapphic (talk) 03:11, 24 October 2008 (UTC)

Sorry to prick the bubble of your "statistics" flurry that was started up at some specially created page last month and aggressively used as an urgent reason for any DA removal to stop—some of which pressure I found offensive in its style and intent. As I kept asking at the time, what on earth is the point of those "statistics", and can they possibly be useful when there are already so many contaminating variables? I was told just wait three weeks, just wait, wait, wait, and we'll show you. ....... But you haven't shown us anything. I'm afraid I don't believe for one moment that leaving WP's articles littered with double square brackets is going to lead to the generation of useful "statistics"—you'd expect some evidence already, but we remain highly skeptical, half-grinning onlookers, confirmed in our initial beliefs that this is a non-issue. As for wanting to retain the ability to convert the double square brackets to some other form of DA in the future: who would want to do that, when DA has caused such a mess to be made of date formatting through concealing from editors what our readers see? And who would want to repeat all of those broken syntaxes and wrong choices in a new scheme? It's the ultimate flogging of a dead horse. Tony (talk) 05:16, 24 October 2008 (UTC)

• To UC_Bill: There is no loss of any data, meta or otherwise –it's just a bunch of square brackets which are being removed. Quite frankly, it takes no grey cells to insert such fixes as and when they are agreed upon, unlike the current need to manually (or script-based) de-linking to ddmmm or mmmdd. Just thinking about which way to go for each article because of the 'first significant editor' rule gives me blurry vision after de-linking more than a handful of articles. The rule has created all manner of inconsistencies between themes, categories, family of articles etc. The only consistency is within the article, which is already something, I guess. Ohconfucius (talk) 08:47, 24 October 2008 (UTC)

The problem with DA was the specific implementation of it when it was added to Mediawiki, not so much the concept. Overloading one function to perform another, as any OOP programmer will tell you, is a BAD thing, but at the time it was added it wasn't obvious, now very much so. A version of DA that allows for linking per instance basis and defaulting to a page-specific date format for all users when not otherwise set, and deal with the issues of BC/AD and the Gregonian calendar is not impossible, it just has to be explicitly developed; every point addressed by those wanting to ditch the current version of DA can be corrected, it is just that there seems to be an impass between the devs and main Wiki wants to get there. But the key point it is not impossible; there is a practical consideration that we can get a working DA in place that does its job and just its job. At that point, we can then separate the issues that we've been forced to connect here: how dates are presented to non-reg/non-pref users, and overlinking on dates. Thus, it makes complete sense to try to retain meta data of some nature such that if a corrected DA is ever presented appropriately, we can bot-transform all meta-stored dates to the new style. Arguably, a bot searching for "MMM DD, YYYY" or "DD MMM YYYY" can still search text via regex to find these, but this is a CPU-intensive process, and there are other types of dates (Month, Year, or Year alone, or date ranges, or dates lacking explicit BC/AD or calendar type notes) that without some type of meta-ness, it cannot pick up. There's a strategy we could do now, still ditching DA and overlinking, but keeping data around to make finding dates more friendlier to bots in the future. --MASEM 14:11, 24 October 2008 (UTC)

We're getting down a rabbit trail here; the merits or detractors for DAing are well and good to talk about, but that's not the primary issue I raise. My point, which others have followed on, is that regardless of the perfect merits of removing date links, is that there is no single point of discussion that has been brought to the full community's attention (limited, but not to the usual places one looks) for a change that affects nearly every single mainspace article, but a decision has been made by < 20 editors worth of input. That's causing other editors to come and complain about it when it is put into action. Normal course of action, as is it a content issue and not behavioral, is to open an RFC to get a wider discussion and determien the best path forward based on this; in this case "Should we avoid using DA?"

All that I'd like to see (again, I completely agree with DA delinking, I'm not arguing against it), is that to basically end further discussion is to put up an RFC announced broadly, and get a large swath of editor feedback. Presuming that most will agree with the various reasonings (which, yes, should be put to a WP page essay for ease of reference) the RFC will show the global consensus towards removal, and those that are against it will have to realize they are part of the minority and live with the change. Mind you, this may show strong resistance to the change as well, at which point we have to consider the fact that a new approach might be needed. I'm going off my 1.5 yrs of experience trying to get something at WP:FICT to a point of consensus: this RFC should be seen as the crowning moment of those that have been arguing for DA removal for the past two years: the reasoning is all sound and perfect and fits with policy. However, the final touch is to present to the community at large and say "are y'all ok with this?" making the process more open as necessary for a volunteer project. This is no ways meant to invalidate the work that those behind the removal of DA to date have done. --MASEM 14:29, 24 October 2008 (UTC)

Many thousands of WPians have not complained at the delinking of dates and date-fragments in their articles, and very few (proportionally) have. The number of enthusiastic approvals in just a four-week period numbers many more than those who contributed to the positive consensus here over months of debate. Want the link? I've put it on this page so often I'm sick of it. Tony (talk) 15:18, 24 October 2008 (UTC)

I don't debate this, but it can be argued to be inconclusive - maybe those thousands just don't care, or they don't know where to reply about the issue, and so forth. Again, we're talking about part of the process of any policy development on WP, following up local consensus with a large-scale RFC to validate the point, I cannot see any harm in doing this. --MASEM 15:58, 24 October 2008 (UTC)

When above I commented that I've come under unfair pressure that I found offensive, I wasn't joking. This was posted by UC_Bill five minutes ago at WikiMedia's Bugzilla:

Tony, you're an idiot who clearly doesn't understand the first thing about technology. You should just leave Wikipedia for good, and stop annoying people. At the very least you should drop yourself from replies on this ticket, since you've made it clear you have no interest whatsoever in a solution to the problem outlined here. Bypassing autoformatting is not the same as fixing it, so your asshole-ish actions of mass delinking aren't actually a "solution" at all. Go away.... I recommend we just completely ignore the MOS-nuts who have some sort of

vendetta against date links (I think everyone knows who those people are) ...

See what I mean ... Nice one, Bill. Tony (talk) 16:29, 24 October 2008 (UTC)

Not a nice one, Tony. On the issue of date linking I disagree with Bill and agree with you. On the issue of your behavior I agree with UC_Bill. In fact, if you had stopped to listen to Bill instead of soapboxing, you might have found that his patch does in fact do exactly what you want, and it does so without needing to strip articles of potentially valuable markup.

You are way out there in WP:POINT land. Its disruptive and destructive, also to yourself: Your feeling "unfair pressure" and "offended" is no less due to a response that you provoked to begin with (comment #156).

So go take a wikibreak or mosbreak or something, but for heaven's sake chill out! That is, if you intend to actually accomplish anything. At the moment all you succeed in doing is pissing everyone off, including those who in principle agree with you. -- Fullstop (talk) 21:27, 24 October 2008 (UTC)

RFC was issued

A Request for Comment was issued; see the announcement. --Gerry Ashton (talk) 16:14, 24 October 2008 (UTC)

That RFC dealt with date formatting and did not address the issue of link deprecation and subsequent de-linking thereof. Shereth 16:20, 24 October 2008 (UTC)

Suggestion

Rather than debate what is a consensus here, why not do it at wt:Consensus where it will have some lasting impact?LeadSongDog (talk) 15:47, 24 October 2008 (UTC)

There is no consensus to deprecate IEC prefixes

Twice now in the last week, a discussion that I started has been prematurely archived, without giving Headbomb a chance to answer the question I put to him. So let's forget the question and concentrate on the issue.

First, "What is an IEC prefix?" I hear you ask. Read this for a brief introduction and this for the case against the deprecation of their use on Wikipedia.

was there ever consensus on this?

There are several reasons to question that consensus was reached for the present deprecation of IEC prefixes:

• the concerns of the 3 minority editors (in the 7-3 vote for the present wording) were not taken into account. All three (Seraphimblade, Thudnerbird2, Woodstone) expressed concerns about exactly the same piece of text in a larger guideline. The piece of text they were concerned about was the said deprecation. The reason for the concern, at least on my part, was that only 2 months previously, 11 editors had expressed a view that use of IEC prefixes should not be deprecated by MOSNUM (to none against).
• I did not see a need to go over all of the reasons for an umpteenth time, as I could not believe anyone would have the temerity of ignoring such an overwhelming consensus against deprecation - I was wrong
• despite this concern, the views of the editors involved in the 11-0 vote (against the present wording) were not sought
• the discussion was held in an acrimonious atmosphere, in which any opposition to deprecation was met with a barrage of ridicule from Greg_L.[1] Some elected to stay away rather than participate in such a mockery of a debate.(under Evidence that editors stay away from MOSNUM due to disruptive behaviour); see also Omegatron's statement

is there consensus for it now?

• Three attempts at starting a discussion were shouted down [2][3][4]
• In the 3rd attempt, at least 5 editors (Jeh, Seraphimblade, Thunderbird2, Tom94022, Woodstone) argued against the present wording. Those who dared to support their view were met with further ridicule from Greg_L:

• • This MOSNUM talk page has officially been declared a “no whining zone”,
  • “I want my IEC prefixes! I want my IEC prefixes! We had them for three years and I want them baaaaack!. Well… tough and
  • Well, Anomie, do you feel better now after that little fourth-graders’ rant? ,
• a similar number argued for keeping the present wording. They were cheered on by Greg_L
  • Hallelujah, amen. Pass the collection plate and let’s get outa here.
• the need to resort to these tactics to prevent even a discussion about the text demonstrates the weakness of the case for keeping it
• the ridicule tactic is not new [5]

After those attacks I requested mediation. An offer of mediation was made by Doug and rejected by Greg_L.

And now, because I dare to question the claimed consensus, Greg_L portrays me as some kind of lunatic[6].

See also the theses of Quilbert and Omegatron on their personal spaces

The following WP Policy statements are relevant:

• Reasonable consensus-building: Consensus can only work among reasonable editors who make a good faith effort to work together in a civil manner.
• Forum shopping: It is very easy to create the appearance of a changing consensus simply by asking again and hoping that a different and more sympathetic group of people will discuss the issue. This, however, is a poor example of changing consensus, and is antithetical to the way that Wikipedia works. Wikipedia's decisions are not based on the number of people who showed up and voted a particular way on a particular day; they are based on a system of good reasons.

In other words, there is no reason to assign any more weight to the 7-3 vote than to the 11-0 vote before it. The dead horse that anti-IEC editors are so fond of quoting simply doesn't apply here, because there has never been a discussion that concluded in favour of deprecation that has not been dominated by abusive remarks from Greg_L. The result is that editors who wish to take part (like Omegatron and Quilbert) stay away from the discussion because they do not wish to be on the receiving end of such abuse. Thunderbird2 (talk) 18:11, 24 October 2008 (UTC)

Demonstrating there is consensus for the guideline text to deprecate IEC prefixes.

IMO there has never been a consensus to deprecate IEC Binray Prefixes and this subject needs to be discussed in this talk page. It is really improper for anyone to revert this section. Tom94022 (talk) 18:45, 24 October 2008 (UTC)

IMO there has never been a consensus to allow IEC binary prefixes (except in articles about the prefixes themselves, or if they're actually used in the source), as they're not used "in the real world". Apparently, even though IEEE accepted the standard, journal authors refused to use it. There's no "there" there.

In spite of the fact that I don't think Greg understands consensus, he was right in this instance. — Arthur Rubin (talk) 23:26, 24 October 2008 (UTC)

Thunderbird2 I see you have not actually tackled the real issue and instead have again misrepresented the actual situation and again have attempted to misrepresent other editors and have again attemtped to use ad hominem. Nothing you have posted above has any substantive valid argument, as I will now demonstrate. For example: You claim you asked a question that had not been answered [7] but that question is fallacious because you know full well there was discussion and that in that discussion you failed to answer questions put directly to you. The question you asked is the same as asking an equally fallacious question such as "Can you point to a discussion proving the moon is not made of cheese?". The "question" you put to him is irrelevant and the answer already known and as such your demand that he answer it is not a valid demand and is wasting the time of Headbomb. You then go on to try to cite really old votes on a tiny issue but there is a much newer larger debate that refutes the older votes you cite because the newer much larger debate provides much stronger arguments than just the vote you cite. Strong arguments make consensus, votes are not strong arguments, and because you keep on repeating this accusation this demonstrates you are refusing to get the point WP:POINT. Then you again repeat the allegation that concerns of some editors were not taken into account, again as the talk archive shows the concerns of the editors were heard but when refuted by much stronger arguments there were no strong arguments in reply (you Thunderbird2 actually refused to give any valid answers many times, this is documented at the end of the talk archive). In the talk archive I can point to at least two key questions asked by two different editors directly to Thunderbird2 where no valid answer relevant to the topic was given. Since those opposing the text made no strong substantive arguments compared to the stronger substantive arguments for the text then the much weaker point of view does not have to be included in the guideline. It is obvious why unsupported weak points of view are not included in guidelines because guidelines need to be made from strong arguments, otherwise we would have guidelines saying "in all articles about the Earth it must be stated that some people believe the Earth to be flat" for example. Read WP:UNDUE because it applies to this situation. Again, repeating false allegations is a violation of WP:POINT. Then you attempt to misrepresent other editors because instead of discussing their arguments you personally attack them instead, therefore you are trying to use ad hominem instead of valid debate. Your actions are a violation of WP:NPA. Also this is not the first time you have posted the same refuted old arguments yet you still continue to repeat them and that is called "beating a dead horse". Since you repeatedly continue to violate policies and guidelines intended to promte valid debate on Wikipedia then this demonstrates disruptive editing. The consensus as demonstrated in the talk archive is that IEC prefixes are not to be used except in very limited situations and that consensus is reflected in the guideline text.Fnagaton 02:13, 25 October 2008 (UTC)

Enough with the tug-of-war

I don't want to see this block of text be inserted and reverted and inserted and reverted anymore. Let it go. It will be automatically archived by the bot, normally, if the interest in it dies down; there is no hurry to archive it. This is getting dreadfully tiresome and close to warranting protection or issuing blocks. I doubt anyone wants to see that happen. If someone wants to keep beating this dead horse into a pulp, let them. Ignore it, don't get involved in an edit war over it. Shereth 22:56, 24 October 2008 (UTC)

Thunderbird2's comments were archived to remove the violations of WP:NPA from the talk page. Since the user is again repeating those same false allegations then the comments should again be archived and removed to avoid cluttering up this talk page with irrelevant comments.Fnagaton 02:18, 25 October 2008 (UTC)

• (*sigh*) …since it’s a simple copy/paste… here is the consensus:

Figure of Merit—Binary prefixes (Purplebox)

Degree of support

User	5	4	3	2	1	0
[[::User:Headbomb|Headbomb]] ([[::User talk:Headbomb|ταλκ]] · [[::Special:Contributions/Headbomb|κοντριβς]]) 05:30, 2 June 2008 (UTC)		X
Greg L (talk) 15:16, 30 May 2008 (UTC)		X
Fnagaton 19:08, 25 May 2008 (UTC)	X
Woodstone (talk) 20:42, 2 June 2008 (UTC)				X
SWTPC6800 (talk) 18:01, 30 May 2008 (UTC)		X
Seraphimblade Talk to me 05:42, 26 May 2008 (UTC)				X
MJCdetroit 19:50, 27 May 2008 (UTC)		X
Thunderbird2 (talk) 07:28, 6 June 2008 (UTC)					X
Dfmclean 19:00, 28 May 2008		X
Pyrotec 22:35 05 June 2008		X

The above was after three solid months of debate. No rule of conduct in a decent and civilized society requires that a single holdout can keep on disrupting a system for so long. T-bird: your objections were heard but your persistent silence, when Headbomb asked you (repeatedly) to explain your reasoning, was deafening. You have no one to blame but yourself for failing to persuade others to your way of thinking. As I stated above, we are done with this issue for now. When there is a change in the reality of the situation and there is actually a fair amount of real-world usage of the IEC prefixes, let us know. Until then, please accept with grace that the consensus is that Wikipedia will communicate to its readership the same way all other encyclopedias and computer magazines do: with terminology and symbols that readers actually recognize. Greg L (talk) 02:47, 25 October 2008 (UTC)

• Nothing more can be said to you, Thunderbird2, that hasn’t already been said ten to a hundred times. You arguments and tactics remind me of a stuck record. Your views were heard by all and rejected as unwise. As for “abusive” remarks, and your implication that such alleged abuse somehow undermined the validity of the entire proceedings: nice try, but you obviously stayed in the thick of it to the bitter end (note your above vote)—even in the face of your perceived slings and arrows of outrageous misfortune. You have no one to blame but yourself for failing to persuade others to your way of thinking.

  Actually, you had an insurmountable objective: arguing a case to make Wikipedia do something foolish and use terminology that no other encyclopedia in the world uses, nor any computer magazine directed to a general-interest audience, nor which any computer manufacturer uses in marketing communications to their customer bases. Further, you were advocating Wikipedia use terminology that you conceded our readership didn’t even recognize (the fifth entry down is your signature). Further, your silence here for six days on this thread and then, after I archived it, your deciding to drag it back here to keep on flogging this dead horse, is just more of the same old stuff from you. Your actions here are tedious at best, and disruptive at the worst, and I will no longer dignify your tactics with any further responses. Goodbye. Forever. Greg L (talk) 23:26, 21 October 2008 (UTC)

Something to take under advisement

This page in a nutshell:

If you are the only person left beating the horse after it has died, consider the possibility that you should stop.

There comes a point in every debate on Wikipedia where the debate itself has come to a natural end. You may have won the debate, you may have lost the debate, or you may have found yourself in an honourable draw. At this point you should drop the stick and back slowly away from the horse carcass.

If a debate, discussion, or general exchange of views has come to a natural end through one party having "won" or (more likely) the community having lost interest in the entire thing, then no matter which side you were on, you should walk away.

If you don't, if you continue to flog the poor old debate, if you try to reopen it, if you continually refer to old news, if you parade your triumph in the faces of others... you're not really winning friends and influencing people. Instead, you are annoying the hell out of everyone nearby.

If you've "won": good for you. Now go about your business, don't keep reminding us of the fact that your "opponent" didn't "win". If you've "lost": sorry, hard luck. Now go about your business, don't keep reminding us of the fact that your "opponent" didn't actually win because of... whatever. If the debate died a natural death: let it remain dead. It is over, let it go. Nobody cared except you. Hard to stomach, but you're going to have to live with it.

So, the next time you find yourself with the body of a horse: please stop beating it. It won't help.

See also

• Don't be a dick
• Wikipedia:Don't edit war over the colour of templates
• Wikipedia:Use common sense
• Wikipedia:Do not disrupt Wikipedia to illustrate a point
• Wikipedia:No climbing the Reichstag dressed as Spider-Man
• Wikipedia:How many legs does a horse have?
• Wikipedia:Lamest edit wars
• Wikipedia:The Last Word

Greg L (talk) 02:47, 25 October 2008 (UTC))

IEC prefixes were permitted (and even encouraged) between July 2005 and April 2008

• Arthur Rubin states IMO there has never been a consensus to allow IEC binary prefixes (except in articles about the prefixes themselves, or if they're actually used in the source), as they're not used "in the real world". This is simply incorrect. Binary prefixes have been accepted by MOSNUM at least since July 2005 (and even recommended for some of that time). The present dispute began in January 2008 when Fnagaton removed the statement that they were permitted without first establishing that there was consensus to do so. In April 2008 Greg_L removed the text that he and Fnagaton were disputing (twice). [8] [9].
• Greg_L presents his tired table of votes for the umpteenth time, which proves that 3 editors voted against the present text, forgetting to mention that the 3 votes were all for exactly the same reason: that it has been established after lengthy discussion that there is no consensus for the present deprecation.
• Fnagaton protests weakly that I have not tackled the real issue. But he is wrong, for from the outset I present a detailed case against deprecation, prepared by Tom94022 and myself. In case he missed it the first time, he can read it again here,

Thunderbird2 (talk) 19:07, 29 October 2008 (UTC)

This has been answered about twenty millions times by now and your lack of tackling the real issue has been well documented. I've asked you well over 20 times to give substantial arguments over 3 months and you've failed to do so every time [See ([10], [11], [12], [13], [14], [15], [16], and on Headbomb's page [17], [18]])]. I've berated Greg for encouraging bad faith when it came to dealing with you ([19]) , saying that I would rather form my own opinions on this. Greg took numerous swings at me and at my supposed agenda for the promotion of the IEC prefixes. I'm a personal proponent of IEC prefixes in the real world, I use them and I love them. The fact that I side with Greg and Fnag (and Pyrotec, and Marty Goldberg, and SWTPC6800, and MJCdetroit, and Franci Schonken, and Jimp, and Rilak, and Dfmclean ...) on this is a testament to both the weakness of your position and arguments and the strength of theirs. Please drop the stick. You are a single-purpose account who spends >95% of his time pushing for binary prefixes. Go away. Headbomb {^ταλκ_{κοντριβς} – WP Physics} 20:36, 29 October 2008 (UTC)

People should stop opposing delinking. It is minor and inconsequential.

People should stop opposing delinking. It is minor and inconsequential. Lets move on. Lightmouse (talk) 16:16, 25 October 2008 (UTC)

• Indeed. Embracing just a 10% dose of Wikipedia:Don't-give-a-f*ckism helps me better deal with the goings-on here. That article is a reminder to observe the two golden rules of life: 1) Don’t sweat the small shit. 2) Everything is small shit. Greg L (talk) 18:09, 25 October 2008 (UTC)

Says you. I won't go in to why (there's already plenty of that elsewhere in the archives here and on this very page), but I disagree completely with date unlinking like this. To whomever had this idea: it was terrible and I fail to see the point in it. —Locke Cole • t • c 19:27, 25 October 2008 (UTC)

• Because sound technical writing practices dictates that links should be topical and germane to the article they’re in. They should anticipate what the readership visiting any given article would be interested in further exploring. The date articles that are linked to are nothing more than lists of randomly chosen trivia that rarely has a thing to do with what subject matter of an article. Too few readers actually bother to read them. If you don’t yet understand this point, try reading these dozen trivia articles, top to bottom, with no skimming: January 1, January 2, July 15, July 16, September 22, September 23, 1925, 1933, 1955, 1965, and 1987. No, seriously. Read all twelve. All of them. Tough, ain’t it? Come on… you didn’t read them. Go ahead… try.

  Furthermore, there were some stupid autoformatting going on where only registered editors could see *pretty* dates like 6 July 2005—the vast  majority of our readership (I.P. users) would see the raw date formats such as 2005-07-06. In other words, it was unwise to have done autoformatting and its associated linking in the first place. Greg L (talk) 20:14, 25 October 2008 (UTC)

  So instead of consistently formatted dates, now we have articles with a mish-mash of date formats (often times in the same paragraph and sometimes even the same sentence). How is this an improvement? Wouldn't it have made more sense to get the date format for IP users changed to something everyone could agree upon? Or better yet, if linking is really your primary concern, come up with a different syntax for date handling that allows dates to be autoformatted and yet remain unlinked. Instead a few people here have seemingly decided their way is the best way and made this a crusade. I really hope this reaches the ArbCom so those involved can be told that making wide sweeping changes without vast community consensus (especially changes which will be visible to our readership) should require the highest involvement of the community. —Locke Cole • t • c 23:15, 25 October 2008 (UTC)

People here have tried for years to get the developers to implement non-linking date autoformatting, to no avail. The devs are either uninterested, or actively resistant. It is not going to happen. Setting a default date format for users who haven't set the preference would fix half of the problem, but would still leave ugly and stupid links scattered through all our articles. Better to rip out date links for now. If unlinked date formatting with a suitable default is implemented later, it will have a new syntax anyway so there is little to be gained by keeping the current broken date links in the meantime.--Srleffler (talk) 05:19, 26 October 2008 (UTC)

Um, no. Nothing that impacts large numbers of articles on Wikipedia is “minor and inconsequential” – least of all when it lacks consensus. The simple fact is that date autoformatting was deprecated, and with little opposition once the reasons for it were explained. Since DA is embedded in the link coding and editors cannot remedy that, the only way to remove DAs is to delink dates. The problem is that there has never been any consensus for “de-linking on sight” (or by bot) of linked dates in general, much less “years in XXXX”.

“Deprecation”, by the way, means “no longer using, but leaving the extant coding in place” – not “removal with extreme prejudice”. I hate to say, “I told you so,” but as I pointed out in August, mass removal (by whatever means) of existing DAs would prove highly disruptive. Those who preferred to be rid of them ASAP wikiwide have pursued their agenda and reaped their due reward, so I have no sympathy for them regarding the tremendous amount of time they’re having to waste defending their position. I will note, though, that I proposed in that posting a far less disruptive approach, which remains a practical option, although it won’t achieve the “ASAP” removal that some here so dearly desire.

As for the “historical years”, again, there is no consensus for their removal at all; this isn’t a function of DA. Their removal is due entirely to the prejudice against this type of linking by some of the same folks, who feel that the product is unencyclopedic and useless. Whether one agrees or disagrees with this position, it is the sole reason for their being delinked at this time. The solution to the uproar over delinking DAs – and, separately, historical year linking – is, of course, to find a consensual path forward … not telling others to “just get over it”. After all, without consensus-seeking, it’s just as fair to proclaim, “People should stop delinking existing DAs. It is minor and inconsequential.”. Askari Mark (Talk) 20:31, 25 October 2008 (UTC)

• Well said Askari Mark. So what do we do to clearly identify a consensus? If there was ever anything that is the subject of dispute, it is when one tries to identify when a proper and true “consensus” has been arrived at. I guarantee you, that some editor who is willing to don orange robes and set himself alight over this issue, will set the bar much higher than does the middle-of-the-road editors.

  Trying to settle upon any course of action on Wikipedia is damned tough because it has no elected representatives. Further, even if 50 editors weighed in here, any decision that lets a bot loose is guaranteed to result in editors coming here wondering why they weren’t consulted on the matter. As a result of this phenomenon, all decision making on Wikipedia tends to be neutered down. For example, it took three years to get Wikipedia to stop being all alone in its use of “256 kibibytes” of memory when everyone else inhabiting this pale blue dot uses and understands “256 kilobytes”. Three years. That was a Wikipedia:Use common sense sorta issue but change doesn’t come about easily here. Maybe it sometimes takes being WP:BOLD to get anything done. I truly don’t know what is the best tact to take here. Greg L (talk) 21:01, 25 October 2008 (UTC)

  IMHO, a change like this is worthy of notice in Mediawiki:Sitenotice for at least a week (preferably a month). You need lots of input, not just MoS regulars, for something like this. —Locke Cole • t • c 23:19, 25 October 2008 (UTC)

• Locke Cole, in response to your 23:15, 25 October post: Yes, it would have been better to have a simple set of rules governing the choice of date format to use in articles. I had proposed a simple one: Use Euro-style dates (6 September 2008) for most articles but use U.S.-style dates if the article is about or is closely associated with the U.S. (and a handful of U.S. territories that would have been listed). But I lost.

  As for a date format that works for I.P. users, we don’t have those parser functions available for use in templates. At least not yet.

  Note that I am an American. But I am also an R&D engineer and I have noooo problem looking at dates in either format. I’m all about writing to produce the least confusion for the greatest portion of our readership. So an article on Chicago should say “January 20, 1985” (which it does) whereas an article on International System of Units should say “1 August 1793” (which it doesn’t;  it uses American-style formatting and they’re all linked!). Don’t blame me. Greg L (talk) 00:13, 26 October 2008 (UTC)

  P.S. I have my date preferences turned to "none". Every editor should set theirs the same way, otherwise, we aren’t seeing what I.P. users see. Greg L (talk) 00:16, 26 October 2008 (UTC)

• What makes me gobsmacked is why the order of month–day or day–month should matter to anyone. In all English-speaking countries, both are used (the US military, for example, uses day–month, Canada uses both, many newspapers elsewhere use US format despite the predominant use of international). Why do you care so much about it when you readily accept "colour/color" and "travelling/traveling"? It's a mystery. Those who moan on and on about the dispensing with date autoformatting seem to believe that it's quite OK for our readers to be exposed to both types of this binary system (just as for what is essentially our binary spelling system). Please take the emotion out and give that in-house indulgence a rest. We have much more important business to get on with. I ask all responsible editors here to assist in the task of removing date autoformatting. As for the linking of trivial date fragments, far better solutions to the provision of gateways into year pages et al. have been provided on this page. They should be embraced and the year-page crowd should spend its efforts improving woefully bad year-pages, not shoring up millions of unused pathways to those pages. Tony (talk) 01:26, 26 October 2008 (UTC)

• 'In all English-speaking countries' except England! Nimbus (Cumulus nimbus floats by) 01:47, 26 October 2008 (UTC)

• Do you read English newspapers? Um ... please have a look. Tony (talk) 01:54, 26 October 2008 (UTC)

• No, I don't read the papers, in most cases they are a waste of wood but I will take a special trip to the newsagent to see. If I want to know the date I go to the clock in the bottom right of my screen where Microsoft tells me it is currently 26 October 2008. I guess their 'regional' setting option is there for a reason. Nimbus (Cumulus nimbus floats by) 02:12, 26 October 2008 (UTC)

• Actually only irresponsible editors would continue to delink dates while there's an obvious lack of consensus. Such editors should be taken to task for their behavior IMO, up to and including blocking for obvious and continued disruption and violations of WP:POINT. Continuing to remove date links only makes the situation all the more urgent for those opposed to such delinking and creates entirely unnecessary drama (especially considering that such date links are, essentially, harmless; the wiki has somehow managed to avoid collapse while having date links all these years, why the sudden urgency to remove them as fast as possible?). —Locke Cole • t • c 09:02, 26 October 2008 (UTC)

• And your sideline sniping at people who actively work to improve WP is not irresponsible? Let me think about that for a moment. You should be ashamed of yourself, calling for the blocking of those who shoulder the work (unlike you); your view of the situation appears to be based on utter ignorance of the long long debate that has gone on here and ended in consensus in August. You're a Johnny-come-lately, perhaps, who's upset at having missed the debate—sorry, but that's just too bad. Maybe you'd like to bone up on the four information and consensus pages linked to at the bottom of this page. Tony (talk) 00:42, 27 October 2008 (UTC)

[Unindent] Greg, I believe we do have consensus on deprecating DA – or perhaps more accurately, further use of DA (not its removal). Rarely have I seen as many editors come to MOSNUM (or any other policy/guideline talk page) originally in opposition and change their mind once the issue was clearly explained to them. Normally, opinions are locked in and loaded, the trenches grow fuller, and the bombardments more intensive. It’s quite understandable that editors would like to be able to read articles in their preferred format, but the argument that it leaves a mess for non-editors – those whom we write here for – was pretty cinching. I think most editors who were involved in the debates appreciate (or came to appreciate) that it would be preferable for something to be written by the developers to enable this without the use of links. There were several threads posted about this issue in the VPP and I know the members of the more active WPs were aware of it, so I think it’s fair to say that there is consensus for the deprecation of DA.

The problem now being encountered is the one I foretold would happen. First, DA is intimately tied to wikilinking. It’s all well and good to say wikilink only where necessary and useful, but there’s a broad range of opinion on what “necessary and useful” really means; wanton – and particularly semi-/fully automatic removal of links – tramples on these viewpoints indiscriminately and most particularly affronts those who didn’t participate in the discussions because, well, they rarely or never do until something impacts them directly. Each one then demands an accounting and explanation/cajoling/convincing, which even the most ardent advocate of delinking dates has no doubt wearied of by now. In effect, these “delinking activists” have decided to “be bold”, forgetting that that approach works best on the small scale of an individual article, but not on the larger scale of a wikicommunity.

To get past this brouhaha, I think what should be done is something along the lines of what I proposed in August. To wit, pursue the broadest consensus on what to do about delinking DAs; I think Locke Cole’s suggestion that it be advertised via Mediawiki:Sitenotice is excellent advice since it wraps it all up in one intense episode and then we can all get back to using a larger portion of our wikitime in constructive editing. (However, here I would insist that “year in XXXX” be explicitly excluded, for the reasons I outlined to Nimbus above.) This proposal should first describe the consensus for deprecating continued use of DAs and why it was justified (telling the story one more time instead of hundreds of times), and offer two options for community comment approval: a) active removal of existing date links by manual, semi- or fully automated means; or b) permit the editors to decide on the fate of these links on the articles they are active in on the respective talk pages. (Of course, if one of these editors wants to be bold on that particular article, they can, but if protested must then abide by the talk page result.) Is this latter option quick and efficient? No, not at all. But what it does take into account is the variable range of tolerance for change among people in general, and the distaste by the majority for diktats by “outside experts” that meddle in their workspace. I think you’ll find that most of these “involved editors” will opt for removal, albeit not as fast as others would like; the advantage is that these editors will then “own” the decision as much as those who have already wrestled it into existence here at MOSNUM.

As for Tony’s wistful cry for a standard date format, … something about “If wishes were horses…” goes here. Choosing one over the other means selecting a winner and a loser … and, naturally, as in locking pages being edit-warred over, the winner will be the “wrong one”. As for myself, I’d settle for all the formats being consistent within each article. That would be major progress over what we have now. Askari Mark (Talk) 03:41, 26 October 2008 (UTC)

Which "Tony" is this? I think you'd better strike out your accusation, which is in total opposition to my view. "wanton – and particularly semi-/fully automatic removal of links – tramples on these viewpoints indiscriminately and most particularly affronts those who didn’t participate in the discussions"—you make it sound like having lewd, animalistic sex on a busy street; but this is more like hard work, requiring the exercise of fine judgements in some cases as to which format is most appropriate. The simple rebuttal is that from many thousands of date audits, only a handle of editors have complained, and most of them have been easily won over when it's explained. There's one of those on my talk page from a few days ago. You appear to be overstating the opposition, which, to my eyes, is coming mostly from IT-trained editors, who tend not to balance functionality over disadvantage in the way that normal people do. That accounts for the concentration of loud complainants and the silent agreement/support by the vast majority of users (let alone readers). Tony (talk) 08:58, 26 October 2008 (UTC)

Tony, I’m sorry you mistakenly took my “wistful” note as an insult; actually, I was agreeing with you. I’m completely comfortable with the international format and have always preferred it. However, on Wikipedia, it’s not what I want – or what you want – but rather what the community wants. Askari Mark (Talk) 22:11, 26 October 2008 (UTC)

Hear hear. If more participants here took such a reasonable approach, I would be more willing to discuss the issue. What I will note is that a blue-linked date doesn't really detract from the article, since there is no chance of the reader being fooled into thinking that the link is relevant to anything but the date itself. I'd also note that article-consistent date formats are double-plus-good, regardless of the status of DA itself, and year linking should be subject to editorial discretion, especially for historical years. Franamax (talk) 04:21, 26 October 2008 (UTC)

Franamax, if you'd engaged with the long debates about the matter, you'd be better informed as to the disadvantages of DA. Tony (talk) 08:58, 26 October 2008 (UTC)

• Thank you Askari Mark, for your careful analysis of the current situation we find ourselves in. I think your above observations best sum up the views “middle-of-the-road” editors who have seen all the evidence. Maybe you and Tony can suggest at a way to move forward. Greg L (talk) 04:55, 26 October 2008 (UTC)

• I think Askari already provided a reasonable suggestion of the way forward: advertise on the site notice before continuing with date delinking. Get widespread community support, then go to work on this (massive) undertaking. Specifically, we should decide on a few options, create a poll, and advertise that on the site notice (or alternately the watchlist notice). I see the options as really being: 1) Deprecate linked dates (with the understanding that removal will be gradual, done during the course of regular editing, not via bots). 2) Eliminate/forbid linked dates (this is (1), but with approval for bots to remove the links immediately). 3) Deprecate linked dates and request that the developers provide a method for formatting dates using some alternate wiki markup/syntax (presumably if this got enough support from the community the devs might feel compelled to make this a priority). 4) Maintain the original status quo, allow linked dates, etc. If there's additional options, feel free to list them (or your modifications). Once we have a list of options we can move forward with a wider community poll/discussion. —Locke Cole • t • c 09:02, 26 October 2008 (UTC)

• A key thing to remember is that there is a trick or two that we can link to full dates without invoking date autoformatting - it would require a bot assistance. Just as the DA implemenetation by mediawiki is convolutes linking and date autoformatting, the arguments on this issue are also convoluted, and sometimes the fact that we can separate the two issues (the dissadvangates of non-reg/non-pref users for DA, and the overlinking due to dates) needs to be stated as well as another option. --MASEM 09:12, 26 October 2008 (UTC)

Um, the reluctance of people to go down that path just might have to do with their feeling that the issue has been discussed to death on this page over the past two years, and resolved. You're flogging a dead horse. Tony (talk) 09:09, 26 October 2008 (UTC)

I don't see anyone else who read the discussion saying it was resolved. And I don't see an explicit consensus, except for date autoformating. — Arthur Rubin (talk) 09:46, 26 October 2008 (UTC)

Date delinking is by definition, minor and inconsequential, it makes no sense to object to such edits. Lightmouse (talk) 10:28, 26 October 2008 (UTC)

OK, I'll bite. If date delinking is, indeed, "minor and inconsequential", it makes no sense to object to such links. — Bellhalla (talk) 12:49, 26 October 2008 (UTC)

er ... Lightmouse is quoting an administrator who recently berated him/her for making "extremely minor, inconsequential edits" (ie date-unlinking edits). Sssoul (talk) 12:56, 26 October 2008 (UTC)

Yup, and it seems like an abuse of admin tools, frankly. The reasons provided are thin and contradictory, and the users involved have seemed to pop up out of nowhere. It's really odd. Tony (talk) 15:02, 26 October 2008 (UTC)

I don't know why some editors who are still discussing this issue, and doing so as if it were a major step back. This debate is not only going round in circles, but I object to editors who are starting 'name-calling'. Specifically, I take issue with "only irresponsible editors would continue to delink dates while there's an obvious lack of consensus. Such editors should be taken to task" – are delinking editors really being irresponsible or just being bold? I defy anyone to to defend why the tons of date links in each article in this category should not be de-linked forthwith. It would be blindingly obvious to all, except those not possessed of it among us, that it's just applying a tiny dose of common sense. Before DA was deprecated, the MOS already said that dates should be linked to only if it enhances the understanding of the article. So even before deprecation, there was little reason for these links to exist. Now that DA nonsense is gone, I say good riddance to these links. Ohconfucius (talk) 14:18, 26 October 2008 (UTC)

Yes, the following editors say that date delinking is minor and inconsequential

• Martinp23
• Tombomp
• xeno

Lightmouse (talk) 17:08, 26 October 2008 (UTC)

I don't really have a solid position on whether or not it should be done - what I know is that delinking dates is minor and inconsequential and should not be the sole purpose of your AWB edits. Don't conflate an AWB misuse issue with a MOSNUM debate. The shaky consensus of delinking dates is another matter altogether, and a can of worms I have no desire to eat from. –xeno (talk) 17:58, 26 October 2008 (UTC)

• Agree, by Greg L. Minor an inconsequential. I think (my opinion; don’t shoot me) that Wikipedians’ reaction has more to do with the perceived insult of “trespassing” where uninvited. De-linking improves Wikipedia for regular I.P. readers and increases the value of the remaining links. Greg L (talk) 18:01, 26 October 2008 (UTC)

  Not really.

  • Delinking makes very little difference to IP readers, since they don't run into the problems that autoformatting produces; it would be a relatively minor benefit to them to make date formats uniform within an article, but this can be done without delinking, and delinking will not produce uniformity - it will merely expose inconsistency.
  • One thing that this long discussion has made clear is that there is no consensus on which date links have value. Tony thinks none do; I value a minority; some editors think almost every year link has value as context. In the absence of consensus, we should, as often, leave well enough alone.

  How about it? Septentrionalis PMAnderson 18:15, 26 October 2008 (UTC)

Lightmouse should stop insisting on delinking. It is minor and inconsequential.

Delinking dates is not the sole purpose of my edits. Lets all be clear, linking of dates was implemented without consensus and was a big mistake. It is only now that delinking is accepted that we are cleaning up the mess on Wikipedia. Not one pro-linking editor has told me that they run a bot to clean up the many easily fixable but common date errors such as [[December 12|12 December]] and [[12 December]] [[1995 in aviation]]. Not one. All talk about how great date linking is, no action to make it work. Lightmouse (talk) 18:22, 26 October 2008 (UTC)

We are not interested in motive; we are not ArbCom. But this is an unfortunate combination of falsehoods: linking of dates and autoformatting was indeed accepted by consensus as a technical fix to a behavioral problem. It was, I agree, the wrong solution; it would have been better to have treated it as we have since treated spelling differences: by agreeing to differ.

The road forward now, however, is not another technical fix, imposed while ignoring the functions of linking in individual articles. It is a behavioral solution: again, leave well enough alone, and talk to editors on talk pages, and see if you can persuade them to delink. We have the arguments for that; if WP:Autoformatting does not persuade editors to delink, we should again agree to differ.

It is bad practice to treat everything as a nail because you happen to have a hammer. It is the same bad practice to run a bot because you can. Septentrionalis PMAnderson 18:34, 26 October 2008 (UTC)

And it's good practice to hammer in nails when it makes a product. I'm sure Lightmouse has no intention of pursuing your ridulous demand that editors at each article be "persuaded" to delink: they don't own the article (nor does Lightmouse, but he at least is improving a major fault in our articles). Tony (talk) 00:38, 27 October 2008 (UTC)

Tony may well be right that Lightmouse has no intention of accomplishing his ends by gathering consensus at the various pages. This would be regrettable, and would suggest recourse to dispute resolution.

But this is not a major flaw; indeed, in many articles linked dates are harmless: they have none of the flaws of WP:Autoformatting lists. All that needs is an article with one or two dates, after 1752, in positions (like the birth-death list in the lead) where the inconsistency between British and American formats on final punctuation does not arise. Septentrionalis PMAnderson 15:05, 27 October 2008 (UTC)

Please give a reference to the consensus for linking. Lightmouse (talk) 18:37, 26 October 2008 (UTC)

It was proposed, and immediately accepted, by Tim Starling in 2003; see archive 6 of this page. It was still consensus in April 2007 (sufficiently so that no-one, in a long discussion on dates and how they should be linked, suggested that it might be changed); see Archive D4. That's more consensus than any poll is ever likely to be. Consensus has changed since. Septentrionalis PMAnderson 19:13, 26 October 2008 (UTC)

O come now, that's the least convincing case I've heard for this C-word. Tony (talk) 00:22, 27 October 2008 (UTC)

Why thank you; that is to say that a rule which has been on a guideline page for four years and was demonstrably non-controversial during that period was never consensus. Will you hold to that postion the next time some of the hare-brained notions that infest MOS are challenged by a mere speaker of English? Septentrionalis PMAnderson 14:52, 27 October 2008 (UTC)

Just to add a voice to the debate: Because of the serious problems with linked/autoformatted dates, I feel that full linked dates should be delinked on sight. The best option is to do so with a bot that is set up to make all dates in an article use a consistent format. Linked years before 1900 should probably be reviewed by a human. The vast majority of them are linked in error, but in rare cases it may be beneficial to put an article's topic into historical context.--Srleffler (talk) 03:01, 27 October 2008 (UTC)

What serious problem are you speaking of? And why is unlinking the dates and removing the autoformatting as it currently exists the better choice when there are other ways forward? —Locke Cole • t • c 05:26, 27 October 2008 (UTC)

Please make a proposal; we could use one. Septentrionalis PMAnderson 14:53, 27 October 2008 (UTC)

The problems with linked dates have been well discussed here and elsewhere. If you're participating in this discussion you should really already know what they are. The worst problem in my opinion is that they have caused date formatting for readers to be inconsistent, by obscuring the true date format from the editors who would otherwise be likely to fix format problems. Delinking dates is a first step to beginning to clean up the mess. Editors won't fix what they can't see. Almost as bad is the fact that linked dates clutter articles with worthless links such as March 20 and 2008. Links to dates without years, and to year articles are almost never appropriate, especially for years in the 20th century and later. If you want to propose another way forward, great, but be aware that ways forward that involve getting the developers to change the software have not proven successful.--Srleffler (talk) 16:40, 27 October 2008 (UTC)

Armchair critics watch while others work hard

It's galling to observe that the noisy complainers here are totally unwilling to lift a finger to improve the woeful state of WP's date formatting. They're quite content to watch from the sidelines while others who care about this significant feature of our articles apply their skill and effort to improving this long-neglected aspect that has been allowed to fall into a state of barely controlled chaos.

Only part of the reason for the chaos has been the ill-considered date-autoformatting toy, which itself was rashly introduced and spread among unsuspecting users without this concocted notion of extensive consensus that is now being tossed about with moral outrage. No, the outrage is that those who huff and puff about the end of autoformatting—and, indeed, about addressing the cancer of linking irrelevant date fragments—have sat by for years while ISO dates have been introduced (and linked, for some reason) in the main text, often inconsistently within an article. No one, not they nor the readers, see those dates in a "favourite" order of month and day. And they have been content to watch for years as our readers have increasingly been exposed to sloppy and inconsistent American and British date mixtures within the same article, affecting more than half the articles in many areas.

The huge number of articles on US Congresses and Acts are a case in point that I have recently visited. These articles are a superb achievement—even if a little uneven in their level of detail—and are a valuable resource for the democratic process that deserves more attention by WPians. While most of their dates are in a reasonable state, some have such bizarre formats as year–day–month, with a vain attempt to autoformat them. Some have ISO dates, surely confusing for many readers, and I just caught one with British dates. Why is it that the carping critics turn a blind eye to such issues while lampooning and insulting those who painstakingly do the work? Yesterday, on a publicly accessible forum, I was called "an idiot who ... should just leave Wikipedia for good, and stop annoying people", and my contributions "asshole-ish actions". Greg L has been similarly insulted at the same forum, and anyone who disagrees with the party line labelled "MOS-nuts".

Quite a few of our most talented people have been working in the background, unsung, making headway in improving this mess. Some of this work has also involved the correction of spelling and typos, over time on a quite massive scale. Really, you people should be ashamed of yourselves. It's high time that you pitched in and did some work on something constructive for a change instead of squealing about some precious order of month and day, day and month. Tony (talk) 12:33, 26 October 2008 (UTC)

And you don't see your own words as part of the problem here? It's all us other people - the enemy? Franamax (talk) 12:41, 26 October 2008 (UTC)

Some of us correct these minor mistakes whilst we do GARs and peer reviews, etc. We also see some of the consequences of unthinking uses of such tools to make bizarre changes to, e.g. titles of reference, citations etc. However, why is date fixing regarded as the only importance task being carried out on wikipedia; and why are these editors considered to be the only talented people at work?Pyrotec (talk) 13:01, 26 October 2008 (UTC)

Pyro, you completely miss the central points: there are lots of talented people at work on WP, but the complainers here refuse to do anything about the issues at hand, content to sit around bagging those who do shoulder the burden. Franamax, I'd respond if there were substance to your post above, but it does appear to be more of exactly the attitude I'm talking about. Tony (talk) 13:14, 26 October 2008 (UTC)

Franamax, your question is spot on. Tony disparages people here, disparages people on his discussion page, disparages people on other people's discussion pages, and disparages people on various article discussion pages. This gossiping behavior, which is an overly generous description of what he does, really should stop. See WP:CIVIL and WP:AGF. Tennis expert (talk) 19:48, 26 October 2008 (UTC)

Tennis expert, I don't need to gossip, and don't bother with it. As to this accusation of "disparagement" on people's talk pages, if you're talking about what I've said about you—it's hardly worthy gossip if it's in public. Now, when are you going to help rather than hinder the project? Tony (talk) 23:30, 26 October 2008 (UTC)

Thank you for demonstrating my point yet again. Your incivility and failure to assume good faith is rampant, not just here but in many other places in Wikipedia. Also, your belief that anyone who doesn't help with your date-unlinking-agenda is "hindering the project" is patently false and incivil. Go ahead and Wikilawyer about whether public gossiping is "real gossip". I have no idea what you're doing in private, obviously. But your public behavior (trash talking everyone who opposes you) is totally detrimental to the project. Tennis expert (talk) 06:24, 27 October 2008 (UTC)

Here are some examples of disparaging names you have called me: "Tennis pest", "Tennis fanatic", "pig", "very eccentric". And those are only the ones I know about. Contrast your hostility toward me with my empathy toward you when an RFC was started about you. Tennis expert (talk) 06:57, 27 October 2008 (UTC)

I've watched this page, and I am pretty sure you will have the backing of most of WP to remove DA and to insist on a common ground for how dates should be formated: your work to get to this point is neither misguided nor gone to waste. There is a sufficiently compelling case to go forward on. However, the reason that people are specifically complaining is likely not because of these end decision, but the methods that the main editors here have approached and started to implement that decision. Yes, it is necessary to be bold and no one cannot criticize the editors here for trying to push forward these changes. But considering the WP:BRD cycle, those bold changes have gotten negative feedback, even if a minor number, and now its time for discussion, but that's not what is happening here.

There are two points that those that don't want to see DA going away that have been brought up: the retaining of some sort of meta-data that can be used to re-establish DA if Mediawiki ever provides a non-overloaded way to doing it, and how and when dates, in general, should be linking (without the use of DA). Neither point has gone through a significant discussion cycle that I've seen in the last week with those complaining from the outside; those that want DA removed keep parroting the same, yet valid reasons, those that are concerned keep asking the ask questions or provide the same alternatives that work towards the same goal those that want DA gone, but get no feedback. It is the method of process that is at issue here, not the actual issue of DA. As no one is moving from their respective corner, that is why it is time to open an RFC and get an idea of the global consensus on losing DA and when dates should be linked. I very much doubt the hard work that those wanting to remove DA will go to waste, but there may be more alternatives or a different type of consensus that has yet to be revealed on exactly we should replace DA or when dates should be linked. Everyone involved should be open to this - this is the fundamental principle of how WP is build, as a collaborative effort. --MASEM 13:27, 26 October 2008 (UTC)

The word "deprecated"

I note above that several people have read the letter of the new autoformatting policy as disallowing new additions of non-contextual date linking, but still allowing the existence of such date linking that was added before the change. More than once it has been claimed that there is consensus for ceasing to use autoformatting links, but not for removing autoformatting links. I do not believe that this is possible, as it would mean that the community wants intentional inconsistency, even within articles. Anyone looking to the spirit of Wikipedia policy rather than the letter understands that this sort of inconsistency, that makes absolutely no sense to our readers, is not helpful to the project. Nevertheless, if people are going to read the letter of the law only, I see their point, because deprecation is a complicated concept, which does not translate all that well to an encyclopedia project.

My suggestion is that we get rid of that word in favour of something clearer, that cannot be misinterpreted this way. If that means we have to discuss this all over again and finally get a watchlist notice for the discussion and everything, then I agree with MASEM on that, why not? There is no deadline. (And just to give full disclosure: I'm 100% with Tony and Greg on this issue, and I still would prefer beyond measure the old wording over this "old links good, new links bad" mess that some people seem to see in the current wording.) -- Jao (talk) 17:06, 26 October 2008 (UTC)

What do you suggest? To me it is so utterly absurd that we should disallow new additions but not permit removals that I can't see how any reasonable person could think that's what it meant. But if people nonetheless do think that (or pretend they do), then certainly a better wording might be preferable (though don't count on getting agreement on the matter easily).--Kotniski (talk) 17:44, 26 October 2008 (UTC)

I don't really see a wide consensus about anything other than autoformatting being deprecated. (I'm not entirely sure there's a consensus on that, but I'll concede it for a point of argument.) I think we need to start over and reach a wide consensus as to exactly what are appropriate date links, or at least as much as possible, leaving some to editor discretion. — Arthur Rubin (talk) 20:17, 26 October 2008 (UTC)

Well, if linking for autoformatting purposes is removed, only linking for context will remain, and the guidelines for that already exist: at WP:MOSLINK and at WP:CONTEXT. But I agree, judging by the above, there seems to me much doubt over how to apply those guidelines in this instance. -- Jao (talk) 20:40, 26 October 2008 (UTC)

actually it's not only linking for context that will remain - people have also been talking about linking dates for "metadata" and linking dates to "highlight" them. Sssoul (talk) 20:45, 26 October 2008 (UTC)

Another commonly quoted reason: 'if date links are removed, date articles will be orphans'. Lightmouse (talk) 21:03, 26 October 2008 (UTC)

Sssoul, no need for the scare quotes — I assure you metadata is a real term. It's not necessary for dates to be linked per se, but they must be marked up in some way to be identified as unquoted dates. That markup — currently represented by links, but not necessarily so — is what's being referred to as metadata. By removing the links and not replacing them with some (inert) type of markup, you're making it impossible to distinguish between an actual reference to a date and a quotation that happens to include a date. The former should always be correct, formatted properly, etc. The latter (quoted dates) need not be factually correct, typographically correct, and they should never, ever have their format changed. It's useful to be able to tell the difference, and that's what the metadata is for. The markup also eases the process of scanning for dates by various programs, either to build fact databases, look for linguistic patterns, generate automatic timelines, or (as some are fond of disparaging) generating usage statistics for various kinds of date formats (or any other purpose involving dates.) --Sapphic (talk) 21:40, 26 October 2008 (UTC)

Jao, a better, more accurate word would be preferable, but unfortunately it is the word proposed and adopted for the consensus. Since some people took the word to mean what it actually does and others interpreted it to mean a ban (and, yes, this problem was pointed out), changing it post facto is tantamount to reopening the consensus achieved for further debate, because an unknown number of people might then want to change their position on it. Personally, I think that’s the last thing most of us who were involved in the process want to have to go through again. Redefining “deprecate” will still not resolve the disputes over the removal of date links, in any case, and that’s the real problem. The simpler, easier, and less disruptive way to resolve them is to seek consensus for a proposed approach for their removal (and at least two approaches have been proposed). Askari Mark (Talk) 22:05, 26 October 2008 (UTC)

I don't see the difference between seeking consensus for changing the word to something that obviously implies removal of non-contextual links (if such consensus is deemed not to already exist), and seeking consensus for the removal of non-contextual links (if such consensus is deemed not to already exist). I'm happy either way, of course. The point is that a guideline that can be interpreted to mean that if you want to add an election date to the sentence "He entered office on March 22, 1995", you must make it inconsistent ("He was elected on January 30, 1995, and entered office on March 22, 1995") has to be changed, one way or another. On a side note, I totally agree with Kotniski that it's astounding that people do interpret it that way, even considering the use of the word "deprecated", but it seems to be a fact. -- Jao (talk) 22:21, 26 October 2008 (UTC)

How exactly are the pages on "March 22" and on "1995" going to help the reader in this case? I can imagine a piped link to a section of a page on the overall election, if any is required. Tony (talk) 00:34, 27 October 2008 (UTC)

Yes, that's exactly why we need a guideline that allows de-linking of "March 22" and "1995". To most people, including me, Kotniski and (presumably) you, that's what the guideline already says, but obviously not to everyone. -- Jao (talk) 10:41, 27 October 2008 (UTC)

What we have here, is a “failure to communicate”

We need to get more editors on the same page. Whenever a bot is let loose, editors who haven’t participated in any of the discussions here (over the last two years) come here and don’t understand any of the technical issues. The first step is for every editor who is weighing in on the issue of date linking and autoformatting to be sure they understand the details of what is really going on with this technology. So…

They can now be directed to Wikipedia talk:Manual of Style (dates and numbers)/Why dates should not be linked. {now at Wikipedia:Manual of Style (dates and numbers)/Why dates should not be linked}

WP:Why dates should not be linked

By the way, if you agree with that essay, please help out to improve it. If you disagree with it, please make your own essay as to why date linking and autoformatting is a good idea.

Greg L (talk) 22:44, 26 October 2008 (UTC)

I agree with the overall conclusion, but not the strong ties to national topic reasoning. I wrote my own essay quite some time go about the flawed design process of date autoformatting. --Gerry Ashton (talk) 22:53, 26 October 2008 (UTC)

(Small suggestion: move the essay to the WP: space at the same name, so that it's talk page can be used to comment on it.) --MASEM 22:55, 26 October 2008 (UTC)

• Thank you. Is now moved and is available at Wikipedia:Manual of Style (dates and numbers)/Why dates should not be linked

  Greg L (talk) 23:10, 26 October 2008 (UTC)

• BTW, I can’t figure out how to get that page’s talk page going. All I did was douche the article in my attempt. Greg L (talk) 23:15, 26 October 2008 (UTC)

Fixed, if that's what you wanted. Garion96 (talk) 23:19, 26 October 2008 (UTC)

• Thank you very much Garion96. That’s what we needed to implement Gerry’s objective. Happy editing. Greg L (talk) 00:06, 27 October 2008 (UTC)

Per WP:Wikipedia essays, the essay should be moved to your own personal space, since it expresses your personal views, rather than "consensus amongst the broad community of Wikipedia editors". The maligning of the date pages, in particular, does not represent WP policy. Jheald (talk) 23:25, 26 October 2008 (UTC)

• No. If WP:Wikipedia essays really says that, then real practice ignores it so that article needs to be updated. Have you seen the WP:Drop the stick essay that has been used for a long time on Wikipedia? It is in article space. Further, that rule makes no sense at all and is schizophrenic in its logic and reasoning. An essay, by definition is:

…a page containing the advice or opinions of one or more Wikipedia contributors. Essays may represent widespread norms or minority viewpoints, and they may be heeded or not based upon your judgement and discretion.

There doesn’t have to be a “consensus” view to have an essay. That’s just absurd. That’s why they’re called “essays.” Furthermore, the point of the essay is to try to develop a consensus as there currently is a serious lack of it here.

Now, I could have put that essay in user space if I wanted to have it pretty much as I see fit. But I elected to put in into article space, and, as I stated in my above posting alert: “By the way, if you agree with that essay, please help out to improve it.” It is an essay for those who share the viewpoint the essay conveys. If you disagree, then, like I also said, make your own essay; the proper response to “bad speech” is “better speech.” And if you put your essay into your user space, it will be all yours. And if you put it into article space, it will be subject to revision by others.

And finally, the purpose of this essay is to educate editors who are new to this debate on the technical goings-on with date linking and autoformatting. It is more than some silly essay on how childish editors should stop acting like children. Greg L (talk) 23:50, 26 October 2008 (UTC)

P.S. After my knee-jerk reaction to your post, I actually read WP:Wikipedia essays. It itself, is an essay. Further, it is in article space. And finally, the essay in question largely conveys facts and points editors in certain directions in a “did you know this”-type fashion. To the extent that it conveys “opinion”, those opinions are shared by a sufficient number of other editors here on Talk:MOSNUM that it clearly falls under the article space of the MOSNUM family of articles. Double “no”. Greg L (talk) 00:27, 27 October 2008 (UTC)

• And as well, there's been this information page, this gathering of consensus over just three weeks, and this essay by His Grace. Some people above seem to have missed all of these widely promoted and linked pages. Tony (talk) 00:25, 27 October 2008 (UTC)

• • I don't give a damn about autoformatting, one way or the other. Actually, it's probably a good thing if it no longer exists. But the trashing of date pages as "trivia articles" that "will only bore our readers" is completely out of order. I for one am very appreciative of the work that many many editors have put in creating and raising these pages to what they now are, and I do not think WP official space is an appropriate place to piss all over what they have done. Jheald (talk) 00:40, 27 October 2008 (UTC)

• Stop exaggerating and get a grip on yourself. The date/trivia articles aren’t being “pissed on”, as you put it. The issue is not whether or not the trivia articles have socially redeeming value and should be available on Wikipedia. The issue is simply whether or not they are sufficiently germane and topical enough in articles to merit being routinely linked to in common uses like birth dates. It’s a legitimate discussion and there is an increasing consensus in the community is that they should not be linked to. That these articles are linked to via a method that also produces autoformatting that only registered editors can see and which often thoroughly screws up articles for normal I.P. users, makes this discussion quite important. Greg L (talk) 00:50, 27 October 2008 (UTC)

• There is an RfC still open up-page which clearly shows no consensus at all to de-link birth dates and death dates. Characterising these pages as "trivia articles" is misleading, POV and offensive -- as is the whole tone of the last section which relates to them. Cut it, rewrite it, or userfy it - otherwise it goes up for XfD. Jheald (talk) 01:08, 27 October 2008 (UTC)

• Earth calling Jheald. That’s the purpose of the essay: to work towards building a better consensus on all maters pertaining to date links. So… you want to make my speech—in an essay of all things—comply with your wishes (“Cut it, rewrite it”) or you will seek to have it censored. Wikipedia doesn’t work that way and the notion that some editor would even think that way truly disgusts me. Last time I checked, Wikipedia doesn’t operate the way Communist China tries to make the Internet work for their citizens. As I said before, the proper response to bad speech is better speech. Try reflecting on what that concept means. If you can’t abide by that simple principle, go cool off and come back when you’re in a better mood.

  So you consider my calling the trivia articles “trivia articles” to be “offensive” to you. OMG! Does the essay also threaten the harmony and social order of the collective? Guess what? Whether it’s in user space or article space, it takes up the same exact amount of hard drive space. The only difference is that in article space, it’s open for others to contribute to. And that’s exactly why it’s there. I created Sewer cover in front of Greg L’s house as an essay in my user space because it was a humorous thing and I wanted it written a certain way. The Why dates should not be linked is in article space so other editors like Tony, Ohconfucius, and Gerry Ashton can participate and make it better.

  Further, I moved it from WT:MOSNUM space to WP:MOSNUM space per Gerry Ashton’s Masem’s suggestion (see his 22:55, 26 October 2008 post) because he thought it would be better if there was an associated talk page where editors can discuss matters. Are you afraid like-minded editors will share ideas in a market place where ideas are shared? Or maybe, the essay will influence others and build a consensus that isn’t to your liking (*sound of audience gasp*). Or perhaps too, you fear having it in article space somehow ennobles it with an implied stature you’d prefer it not have? Because you disagree with the message and find it’s message “offensive”? You want it to go away or somehow be diminished. Your bias is clear. Your tactics here are beneath contempt. Go write your own essay or make better arguments here if you want to influence others’ thinking; that’s how others go about it. Greg L (talk) 02:27, 27 October 2008 (UTC)

• Wikipedia space may be an appropriate home for this essay but certainly not as a subpage of the MOS. I've moved it to WP:Why dates should not be linked. –xeno (talk) 02:44, 27 October 2008 (UTC)

• Thank you Xenocidic. I’ve revised (again), my link notice above. Happy editing. Greg L (talk) 02:48, 27 October 2008 (UTC)

• Actually, it was MASEM who suggested moving Greg L's essay so it would have a talk page, but due to the way his or her post was indented, it sort of merged into my post. It does not matter; I also thought the essay should have a talk page. --Gerry Ashton (talk) 03:26, 27 October 2008 (UTC)]

• Ahhh. Now fixed. Thank you for your support, Gerry. Greg L (talk) 03:30, 27 October 2008 (UTC)

• I'd be out of a job if all engineers could write as well as that. Thanks, Greg! Tony (talk) 10:54, 27 October 2008 (UTC)

• Those few kind words made the effort (and the associated frustration) well worth it. I appreciate it. And, you’re welcome. 14:32, 27 October 2008 (UTC)

Proposal to strengthen the watery proscription of year–month–day numerical dates

A user has brought to my attention this little mess, in which the numerical dates (1) jostle with the numerical publication numbers, (2) are probably not comprehensible by most of our readers (is "1943-09-08" September 8 or August 9? It's made harder by the non-linear US medium–small–large date format (month–day–year) as a reference point for many readers.), and (3) are inconsistent with the US format used elsewhere in the main text.

Currently, MOSNUM says this:

YYYY-MM-DD style 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. Because some perceive dates in that style to be in conformance with the current ISO 8601 standard, that format should never be used for a date that is not in the (proleptic) Gregorian calendar, nor for any year outside the range 1583 through 9999.

What the hell is "ease of comparison"? No one has ever explained that bit.

I propose that this be strengthened thus:

YYYY-MM-DD style dates Wholly numerical dates such as "1976-05-31" are uncommon in English prose, and are generally not used in Wikipedia the main text of Wikipedia articles. However, they may be useful in long lists and tables for conciseness and ease of comparison. Because some perceive dates in that style to be in conformance with the current ISO 8601 standard, that format should never be used for a date that is not in the (proleptic) Gregorian calendar, nor for any year outside the range 1583 through 9999. In tables in which there is a shortage of space, three-character abbreviations of the months may be used in the order that is consistent with the prevailing date format in the main text (e.g., "Aug. 5, 1961" or "5 Aug. 1961").

I'd rather ban them in all tables, but that might be going too far. Tony (talk) 02:17, 27 October 2008 (UTC)

I see no reason to use them in tables when it seems to be ok to use "12 Nov 2008" or "Nov. 12, 2008" which is exactly one or two characters longer than "2008-11-12" but reads much easier. ISO should only be used in references and that until we work out how to replace DA. --MASEM 02:24, 27 October 2008 (UTC)

• Agree with Tony. Greg L (talk) 02:29, 27 October 2008 (UTC)

Your rewrite of the ISO 8601 part, in my view, changes the meaning. Before, the date 1582-10-14 was totally unacceptable for any purpose and could be changed on sight, because it is in the YYYY-MM-DD format and the year was less than 1583. The new statement says the ISO 8601 standard should never be used for years outside the 1583-9999, so it might be ok to use the format so long as it was not accompanied by a statement that it conforms to the ISO 8601 standard. --Gerry Ashton (talk) 02:29, 27 October 2008 (UTC)

The "ease of comparison" bit is straightforward and should be retained. Basically, the yyyy-mm-dd sequence is hierarchical, inspectable and sortable. It's more easy to see that 04 comes before 05 than to see that Apr comes before May. Franamax (talk) 03:53, 27 October 2008 (UTC)

The "sort" template deals with this. eg sort|2008-11-12|Nov. 12, 2008. --MASEM 04:07, 27 October 2008 (UTC)

The documentation for the sort template is horrible, and in my book, a template is no better than its documentation. --Gerry Ashton (talk) 04:14, 27 October 2008 (UTC)

It's pretty straight forward: two parameters, first is the sort key, second is the displayed text. It basically expands to a bit of display:none CSS for the sort key. The template doc can be fixed but I've used the sort template with iso dates as the sortkey but normal dates for songs (see, for example List of songs in Guitar Hero III: Legends of Rock) --MASEM 04:18, 27 October 2008 (UTC)

Masem, I've checked that article and I find no instances of {{sort}} - perhaps you subst'd the instances? I do see the sort technique though (rather clunky) and I see another template now {{dts}}, so that's good to know.

More worryingly, even though I figured there was a way and I'm somewhat used to the wiki-hunt, how was I supposed to know about those templates? I naturally checked at Help:Table, which leads to Help:Sorting and both are Meta pages and seem to contain no links to the en:wiki sorting templates. So not only is the template docs not so good, I can't even find a way to get where I'm supposed to be confused! Asking the average editor to figure all this stuff out whilst forbidding the simple yyyy-mm-dd format so they can get their table to work at all just seems a bit much. I dunno, was there some obvious clue I should have picked up on? I'm not that swift sometimes... Franamax (talk) 05:30, 27 October 2008 (UTC)

That's a notable, but fixable problem, if we decide to completely rid ourselves of ISO-type dates in the article body; both sort (which is general purpose for any sortable key, so those lists use for "the"-less sorting, the dates as given are effective what it would expand out to) and dts help achieve this for the only place where ISO dates (outside of the ISO date article itself) are left being used. These do need to be discused in at least the help for sortable tables, but that's easily fixed. --MASEM 05:43, 27 October 2008 (UTC)

It's time to call this hyperbole that's been repeated here way too many times. The assertion that YYYY-MM-DD formated dates such as 1943-09-08 "are probably not comprehensible by most of our readers" is unreferenced, weasel worded, and just plain insulting to the reader's intelligence. Can anyone furnish reliable sources that show that "most" of any group similar to the users of the English WP: (Internet-using, English-literate, and globally distributed) have problems understanding that numbers are normally written most significant digit first? LeadSongDog (talk) 04:43, 27 October 2008 (UTC)

I personally have no problem at all, when I see a leading year, I switch over to hierarchical mode, so I would read the above as Sep. 8, 1943. However, I'm Canadian, so I'm used to seeing multiple date formats, maybe that's not a universal experience? Certainly 02/03/04 can be ambiguous, but to me the leading four-digit year in dashed YMD notation is not confusing. (And *ahem*, that's where linked day articles can make things very clear - but I digress :) Franamax (talk) 05:02, 27 October 2008 (UTC)

• It’s not that they can’t be understood. They’re just not normally written that way so they are less intuitive and takes more effort to read. People read by looking at words. That’s why “July 4, 2001” is much easier for Americans than parsing out 2001-07-04. The latter is ugly to boot. What Tony proposed is simple common sense. Greg L (talk) 04:54, 27 October 2008 (UTC)

• I think we need to sort out date autoformatting before we start banning certain date formats entirely as seems to be the proposal here. Strongest possible oppose to this proposal. —Locke Cole • t • c 04:59, 27 October 2008 (UTC)

• Now that we're down the road of more than one acceptable date format (Int and Amer), I'm going to start pushing for yyyy-mm-dd format for all China-related articles. No more Mr Nice Guy ;-) Ohconfucius (talk) 07:46, 27 October 2008 (UTC)

I don't know how you could get anything but 8 June 2004 out of 2004-06-08: for those used to month-day-year, the year is just moved in front and the rest is as usual; for those used to the logic of little to big, this is just as logical only in reverse. But I can't speak for everyone on this. The point is though that this format is uncommon in English ... how they do it in Chinese doesn't matter since this is the English WP. ISO dates serve no purpose here on WP. There are templates to deal with sorting (if they are unfindable, make them findable) and there are three-letter abbreviations to save space. I strongly support Tony's proposal—in fact I'd be happy to make it an all-out ban. JIMp _talk·cont 09:08, 27 October 2008 (UTC)

Rejoinders from Tony:

• Gerry, I was going to ask you to suggest a rewording of the last sentence in the proposal, given your note above; but other suggestions here may have superseded that. Are you in agreement with the reworded proposal?
• I'm an experienced reader/editor, but I still have to dwell on these numerical dates to mouth out in words the month and the day. I guess it's confusing that throughout the English-speaking world we use both month–day and day–month order, which is just fine when the month is spelled out. But as Americans who travel to the UK or Australia soon discover, 2/5/08 means May 2, 2008, not February 5, 2008. It's just asking for trouble on an international site not to spell out the month. We need dates to be simple and explicit on WP. This numerical gobbledygook should be flushed down the pan for the sake of our readers.
• I see much to commend Masem's point that a three-character abbreviation of month is just fine for saving space in tables. I'm going to be bold and change the proposal to this effect.
• Franamax, I don't see why "04" and "05" has any advantage over "Apr" and "May", since every English-speaker is adept with the sequence of months (as with the days of the week). The problem is readily identifying which of the digits is the month and not the day (emphasis on readily—we can all stare at it like a chess puzzle and work it out, but I suspect many of our readers simply glaze over when they see dates that look like telephone numbers.
• Locke Cole, this is quite a separate issue from DA, of course. Tony (talk) 10:24, 27 October 2008 (UTC).

Please compare these three examples of lists starting with dates, from the "mess" I linked to at the top of this section. Feel free to comment below. Tony (talk) 10:47, 27 October 2008 (UTC)

Example A: numerical dates:

• 1943-12-17 — Magnuson Act (Chinese Exclusion Repeal Act of 1943), Sess. 1, ch. 344, 57 Stat. 600
• 1944-02-03 — Mustering-out Payment Act, Sess. 2, Pub. L. 78–225, 58 Stat. 8
• 1944-06-22 — Servicemen's Readjustment Act of 1944 (G.I. Bill), Sess. 2, ch. 268, Pub. L. 78–345, 58 Stat. 284
• 1944-06-27 — Veterans' Preference Act, Sess. 2, ch. 287, Pub. L. 78–359, 58 Stat. 387
• 1944-07-01 — Public Health Service Act, Sess. 2, ch. 373, 58 Stat. 682
• 1944-12-22 — Pick-Sloan Flood Control Act, Sess. 2, ch. 665, Pub. L. 78–534, 58 Stat. 887

Example B: dates with abbreviated months:

• Dec. 17, 1943 — Magnuson Act (Chinese Exclusion Repeal Act of 1943), Sess. 1, ch. 344, 57 Stat. 600
• Feb. 3, 1944 — Mustering-out Payment Act, Sess. 2, Pub. L. 78–225, 58 Stat. 8
• Jun. 22, 1944 — Servicemen's Readjustment Act of 1944 (G.I. Bill), Sess. 2, ch. 268, Pub. L. 78–345, 58 Stat. 284
• Jun. 27, 1944 — Veterans' Preference Act, Sess. 2, ch. 287, Pub. L. 78–359, 58 Stat. 387
• Jul. 1, 1944 — Public Health Service Act, Sess. 2, ch. 373, 58 Stat. 682
• Dec. 22, 1944 — Pick-Sloan Flood Control Act, Sess. 2, ch. 665, Pub. L. 78–534, 58 Stat. 887

Example C: full dates in a standard WP format:

• December 17, 1943 — Magnuson Act (Chinese Exclusion Repeal Act of 1943), Sess. 1, ch. 344, 57 Stat. 600
• February 3, 1944 — Mustering-out Payment Act, Sess. 2, Pub. L. 78–225, 58 Stat. 8
• June 22, 1944 — Servicemen's Readjustment Act of 1944 (G.I. Bill), Sess. 2, ch. 268, Pub. L. 78–345, 58 Stat. 284
• June 27, 1944 — Veterans' Preference Act, Sess. 2, ch. 287, Pub. L. 78–359, 58 Stat. 387
• July 1, 1944 — Public Health Service Act, Sess. 2, ch. 373, 58 Stat. 682
• December 22, 1944 — Pick-Sloan Flood Control Act, Sess. 2, ch. 665, Pub. L. 78–534, 58 Stat. 887

Comment—Yes, the numericals in A are the neatest visually and are compact, but the abbreviated version is almost as neat and compact. These examples do show the full dates in C in the poorest possible light, and I think they work best in most contexts. But a bit of deft formatting can improve their look—I'd be inclined to put the full dates last, in parentheses, like this:

• Magnuson Act (Chinese Exclusion Repeal Act of 1943), Sess. 1, ch. 344, 57 Stat. 600 (December 17, 1943)
• Mustering-out Payment Act, Sess. 2, Pub. L. 78–225, 58 Stat. 8 (February 3, 1944)
• Servicemen's Readjustment Act of 1944 (G.I. Bill), Sess. 2, ch. 268, Pub. L. 78–345, 58 Stat. 284 (June 22, 1944)
• Veterans' Preference Act, Sess. 2, ch. 287, Pub. L. 78–359, 58 Stat. 387 (June 27, 1944)
• Public Health Service Act, Sess. 2, ch. 373, 58 Stat. 682 (July 1, 1944)
• Pick-Sloan Flood Control Act, Sess. 2, ch. 665, Pub. L. 78–534, 58 Stat. 887 (December 22, 1944)

After all, the names of the Acts are what should come first—they're a much better point of departure than the exact date of passage. Tony (talk) 10:47, 27 October 2008 (UTC)

Tony, I don't necessarily disagree with any of your above, except the bit about the chess puzzle. YYYY-MM-DD is really not all that puzzling. In fact, here in Canada at least, where we're headed to one-day financial clearing, cheques now have that same format (sans dashes) Image:CanadianChequeSample.png. Now it's true that Americans still measure in cubits, but I don't see a huge stretch to accomodate three rather than two date formats, namely "dd Mon yyyy", "Mon dd, yyyy" and "yyyy-mm-dd" (and expansion of "Mon" to "Month", since someone will always want September 11, 1918 rather than 11 Sep 1918). The YMD usage should be discouraged in articles, but not forbidden in tables where it can be used to good effect and by it's own context in the column solve the chess puzzle anyway. Franamax (talk) 10:58, 27 October 2008 (UTC)

And my pick for conciseness and readability, unfortunately, is your first set of examples. :( Franamax (talk) 10:58, 27 October 2008 (UTC)

I'm not sure how many e/c's I've fallen behind now - anyway, look at your Example A: there's a good temporal layout which is intuitive and obvious for any reader, by the very nature of the progression of dates. Of course, this would only be applicable when the editor wished to demonstrate a temporal progression of events, but I think example A makes that progression most easily clear. Franamax (talk) 11:06, 27 October 2008 (UTC)

Not intuitive for every reader, experto crede; and B and C are less clear only for those readers who have trouble with the months of the year. This is not the Simple English Wikipedia. Septentrionalis PMAnderson 16:11, 27 October 2008 (UTC)

Of course it's not a separate issue from date auto formatting, because if DA worked as expected for all users (logged in or not) then the matter of what date format to allow/disallow within articles would be a moot point. So please get back to helping resolve DA rather than plowing ahead with other changes. —Locke Cole • t • c 13:25, 27 October 2008 (UTC)

I oppose a ban on ISO compliance within Wikipdedia. It seems to me that Wikipedia and international standards should be friends by default, not enemies. I will support a ban on using ISO format for non-ISO dates (the Gregorian problem referred to by Gerry Ashton). As much as I respect Tony on many issues, I think he is wrong to ban them. I wonder what the citation people think. I think MOS guidance should have a 'clear and present problem' test and this would fail such a test, I think. Lightmouse (talk) 13:19, 27 October 2008 (UTC)

• I like option B. It’s very human-friendly, and it’s compact enough for this backwater use. To me, it is a superior trade off. I haven’t thoroughly looked over the ISO standard, but I’m quite sure they weren’t suggesting that all-numeric dates should be the preferred way for dates to be written out for humans to read in body text. I’m confident the point of the standard was to bring order (standardization) to those instances where all-numeric dates are normally used. You see all-numeric dates in computer databases, on blue prints (technical drawings), tables of data, etc. I’ve seen some editors get awfully enamored with ISO dates and defend them for use in body-text like “after the 1941-12-07 attack on Pearl Harbor”. That’s not what they’re for. Greg L (talk) 16:01, 27 October 2008 (UTC)

• I'm indifferent whether all-numeric dates are declared contrary to this guideline, and our "blessing" is transferred to three-letter abbreviations. If three-letter abbreviations are adopted, we should decide whether or not a period is to be used after the abbreviation. As an aside, I was watching the World Series, and I noticed an advertisement displayed on the wall behind the batter (it was actually added in the television production process; the real wall is a green screen. The ad was for some product or movie due out in December, and the date was formatted something like "12.15.2008". (I'm certain periods were used as separators, I'm not sure which date in December the thing is becomming available). So apparently even mass-marketers are willing to believe Americans have some flexibility reading date formats. --Gerry Ashton (talk) 18:16, 27 October 2008 (UTC)

Interesting, Gerry. I think it's foolish of the advertising company: I'd ditch the year (it's obvious) and say "out December 15", or if desperate for space, "out Dec. 15". From a communications standpoint, the format they used was a big mistake. Tony (talk) 01:16, 28 October 2008 (UTC)

• OK, I have a suggestion on whether the abbreviated months should have periods or not. I note that my local paper, which probably follows one of the major U.S. newspaper style guides (NY or Chicago), handles it this way: When someone writes a letter to the editor, they will insert a parenthetical pointing to which issue the reader is referring to. So you see text like this:

I see from your Sunday article (“Americans shit their pants when they see metric stuff,” Sept. 28), that you left no stone unturned…

Again, note that this is being used in a parenthetical, which simulates the back-water nature of publication citations. Note that the paper also uses flex, three-to-four-letter abbreviations and periods. This allows months like June to be written out. Thus, I believe the progression goes as follows: Jan. Feb. Mar. April May June Jul. Aug. Sept. Oct. Nov. Dec. It abandons the rigid, three-letter-only notion in order to make it more human-friendly. Greg L (talk) 19:53, 27 October 2008 (UTC)

If we write April in full, would we not also write March in full? –xeno (talk) 20:16, 27 October 2008 (UTC)

• I’m not sure about the exact practices of the newspaper. I’ve written a number of letters to the editor (go figure) and have them archived. So I know for a fact about some of them. They might indeed spell out March. I am assuming that they follow an industry-standard practice and I would recommend that we simply adopt that practice since I can see what the goal is and it makes sense. I happen to have the phone number of the gal at the paper who calls to confirm whether it was really me who sent in the letters. I think I’ll call her and ask her how the progression goes. Greg L (talk) 21:31, 27 October 2008 (UTC)

If the only concern is reducing the amount of space taken up by months, we could achieve the goal nearly as well by allowing short months (e.g. June) to be written in full. However, if all abbreviations are exactly three letters, tables can be formatted with fixed-width fonts to force all the elements of the date to align vertically. I suppose exactly how the editor envisions the reader will use the table will determine whether the improved alignment is worth tolerating the ugly fixed-width fonts. --Gerry Ashton (talk) 20:38, 27 October 2008 (UTC)

• Addressing Xeno and Gerry: For tables, where exact fit and alignment are important, I don’t see any reason at all why we can’t have three-letter abbreviations. For sidebars and citations, and similar non-body-text uses where tight fit isn’t a consideration, I would suggest flexible abbreviations like newspapers use for parenthetical use when citing a specific edition. I just got off the phone with the editor of our local paper. They use the Associated Press manual of style. He said all U.S. newspapers follow the AP guideline so this abbreviation method is consistently used across the country. It is as follows:

• Jan.
• Feb.
• March
• April
• May
• June
• July
• Aug.
• Sept.
• Oct.
• Nov.
• Dec.

Greg L (talk) 21:46, 27 October 2008 (UTC)

Thus, we have…

Example D: Associated Press-style abbreviated months:

• Sept. 17, 1943 — Magnuson Act (Chinese Exclusion Repeal Act of 1943), Sess. 1, ch. 344, 57 Stat. 600
• Feb. 3, 1944 — Mustering-out Payment Act, Sess. 2, Pub. L. 78–225, 58 Stat. 8
• June 22, 1944 — Servicemen's Readjustment Act of 1944 (G.I. Bill), Sess. 2, ch. 268, Pub. L. 78–345, 58 Stat. 284
• July 27, 1944 — Veterans' Preference Act, Sess. 2, ch. 287, Pub. L. 78–359, 58 Stat. 387
• April 1, 1944 — Public Health Service Act, Sess. 2, ch. 373, 58 Stat. 682
• Dec. 22, 1944 — Pick-Sloan Flood Control Act, Sess. 2, ch. 665, Pub. L. 78–534, 58 Stat. 887

Greg L (talk) 21:52, 27 October 2008 (UTC)

• Now that we have the above AP-style abbreviated month progression captured above, wouldn’t it be good to capture this information in MOSNUM? I’d offer to do it, but I’m a hundred times more familiar with Talk:MOSNUM than MOSNUM itself; I’d probably put it in the wrong spot. Greg L (talk) 00:11, 28 October 2008 (UTC)
  • Does it serve any purpose? Month abbreviations are well known, and anybody who wants can reconstruct them as easily as we constructed them. Septentrionalis PMAnderson 00:56, 28 October 2008 (UTC)

• Maybe they are well known to you, but I doubt that one editor in ten here can properly write down the AP version of the month abbreviations without looking. Greg L (talk) 01:06, 28 October 2008 (UTC)
  • Is there any profit to the encyclopedia to insisting on March and April over Mar. and Apr.? Would there be any to insisting on the other way around? If not, let us abstain. Septentrionalis PMAnderson 01:09, 28 October 2008 (UTC)

• It’s so abbreviations here don’t somehow look awkward. I’ve always wondered why my casual attempts at monthly abbreviations looked odd. I’d try Jan. Feb. Mar. Apr. Jul. Sep. and they’d look *funny*. Then I’d try four-character abbreviations and they’d look odd. It’s because newspapers (and probably many magazines too) observe a specific practice where they have variable number of characters. The benefit to Wikipedia is a style that looks natural and reads fluidly. I would simply propose that when there are tight space considerations and a particular need for columns to line up, three-letter abbreviations are fine. If it is in main body text, spell out the name of the month. If it is side-bars, citations, and other such uses, follow the AP’s flexible-character abbreviations observed by every subscribing English-language newspaper in the world. Why is there any controversy to this?

  Maybe the AP-style monthly abbreviations are well known to you, but I doubt that one editor in ten here can properly write down all twelve months without looking at the above list and without error. After all, the editor of a major city newspaper had to go to his AP handbook to provide them to me without error. Besides, MOSNUM isn’t only for über-experienced editors like you; it is here to help editors with minimal editing experience (which is why many would come here in the first place). I’m curious why you would even question the value of memorializing the abbreviations here. Wikipedia is authored largely by novices. The Associated Press is populated by writers, all of whom have journalism degrees and they saw fit to put it into their manual of style. Greg L (talk) 01:24, 28 October 2008 (UTC)

  • The AP stylebook is intended for reporters: writers who are in a hurry, and whose stories will not be revised after publication, usually in a matter of hours. We have more lesiure.
  • We also have a Manual of Style which is now infinitely less comprehensible than the AP 's, because we have crammed more into it than our organization can make clear. The solution to that is to be more hesitant about cramming even more into it.
  • Our MoS was never intended to cover all the topics of any stylebook; it was intended to say what was necessary for our purposes: stuff they omitted, matters related to wiki format, content rules needed because anybody can edit us (like ENGVAR), decisions where stylebooks disagree (including agreements that we may differ from article to article), and those purely stylistic points which are most important to us and so worth spending space and complexity on. . Septentrionalis PMAnderson 03:05, 28 October 2008 (UTC)

Common Use doesn't make it right It's common-use in US English prose to call Eastern Time: EST even though that's not correct (in other words NY time gets called EST even in the summer). Therefore the argument that common-use in US English prose should determine Wikipedia convention is a weak argument in comparison to the unamibugauce ISO standard (even for non-traveled US citizens). The idea that anyone could get Aug 9th out of 1943-09-08 is unlikelyhood piled on top of unlikelyhood. I agree with that statement that it's "insulting to the reader's intelligence" to think they can't understand 1943-09-08. —Preceding unsigned comment added by Zaurus (talk • contribs) 02:20, 28 October 2008 (UTC)

People who proclaim that "no-one could possibly misunderstand what I meant" should at least learn how to spell first. Septentrionalis PMAnderson 03:05, 28 October 2008 (UTC)

Agreed, as far as I know, no dating system in the world has ever used YYYY-DD-MM. Short of spelling out the month, it's the most disambiguous way to give a date. –xeno (talk) 14:18, 29 October 2008 (UTC)

• The harm, in posting a list of AP-style month abbreviations here is… what? Greg L (talk) 03:25, 28 October 2008 (UTC)
  • Every additional sentence makes it that much more difficult to find stuff, and that much easier for eyes to glaze over. Small harm, but so is the benefit. It's laudable of you to dig this up; but think of it from the next reader's point of view, when this is just another bit of clutter. Septentrionalis PMAnderson 03:34, 28 October 2008 (UTC)

• Maybe I'm wrong, but I think newspapers hardly ever publish tables containing month names (or numbers) where it would be nice to have them align vertically. Wikipedia is more likely to do so. So while the newspaper could confine itself to two sets of month names (full and abbreviated), we would need three (full, AP abbreviations, and three-letter abbreviations). I would prefer to stay with just two sets, full and three-letter abbreviations, even if the latter don't read quite as smoothly. --Gerry Ashton (talk) 03:34, 28 October 2008 (UTC)
  • Not that much more smoothly; most users will be seeing proportional font. Septentrionalis PMAnderson 03:36, 28 October 2008 (UTC)

• Another point against ISO-like formats is : Won't someone thing of the children?? Seriously, if we're trying to write a work that is accessable across as much regional variances and ages as possible, I will tell you that the ISO-like format is not something that a grade schooler will be able to recognize, but I know in the US that the months of the year are pre-school if not earlier (and I presume in the EU and other parts of the world, taught in DD MMM YYYY format). I have no problem if ISO is used in wikicode, but presented work should avoid this format save for discussions of what the ISO format is. --MASEM 14:30, 29 October 2008 (UTC)

Exception

Any such re-writing needs to allow for the fact that templates like {{Start date}} and {{Birth date}} have parenthetical YYYY-MM-DD format dates, hidden by CSS (and thus visible to people with CSS unavailable or disabled as, say, 31 May 1975 (1975-05-31)) in order to output the ISO-format dates required in microformats. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 11:27, 27 October 2008 (UTC)

Note that the {{TL:Birth date}} provides a date in the format required by HCard. That relies on vCard, RFC 2426, which is intended for directory, "white-pages" information, as would be found on a business card. Thus it is only intended for living people, or people who have died recently. Using it to describe people who have been dead for millenia can be expected to cause trouble, and sure enough, it does. The vCard spec relies on RFC 2426 to define some of the terms it uses, and in RFC 2426 we find " 'date', 'time', and 'date-time': Each of these value types is based on a subset of the definitions in ISO 8601 standard." This means that it is wrong to use {{Birth date}} (and most likely, any of its cousins} for any person who's birth date is written in any calendar other than the Gregorian calendar. --Gerry Ashton (talk) 23:17, 27 October 2008 (UTC)

hCard is based on vCard, but is used for more than just "white pages" info, and hCards are legitimately used for more than living (or still-warm) people. The issue of pre-Gregorian hCard dates (which I raised on 19 April 2007) remains to be resolved (the microformat community is notoriously and lamentably slow in such matters). While your latter point may arguably be valid, it has no bearing on the point I made. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 23:29, 27 October 2008 (UTC)

If there is to be an exception, my post has bearing on how the exception is worded. The exception might or might not use the phrase "ISO 8601". The exception might recklessly claim that the format YYYY-MM-DD is permitted for invisible machine-readable purposes, but Wikipedia specifically rejects any implication that ISO 8601 applies, and makes no statement about what calendar the date is written in. Or the exception might say these microformats are OK so long as the year is greater than or equal to 1583, or 1753, but is forbidden for earlier years. --Gerry Ashton (talk) 23:42, 27 October 2008 (UTC)

The use of abbreviated month names in tables has to my mind some limited merit, but considerable associated complication. Because most editors will not take the trouble to memorize the list of approved abbreviations and their punctuation a (semi-)automated spelling repair would have to be implemented to catch all the variations and bring them into conformance. Should US/international variants also apply, then that automation will need to be sensitive to the choice of variant that is adopted for each article in question. The usual argument for the US style "Jan 1, 2000" (that it avoids starting sentences with a numeral) does not apply in tables. I see no merit in using the abbreviations in citation/cite template inputs or outputs. These should be done with outputs fully spelled out (should that be the ongoing consensus) and with inputs in any unambiguous form that parses correctly. I agree with Gerry that the misapplication of the term ISO 8601 to describe simple YYYY-MM-DD form is problematic, particularly for historic dates and for timezone sensitive articles, such as those dealing with aviation events. We should explicitly distinguish the form YYYY-MM-DD from the ISO 8601 standard.LeadSongDog (talk) 17:54, 29 October 2008 (UTC)

• Rant time: I’m not saying that only abbreviated months should be used and that all-numeric dates should not be used; each has its place. I’m just saying that abbreviated months have their uses, and when they are used, they should be properly written. There is far too much tendency from certain elements here to semi-invent special formats for use here on Wikipedia, or to adopt some cockamamie standard (like all-numeric dates) just because it has been anointed by ISO, BIPM, or whomever. The ISO only standardized how numeric dates should be formatted; it didn’t suggest they should be used in normal prose. Newspapers do too use abbreviated months. And that’s why the Associated Press, whose practices are observed by all newspapers throughout the English-speaking world, has a guideline governing how to write them. It’s a logical system, is what we are used to reading, and is what we subliminally recognize as looking right. Short of some use like cramming abbreviations into a super-tight chart or some other unusual situation like that, you never abbreviate all the months names into either fixed three or four-letter groups. The proper abbreviations are as follows:

• Jan.
• Feb.
• March
• April
• May
• June
• July
• Aug.
• Sept.
• Oct.
• Nov.
• Dec.

It is not Apr., Jun., and Sep. Nor is it “Sept. Octo. and Nove” (only “Sept.”). Proper practice is to use variable-length abbreviations. Further, I utterly reject any argument that there is “too little room on MOSNUM for this list,” or that it “lengthens MOSNUM to the point where readers can no longer find what they need,” or that the “AP has professional writers with journalism degrees who are so damn busy that they need such a guideline but Wikipedians have more time and they don’t need such a guideline”; those arguments are totally absurd. Beyond absurd.

What we’ve got here are some editors who are far too damn pushy and are blocking common-sense adoptions of simple guidelines that enjoy world-wide recognition. Some editors here behave as if MOSNUM and WT:MOSNUM are some sort of private reserve for them to make their personal and indelible imprint upon. What I really think is that some editors here have been doing things a certain (wrong) way, that they A) are completely convinced their way is logical and best, and B) don’t want to see their work eventually changed in any way. It’s not that I’m suggesting some editors here lack bad faith; I just think they are so biased on this issue, that they are desperately resorting to arguments that are utterly fallacious. That’s my opinion… so shoot me. We are not going to be inventing new home-grown conventions for abbreviating months in citations, sidebars, and parentheticals. What’s with this tendency on MOSNUM that it enables editors to come here and push through practices (like “256 kibibytes”) that no one else in the real world observes on this pale blue dot? It’s time to stop pushing through guidelines—home-grown in some cases—just because someone thinks they’ve figured out a *better way to a brighter and more logical future* or are misinterpreting what some standards organization is really trying to accomplish.

After I get this simple damned list of how to abbreviate the months on MOSNUM, I’m going to stop wasting oodles of time fighting every common-sense minor point. Instead, I intend to focus on fixing the very underpinnings of how MOSNUM functions, arrives at a consensus (or not), identifies when a consensus has truly been reached, resolves conflict (which is pretty much all that occurs here), and avoids making half-baked decisions. Right now, this place is broken beyond all recognition, requires that editors invest far too much time arguing, and it gives far too much weight to the editor who shouts loudest, longest, and is most willing to push the boundaries of edit warring. Greg L (talk) 18:20, 29 October 2008 (UTC)

Time for decaf;/? I've said before, this is the wrong venue for discussion of what constitutes consensus. That's Wikipedia talk:Consensus (which desperately needs fixing). If fixed there, it'll spill over here. Some samples from todays news here, here, here, here and here make it pretty clear there is zero consistency of date style in the real world press, whatever the AP stylebook says. Articles even use different styles in their RSS dateline from the displayed dateline ensuring that aggregators like google news get something different. 18:48, 29 October 2008 (UTC)LeadSongDog (talk)

Help with terminology

Can anyone explain to me whether something can be simultaneously:

• minor and inconsequential
• controversial
• too important to build into 'general fixes' of AWB

It seems to me that people are mangling the english language by saying that date delinking fulfils all three bullets. Lightmouse (talk) 12:46, 27 October 2008 (UTC)

As I've said to you above, you need to stop conflating an MOS debate with an AWB misuse issue. Suggestions on how to properly work in these minor/inconsequential delinking edits have been given on your talk page - simply ensure that you enable RegEx typo fixing and skip when no typo is fixed. This also means you actually have to check every edit, which provides the added benefit of lowering your false positive rate. As to why delinking can't be built into general fixes: because there is no way for AWB to tell when a date link is provided for context vs. when it was provided for autoformatting. –xeno (talk) 12:51, 27 October 2008 (UTC)

If I understand much of the previous commentary, the explanation you seek lies within context and scope. Few would argue that the removal of date links in any given article is not "minor and inconsequential", while the act of doing so on a broader scope can very well be "important and controversial". By way of an illustrative example : cutting down any given tree in a forest is minor and inconsequential, but cutting down every tree in the forest is important and controversial. I cannot speak to the AWB issue directly (I have never used AWB and have no experience with it) but it is fairly clear that even the most minor sort of edit becomes a big deal when it is performed many thousands of times. Shereth 13:56, 27 October 2008 (UTC)

• Are you implying that DA and date-linking is like the trees in a forest? If so, I believe you and I are on completely different planets, so th speak. For me, one date-linked page is like one plastic bottle left to litter a beautiful beach. We should pick up and dispose of all the plastic bottles, and restore the beauty of our beach. Cleaning up is important and should not be controversial at all. Ohconfucius (talk) 15:21, 27 October 2008 (UTC)

• LM asked for an example where something could be considered "inconsequential" yet "important" at the same time; I provided one. Your example serves equally well. I am neither advocating nor opposing the delinking of dates. I am merely interested in seeing a solution develop out of community consensus rather than a few interested editors, given the scope of the work being proposed. Cleaning up is important and should not be controversial, but it has generated controversy nonetheless and the question of whether or not this is beneficial cleanup of litter or harmful razing of trees is one the community should answer. Shereth 15:26, 27 October 2008 (UTC)

• Thanks for the clarification. Ohconfucius (talk) 16:08, 27 October 2008 (UTC)

No. The whole purpose of AWB is that it allows many small edits to be made. The 'minor and inconsequential' rule serves to avoid server burden for edits that are hardly noticeable, such as removing double spaces. It was not put in place to permit those with coercive power to censor pro-MOS edits that they don't want done. You suggest that autoformatting links are trees in a forest, well, the MOS says those trees need chopping down. You can't have an MOS that can't be implemented. Lightmouse (talk) 14:19, 27 October 2008 (UTC)

the MOS says those trees need chopping down. No, it doesn't; this is not The Most Important Thing. The marginal benefits involved in removing many (not all) date-links do not justify thoughtless and uncivil mass editing; the equivalent of a forest-fire. If someone is editing an article anyway, I would encourage xim to trim overlinking while xe are about it, and that is what is this guideline is supposed to produce. No more; that's why it's a guideline. Septentrionalis PMAnderson 14:47, 27 October 2008 (UTC)

• Then how about thoughtful and civil mass delinking? By my reckoning, lists are responsible for more than 50% of the date links within wikipedia. By selectively culling date links in articles like I have done in Category:Lists of television series episodes, I believe I am are carefully, selectively, thoughtfully clearing the dead wood so that trees may be seen and appreciated for their true beauty. BTW, these hundreds of edits I have now performed have caused no more than 2 editors to complain, and only about the wrong destination date format - pretty trivial, if you ask me. Ohconfucius (talk) 02:17, 29 October 2008 (UTC)

Date links themselves are of marginal and rare benefit. At least some classes of date links are suitable for mass-removal with little human oversight. Links such as March 20, and links to years since 1900 are almost entirely inappropriate. Cases where those links would be justified are so exceedingly rare that I doubt they justify the great increase in effort required to do manual link removal rather than automating it.--Srleffler (talk) 17:22, 28 October 2008 (UTC)

• Real guidelines express what any random dozen Wikipedians may be expected to say on a given issue. They exist because the majority of the dozen will need to explain themselves to the dissentient minority and the newbies, and they are tired of repeating themselves; so the majority bungs down what they agree on in Wikipedia space, and work on making it clearer over time. Therefore they don't need "enforcement", much less an enforcer; the real rule rests in the general consensus, and another random dozen Wikipedians will support it when necessary. The phrasing in Wikipedia space is an approximation of this, stuck together with Scotch tape and piano wire; just as WP:PRO says.
• What rule about date linking has that degree of support is unclear; the stream of complaints out of the wordwork casts doubt that any statement short of not too many links would. Septentrionalis PMAnderson 01:05, 28 October 2008 (UTC)

• And how long, exactly, is a piece of string? Your prescription is far too vague to be of use in a guideline. If the above approach is adopted, we enter into a very subjective realm with little or no clarity. I don't think anybody is suggesting banning all links to date/year articles any more. I believe most people seem to accept that some small number of date/year links may be valid in certain biographical and historical articles, and it has been actively proposed that these dates should be piped or displayed differently. We need not get into circuitous debate if we can just agree on that. The reason no-body has come up with a 'how', yet appears to be because most who favour retaining some sort of linking are resistant to anything but a direct wiki-link to the date and year article, without piping and without the use of templates. This is not about scotch tape any more... Ohconfucius (talk) 02:37, 28 October 2008 (UTC)
  • If there is no consensus on a clear line, we must be vague - or, what is often preferable, say nothing. Doing otherwise involves both bullying and scotch tape; some editors may prefer that, but I don't. Septentrionalis PMAnderson 02:52, 28 October 2008 (UTC)

• • Vague is as good as useless. You seem to be implying that the existence of rules equates to bullying. I see a consensus, but with isolated pockets of resistance. Having clear rules in such a context cannot be considered bullying, surely. Ohconfucius (talk) 02:56, 28 October 2008 (UTC)
    • Vague is as good as useless. Not among civilized editors, who can take a hint. But that is the case for silence. Septentrionalis PMAnderson 03:15, 28 October 2008 (UTC)

If there is no consensus - and in the sense I have outlined there is none here - then any definite rule will be opposed by many editors. A rule so opposed cannot be imposed without bullying. I have said no more. Septentrionalis PMAnderson 03:10, 28 October 2008 (UTC)

Val and delimitnum.

I've improved {{delimitnum}} a bit. It now supports the most numbers it can and a lot more leading 0s. See [20] for details. It is still subject to the rounding problems and ending 0s however.Headbomb {ταλκ – WP Physics: PotW} 18:16, 29 October 2008 (UTC)