- How much of a huge, bold, flashing editnotice would be enough to convince people that this section is about autoformatting, not linking? Or do we need an AbuseFilter warning on the word "link"? Anomie⚔ 00:12, 30 March 2009 (UTC)
- No-one above appears to have conflated the two. Ohconfucius (talk) 01:56, 30 March 2009 (UTC)
- A majority of the "oppose" !votes appear to have conflated the two. — Arthur Rubin (talk) 02:07, 30 March 2009 (UTC)
- To quote someone I know: "I wasn't going to reply unless an idiotic argument was presented by someone. Congratulations." Ohconfucius (talk) 03:21, 30 March 2009 (UTC)
- Wow. I am currently seeing 3 who mention linking in the rationale as if it was implied by autoformatting, but also give unrelated reasons (Awadewit, Bishonen, Bzuk), 3 who give no rationale (Donald Albury, Juliancolton, Chrishomingtang), and 29 who give a rationale that doesn't suggest they are confused about this. Since when is 3:29 a majority? Perhaps you are confused and think that this is about autoformatting without markup? Then read above under "What is date autoformatting?" It is not presented as an option. --Hans Adler (talk) 08:26, 30 March 2009 (UTC)
- Arthur Rubin is a mathematician, so I'm sure he knows what "majority" means.--otherlleftNo, really, other way . . . 03:38, 31 March 2009 (UTC)
- And why must I give yet another rationale when so many have already been given. This is not something like AfD, where the weight of arguments invoking policy should be decisive, this is a preference poll. -- Donald Albury 12:34, 30 March 2009 (UTC)
- I've stated my opinion several times in the past, so I feel there's no need to provide a rationale. This is, as the title suggests, a poll. –Juliancolton | Talk 19:22, 30 March 2009 (UTC)
- Agree that this is a poll and tons of reasons have been given by others already.—Chris! ct 20:24, 30 March 2009 (UTC)
- Sorry, I meant no offence and no rationale is of course perfectly OK. The only problem is that in a situation where some editors give rationales that prove they are confused ("I hate autoformatting because I hate the sea of irrelevant blue links") with no clue how they would have voted if they were not, every vote without a rationale will potentially be discarded by those who don't like the result of the poll.
- Ah, ye of little (good) faith. -- Donald Albury 16:11, 31 March 2009 (UTC)
- I'm concerned about this too. If the proposal is defeated, so be it, but it would be a real shame if it were defeated because a large number of editors were still under a wrong impression of the implementation details. Tim Pierce (talk) 13:13, 30 March 2009 (UTC)
- Why is it that so many people seem to think that marking up millions of dates requires a great deal of work, but removing markup from those dates is no work at all?-Jeff (talk) 01:09, 30 March 2009 (UTC)
- Bots can easily recognise and remove double-square brackets around dates. Adding date coding (especially complex coding) must be done contextually (instance-by-instance). HWV258 01:53, 30 March 2009 (UTC)
- Actually, I'd bet that a lot of the basic markup could be done by bots, if all they have to do is recognize dates and enclose them in something trivial. The actual choice of what format to use would need human intervention, but that could be separated from the markup around particular dates. — Gavia immer (talk) 02:53, 30 March 2009 (UTC)
- Agreeing with the above. Bots should be able to easily recognize dates, marked up or not, through the use of regular expressions.-Jeff (talk) 03:13, 30 March 2009 (UTC)
- (To the above two posts) The edit comment of "Most of the work could be bot-accomplished" is the key. How is most defined? Also, "The actual choice of what format to use would need human intervention" is a whopper in terms of slowing down the process. Anyhow, doesn't change the basic point that bot-removal of date formatting is trivial (to get back to the original post). HWV258 03:23, 30 March 2009 (UTC)
- If you separate the recognition of formattable dates from the format control, it's no big deal - just add a {{DEFAULTDATEFORMAT}} (or whatever name) parameter that works like {{DEFAULTSORT}}, appearing once per article to control the default date display format. Then you can add separate markup to actually set off dates, with an option to set the default display format for that one date. The sticky bit is that dates in direct quotations shouldn't be autoformatted, so a bot solution would have to recognize when to skip marking those up. A bot that got this even 90% right would leave very little work for human editors to slog through. — Gavia immer (talk) 04:20, 30 March 2009 (UTC)
- With your response above, you've confirmed the validity of my initial response to the original question (you do remember the original question?). Please take this up on a talk page somewhere—we've been discussing these sort of issues for months now. Thanks. HWV258 06:10, 30 March 2009 (UTC)
- I guess we need to add to the bold flashing editnotice that "unregistered users see inconsistent dates" can be easily fixed, so opposing on that basis is rather … misguided. And edit wars would occur just as often in the absence of a magic word with autoformatting, but they'd be worse because the warriors would be messing with dates all over the article (and possibly missing some each time). Anomie⚔ 02:48, 30 March 2009 (UTC)
- Prevention of edit-warring? The notion I see raised above by one or two editors that autoformatting is needed to avoid edit-wars over which format is chosen for an article is, I believe, barking up the wrong tree. Yes, apparently the original system was a response to friction on this matter, but 2003 was early days for the community, and we had established proper rules for neither date formatting nor ENGVAR spelling. We now have well-established practices for both (MOSNUM, MoS), and they are highly successful, by all accounts. Tony (talk) 08:41, 30 March 2009 (UTC)
- And yet still not as simple as letting editors enter dates in whatever format they like and letting the system auto format them to what readers prefer... at some point in the chain (editing/reading) you'll have someone using or seeing a date format they don't prefer when it needn't be that way. —Locke Cole • t • c 10:56, 30 March 2009 (UTC)
- With regards to these "well-established practices" for date formatting, if they are really "highly successful", why has Wikipedia's Chief Technical Officer called for a rewrite of that guideline to use only one consistent format site-wide? --Ckatzchatspy 11:07, 30 March 2009 (UTC)
- I am on record as wanting a mono-format for dates. Unfortunately, I am in the minority. It's not going to happen, because people believe WP:ENGVAR works, so we have to live with it. Having date-autoformatting could be likened to slicing off part of your feett to fit the new undersized shoes you just bought (in other words, don't make a mistake to compensate for a dumb move. Ohconfucius (talk) 13:51, 30 March 2009 (UTC)
- Note that at the time of my comment above, there was only one person alluding to autoformatting as a solution for edit warring (and that in a vague, "it would have helped in this one case" way), while there was one "oppose" specifically claiming that autoformatting with a magic word for setting a default format for the article would directly lead to edit warring over the default format setting with the implication that there would not be such edit warring otherwise. I agree that autoformatting is not needed for edit war prevention; although it would likely prevent some would-be edit warriors from ignoring or fighting over how to apply WP:ENGVAR, I personally regard that as a side effect rather than a major driving reason. Anomie⚔ 12:05, 30 March 2009 (UTC)
- (In response to Neutral #1 above) Would something along the lines of "{{d|30 March 2008}}" be simple enough? Keep in mind that the function "#formatdate" can easily be called from a template with a much simpler name. --Ckatzchatspy 06:26, 30 March 2009 (UTC)
- Considering the millions of times such a template would be called, that would be an enormous waste of server resources; but we could ask that the #formatdate function be renamed to #d, for example. --A. di M. (talk) 08:48, 30 March 2009 (UTC)
- Expansion on my rationale. One of the arguments for autoformatting is to "present a consistent date format". I think we should have consistency, but I don't believe autoformatting is the way forward. Personally, I would love for every article to use the fairly international style of day before month, but I fear that people like arguing about the pointless stuff too much for that to ever happen. Rambo's Revenge (How am I doing?) 13:36, 30 March 2009 (UTC)
- Isn't that an argument for auto formatting? Here we have a system that removes the need to argue (ever!) over which format to use, and it's a simple software solution. —Locke Cole • t • c 13:44, 30 March 2009 (UTC)
- No it is not. If IPs—the majority of Wikipedia readers—cannot choose their preference then autoformatting is not a good option. The only way to prevent IPs from getting a horrible mess of different formats is to choose what they see. If we do that, we might as well choose the style for everyone or completely standardise all dates into one format. Rambo's Revenge (How am I doing?) 14:23, 30 March 2009 (UTC)
- Part of the current proposal is that there would be a Wikipedia-wide default format setting (most likely DMY) for everyone who has not set a preference, including IP users, and a magic word so a particular article can change the default (i.e. to MDY) when that is appropriate per WP:ENGVAR. Then user preferences override the default for that user. This has been said time and time again, which is why can't in good faith understand why opposers keep claiming IP users will see some sort of mish-mash. I also find a claim that every IP user has to be able to set a date preference spurious, as we have an easy way for anyone to set any preference: register an account. Anomie⚔ 15:05, 30 March 2009 (UTC)
- But the point you are missing is that in order to maintain consistency for all users, every date on a page needs to be coded (in order to be rendered properly based on various preferences). The problem with "every date" is that it is extremely difficult to define rules needed in order to detect all the different types of date formats found on WP. Date ranges and slashed-dates are just two examples, but also difficult is to precisely detect the comma in US date formats. These issues remain unaddressed—after months of debate, and lots of examples demonstrating the problems can easily be found. Many people voting "support" are unaware of the technical issues involved. (Incidentally, I don't blame them so much as these issues are not easily grasped by people who have not been involved with the debate for some time.) HWV258 00:46, 31 March 2009 (UTC)
- Hi, I'm a computer programmer. Are you? The reason I ask is that you assert that "it is extremely difficult to define rules needed in order to detect all the different types of date formats found on WP", which strikes me as a statement that would be made by someone who doesn't actually know what they're talking about. There is no real "detecting" necessary; someone just needs to come up with a syntax for specifying date ranges, slashed dates, trailing commas, and the like. For example, {{#formatdate:1 January 2009/2 January 2009}} could be the format for a slashed date, and output as "1/2 January 2009", "January 1/2, 2009", and so on based on the active format; {{#formatdate:1 January 2009–10 January 2009}} or {{#formatdate:1 January 2009|10 January 2009}} could be the format for entering a date range, and output appropriately (there could even be a preference for "1 January 2009 – 10 January 2009" versus "1–10 January 2009" style output, if people wanted it). With a little more effort, any of the (unambiguous) output formats could also be accepted as input. The need for the trailing commas could easily enough be specified as {{#formatdate:1 January 2009|,}} or {{#formatdate:January 1, 2009,}}. Something else that I personally would like to see is a "{{#formattabulardate}}" function, so people who would prefer to see dates in tables and lists as 2009-01-01 versus 1 Jan 2009 versus the full 1 January 2009 could set a preference for that. It's not particularly hard to do any of that, but why should someone bother when the discussion is full of people (not necessarily you, don't take this personally) who will oppose everything without end, drowning out any other discussion, to the point of WP:POINT? Anomie⚔ 12:26, 31 March 2009 (UTC)
- Actually, I am a programmer (and your question indicates to everyone that you haven't been following the debate over the previous few months). I was the only one who actively push to get a specification for auto-formatting established (here), however nothing came of it (despite a myriad of suggestions in various locations—similar to yours above). You do understand that all your examples above simply will be thrown into a pot—a pot that is already full to the overflowing with similar suggestions—however a pot that has so far failed to provide anything nutritious (or in the least bit edible) for the community. As you have not responded to my point "every date on a page needs to be coded" I'll assume that you have understood that, and are in favour of it. Please consider a page such as this one that has over 700 dates (in many different formats), so to apply formatting to all dates (similar to {{#formatdate:8 January 1705}}) is quite an undertaking (an undertaking that has had no analysis in terms of viability). The people you mention "who will oppose everything without end" are actually people who have thought through all these issues and have come to the considered conclusion that simply entering dates using plain text solves all significant issues, and has no syntactical complexity for the wider editing community. In addition, I belong to the group of programmers who believe it is inappropriate (and downright unprofessional) to commence coding without (at least a functional) specification. A large part of the reason for the mess we are currently in is the (well-intentioned) introduction of code that had no specification, let alone community consensus. HWV258 23:19, 31 March 2009 (UTC)
- Of course I haven't followed the debate, certain people have made the situation akin to diving for pearls in a cesspool. I don't understand how you can claim it's so difficult for code to solve the problems you raised in your paragraph above, though, since you state that you are familiar with programming. I also don't see any particular problem with List of compositions by George Frideric Handel; sure, it might need a bit of human attention to get everything marked up, but I see nothing in there that would cause any trouble for a well-written formatter; or is the whole of your concern there that someone would have to do a little bit of formatting work? As for those "who will oppose everything without end", either we're talking about different people or our views are so opposed that we will never reach an agreement on the matter. Anomie⚔ 02:10, 1 April 2009 (UTC)
- I didnt' claim "it's so difficult for code to solve the problems"—rather it is looking increasingly likely that it is impossible to specify the problem in the first place. HWV258 04:18, 1 April 2009 (UTC)
- How is it impossible to specify what the a bot would do? The bot would find plain dates, pre-existing or entered by new users, and convert them to a standard form inside an autoformat bracket. Regular expressions already exist for finding dates in widely used forms. If the date was ambiguous, such as 07-03-02, it could inform wikignomes via wiki: page or any other method, who could manually fix it, helping in the already existing task of removing ambiguity from Wikipedia. The scope of the bot and implementation seem completely clear, maybe 5 hours coding tops on an existing framework.Gsonnenf (talk)
- I think the bot would have to be quite conservative. Spend an hour or so trying to devise a syntax for date formats. After doing so, how to work out ambiguous dates— or at least discover that they are ambiguous and mark them so. Then partial dates such as "April 2009" or "2009". (oh, whoops, it 2,009 kilometres (1,248 mi)). Then quoted dates which must be left alone. Then perhaps the French Revolutionary Calendar. It is *not* trivial. I am in support but I think it is best left to human markup rather than a bot. For sure, have a bot gather the info after it's been marked up, I'm all for that. But not guess what is, or is not, a date. SimonTrew (talk) 03:41, 8 April 2009 (UTC)
- I agree the bot would need to be conservative. As I understand it partial dates aren't used with date autoformat, please correct me if I'm wrong. When you say "Spend an hour or so trying to devise a syntax for date formats." do you mean thinking of how to have the parser recognize them? If so, this problem has already has a published solution. Nested quotation recognition is also a solved parser problem. The logic code for determining ambiguity of format in recognized dates is simple as a couple if/then/and statements. This still looks like a simple task, I'm uncertain as to what you mean by "trivial" as it has different meanings in different science fields. I could see how a new coder or one without exposure to modern parsing libraries or parsing exposure may see this problem as complex. With proper knowledge, which I'm sure many coders on Wikipedia have, the solution is simple to implement. I would prefer conservative date changes by bots with human confirmation. This would ease the burden and time constraint on humans. Manually verifying a date on a page would probably take 5-10 seconds, where in going through the sequence of changing them could take about 5 to 10 times as long (depending on server conditions).Gsonnenf (talk)
- "How is it impossible to specify what the a bot would do?"—this is wandering from the original point (i.e. I didn't mention bots). Please re-read my points. HWV258 02:00, 9 April 2009 (UTC)
- I re-read your points and find them moot. The vast majority of date forms can be easily recognized. I disagree with your statement that "every date on a page needs to be coded". Date forms that are intentionally abiguious or not well formed (e.g. year 197x ) should either be clarified manually or left alone. Autoformatting would only need to be applied to well formed dates. Any date that can be recognized unambiguously, (jan 16, 1973, 10/22/1001) could be easily recognized by both humans and parsers. So in this case it is not longer "extremely difficult to define rules needed in order to detect all the different types". The comment "Date ranges and slashed-dates are just two examples, but also difficult is to precisely detect the comma in US date formats" is not correct, these can be easily identified using a parser and regular expressions. The above comment is also very suggestive of a bot, which contradicts your statement "I didn't mention bots". Unless of course you meant that humans would have a hard time "detecting the comma in US date formats", which is complete nonsense. So as it turns out my above argument does not wander from the original point. Unless of course you wandered from the original point in your above statements. If that is the case, please present your original point and present an argument that does not "wander from the original point".Gsonnenf (talk)
- "I disagree with your statement that "every date on a page needs to be coded"" leads me to believe that we are talking about different things (and your disagreement probably indicates you have not been following the debate over the previous number of months). The point about coding all dates is that it is impossible to guarantee date rendering consistency in an article that contains at least one coded date—but with not all dates being coded in the article. It has been clearly understood by all that you either code all dates in an article, or you code none. If you are suggesting that coding isn't necessary because a page pre-processor can parse (and reformat) all dates in real time—then you are probably the first to do so. These current RfCs are not to do with bot activity (that debate will come after these RfCs are complete). HWV258 00:16, 12 April 2009 (UTC)
- It sounds like your suggesting that "It has been clearly understood by all" that dates containing (in no particular order) Months/day/year and month/day and all other forms, must be subject to autoformating. This is not consistent with the autoformating Background statement that addresses only MDY/DMY format under its definition "What is a date format?". The entire primary statements of Summary and Pro's only mentions "DMY/MDY" format. The con's section specifies incomplete dates forms of type MD as a possible extension to the proposal, "(double the number of keystrokes—even more if |dmy/md is added)", demonstrating its not part of the initial scope. So when you say "its clearly understood by all that you ... code 'all' dates", you are contradicting the primary background statement if you include partial dates in your definition of dates. It appears you are constructing an argument via false dilemma by asserting all people recognize that ALL dates, included partial dates must be encoded or none at all. This of course is not true as demonstrated by the primary statements on autoformatting.Gsonnenf (talk)
- Please read the history of the debate (over the previous many months). You are "arguing" in isolation and clearly don't have all the background information at your fingertips. All this has been covered, and I have no appetite for repeating here what has been covered (to death) many, many times before. If you need more help to understand the background of the debate, please take it up in a different forum—I'll be more than happy to attempt to bring you up to speed on the salient points. Cheers. HWV258 22:56, 13 April 2009 (UTC)
- The IPs would not get a bunch of different styles since the feature would provide a standard default for them. Also, developers have mentioned using Javascript to allow IPs to set a preference, though no developer has yet worked on that solution.—Ost (talk) 15:06, 30 March 2009 (UTC)
- *sigh* If you actually read my last post I never claimed IPs will get a mish-mash of styles. I said that the only way to prevent them getting that is to choose their preference for them. I then said that there is no point in choosing a standard style for them, because it would need to be agreed upon. If we can agree a choice for that, we should just implement that choice as the date style across Wikipedia without autoformatting, because that would give real consistency. I am well aware of what is going on and am not fond of other users trying to twist my words towards their !vote. Rambo's Revenge (How am I doing?) 16:55, 30 March 2009 (UTC)
- You have presented a false dichotomy in assuming that we would have to pick one format for all of Wikipedia for IP users, with no possibility of overriding that where WP:ENGVAR calls for another format. So yes, you're not quite claiming "IP users will get a mish-mash of styles"; you're claiming "IP users will get a mish-mash of styles unless we get rid of WP:ENGVAR on this issue". Anomie⚔ 22:19, 30 March 2009 (UTC)
- (In response to Oppose#89 above) FYI, we are discussing that; the recent changes to the system now allow autoformatting to apply a default format for unregistered users, and there is a patch in the works (I'll try to get the Bugzilla link ASAP) that would add the "per-page" option you've mentioned. --Ckatzchatspy 21:26, 30 March 2009 (UTC)
- Um ... no, Josiah is talking about autoformatting in the sense of allowing a few privileged editors to select a month-day or day-month order. A consistent fixed-text format for IP users is the reality in the thousands of article that have dispensed with autoformatting. Tony (talk) 08:23, 31 March 2009 (UTC)
- Actually, you're both right. I oppose autoformatting as it currently exists, as a special option for a few editors. However, I would support autoformatting as something that could be applied universally with appropriate regional variations. For example, if a citation using the ugly 2009-04-05 format were used in two articles, one dealing with a US-based subject and the other dealing with a UK-based subject, it would appear (to all readers) in the US-based article as April 5, 2009 and in the UK-based article as 5 April 2009. I'd support an autoformatting system that automatically translated dates in that fashion, as long as there was a parameter that allowed for articles to use appropriate formats, in keeping with WP:MOSDATE. In other words, I'd support autoformatting as a method for producing consistency within articles, but I oppose autoformatting as a way to enforce consistency throughout the encyclopedia, or as a way to make dates appear consistent to a small minority of readers. —Josiah Rowe (talk • contribs) 06:06, 5 April 2009 (UTC)
- I'm in favor of the idea of autoformatting, but against the requirements it will make on editor and hardware. Debresser (talk) 16:50, 30 March 2009 (UTC)
- I'm with you. I think the server should do all the work, recognizing dates for what they are and regurgitating them for viewers to match their preferences. Editors should only have to make sure that dates are consistent throughout the article. I'm against using some kind of template or marker or formatting. Binksternet (talk) 06:01, 3 April 2009 (UTC)
- That is impossible: It would result in changing of dates in exact quotes which contained dates by default. You should never change an exact quote, but no software is clever enough to unambiguously spot a quote, even if put in by a newbie that formats it in a slightly unexpected way. Creating a way to bypass this is not a solution, because the mechanism to do so is going to be obscure, and most editors won't know it. It ruins usablity for editors, in order to allow a tiny minority of people to select which of two date formats they like. Shoemaker's Holiday (talk) 06:57, 5 April 2009 (UTC)
- At the present time (61 support, 91 oppose), support #42 and Oppose #3, #12, #13, #22, #42, and #73 think that it's talking about linking. A number of further comments seem think it's talking about things looking different for editors and IPs; which is (1) always true that editors can adjust viewing format, using gadgets, preferences, and Javascript; and is not necessarily true of autoformatting. — Arthur Rubin (talk) 00:09, 31 March 2009 (UTC)
- why are you interpreting Oppose #3 and #73 as "think[ing] that it's talking about linking"? i agree that there are a couple of !voters who appear to be confusing the two issues and a couple of "hanging chad !votes", but both Oppose #3 and Oppose #73 appear cognizant that linking and autoformatting are two different issues.
- either way it's not a high rate of confusion - that's good! and the "hanging chad !voters" still have time to clarify their views if they care to.
- meanwhile, it's right that some !voters are saying they don't want logged-in editors to see different content than unlogged-in users. you disagree with that viewpoint, but that doesn't mean their !votes "don't count". Sssoul (talk) 07:21, 31 March 2009 (UTC)
- Sorry; either I made a mistake; they changed their !votes, or the !votes got renumbered. Still, the purpose of this poll is to find WP:CONSENSUS, and I don't see even a clear supermajority for oppose (yet). As not wanting logged-in editors to see different content than unlogged in users, one might argue that it's a reason contrary to policy as defined by the developers. We would need to remove the gadgets and javascript customization. I'll post more on the talk page. — Arthur Rubin (talk) 16:06, 1 April 2009 (UTC)
- The "only for editors" claim has been bandied about throughout this debate over the past year, simply because it is an easy rallying cry for the three or four core editors who are really pushing the "delink" campaign. Thing is, there's never been any proof for the numeric claims (despite repeated requests), and never any acknowledgement of the readers who may have registered an account not to edit, but in order to access features such as date formatting, watchlists, and gadgets. (The "only for editors" mantra also fails to mention that IPs can't access any special perks, not just autoformatting.) --Ckatzchatspy 02:24, 31 March 2009 (UTC)
Good grief! how hard can it be?. The whole formatting thing is a no-brainer. Simple fact is, if functionality is provided for autoformatting it can be provided to most IP users too with simple preferences option for them via cookie system or countless other ways. Search engines, most news sources and millions of commercial sites have been doing it since Jesus was a small boy. Come down to it, autoformatting also allows the potential to do away with the linking argument completely because it can be controlled by a users preferences as well. Want to link dates? switch it on via preferences. don't want them linked? leave it off. And don't bore me with the non-logged in users rubbish - most of them a: don't care, b: have cookies turned on (or are unaware of their presence) or c: know how to allow for cookies by site. BOTs to do the bulk conversions to simple syntax (say YYYY-MM-DD or #dYYYY-MM-DD) and wikignomes will make short work of the rest. Default presentation for those who haven't set it (maybe with a bit of fancy footwork detecting where they are accessing from to feed them American or ISO format by default). Doing something is unlikely to make it more inconsistent than it currently is.--ClubOranjeT 08:08, 3 April 2009 (UTC)
- Has anyone thought of having a bot do autoformatting? That would nearly neutralize editor requirements, just invest some effort in the bot, and sit back and watch! Oldlaptop321 (talk) 23:58, 6 April 2009 (UTC)
- I think it is accepted on all sides that there will be bots and templates to assist (and hopefully not to destroy). If it is relatively relaxed and gets it right most of the time, like Template:Coord or Template:Convert, I don't see it being a big problem. But like those, you don't *have* to use them, hence my support; they are nice things to have. In a way I can't see why this even goes to a vote since I can't think of much of a reason, beyond finding user preferences, why this can't all be done in a template. SimonTrew (talk) 01:46, 7 April 2009 (UTC)
- Actually now I am wondering whether this is in the purvue of Template:Convert. A date is only a measure after all, and its means of expression is its unit of measure. That's probably outside the scope of this vote, though. It's an interesting question to ask those in the "Oppose" camp; do you want to get rid of Template:Convert, even if you don't use it on every single measure you write? SimonTrew (talk) 01:50, 7 April 2009 (UTC)
- SimonTrew, this hasn't gone to a vote. We don't vote here. See the "!vote" terminology people are using? That's why. Also, I don't agree with your comparison of this present issue to Template:convert. The similarities are superficial and few, as it seems to me. —Scheinwerfermann T·C02:32, 7 April 2009 (UTC)
- OK, so I suppose that's why it says at the VERY TOP of the article Please indicate your vote under ONE option. I never claimed it was a formal, binding, vote.
- The similarities are superficial and few precisely because time is just another dimension that can be relatively easily incorporated into the Convert template, or if not, another template can be made analagous to it. And I have worked designing and implementing very complicated units of measure conversion systems, and I know they are not easy. SimonTrew (talk) 02:51, 7 April 2009 (UTC)
All Support entries that are founded on reader choice should be disregarded when finalizing the findings of this RfC. What reader choice? The reader gets the site default, or the contributing editor's choice. Remember that registered users = subset of editors, and in turn editors are a tiny subset of readers. All Support users who clamor that the reader need only edit a setting should be deliberateley ignored by the rest of the community for a period no shorter than the time it takes defeat any and all proposals in support for auto-formatting. --Karbinski (talk) 21:38, 9 April 2009 (UTC)
- Sorry, but that's not valid. There is no requirement that you must edit if you register; anyone - anyone - can register an account, and if they choose to use that account to read only, so be it. Accounts can be useful to readers who wish to use the gadgets, the watchlist, different skins, and so on. --Ckatzchatspy 21:49, 9 April 2009 (UTC)
- Sorry about that, but if we leave out the middle, it is correct: Remember registered users are a tiny subset of readers. Not that it has bearing, but registered users are also a subset of editors. Karbinski (talk) 21:57, 9 April 2009 (UTC)
Many of the contra arguments speak to the issue that most readers are anons. I know from experience that people assume the only reason to register for an account is to make edits. If there were tangible benefits (like getting standardized dates, watchlists, and so on) and those benefits were promoted to readers, we might get more editors. And once they have an account, they might be more obliged to make an occasional edit. And that could lead to editing as a hobby, until they are fully assimilated into the collective. (Oops! Been watching too much Star Trek perhaps?). The point is, anything that is good for editors is ultimately good for the readers, too. --Willscrlt (→“¡¿Talk?!”) 14:14, 13 April 2009 (UTC)
Responses
- Several users have indicated that autoformatting is the only way to achieve consistency
through in one article, and this is simply not true. EDIT: This may be true for for achieving site-wide consistency, but that begs the question of why the way the dates look is such a big deal. We can all look at DMY or MDY or even YMD (2000 January 1) and fairly easily determine what the date is.
- User:This flag once was red, using the ambiguous 3/2/03 type format is not an option anyway.
- I think you've misunderstood me (or I've not been clear) - my thinking is that by using a template for dates it will discourage the entry of dates as DD/MM or MM/YY, not that dates will no longer be displayed as DD/MM or MM/YY. I agree that this problem isn't going to go away completely, but it should be alleviated as editors get into the habit of specifying dates in terms of {{... day=2|month=3| ...}}, instead of just typing "2/3/09" and hoping that whoever cleans up after them magically knows whether the date is DD/MM or MM/DD. Cheers, This flag once was redpropagandadeeds 12:53, 31 March 2009 (UTC)
- "Before autoformatting was introduced, there were lengthy rows about how to format dates. This seems recently to have come back, just as some started delinking dates." from docu (talk · contribs); can someone point me to these length rows on date formatting? I haven't seen any. Not referring to linking here.
- "Also, autoformatting helps prevent edit warring" A an over-used argument with little evidence provided. Actually, these edit wars are few and far between.
- To the users who cited metadata as a reason for needing autoformatting, can someone provide examples of how this metadata could be used? Dabomb87 (talk) 02:16, 31 March 2009 (UTC)
- Other than imposing a site-wide single format, as the developers have suggested, how do you propose to ensure all articles are consistent with one another? Either we go to a single standard, or we persist with the first-past-the-post "this is American no it's international" methodology. If the latter is the case, it is only prudent to provide some method of presenting a uniform, consistent look. --Ckatzchatspy 02:28, 31 March 2009 (UTC)
- Sorry, I'll be more clear. I edited my above post. Dabomb87 (talk) 12:47, 31 March 2009 (UTC)
Most of the arguments against autoformatting seem to boil down to "this would be a lot of work, and there is no advantage in it for me". I fail to see how this is an argument for forbidding those who do care about consistency from doing the work. Why does this have to be framed in terms of either "everybody must use this markup" or "nobody is allowed to use this markup"? That's not the wiki way. The wiki tradition would be to encourage people to contribute even if they can't be bothered to use the more complex markups, and then let others insert advanced markup if they care to.
What active harm do all of you opposers suffer because someone else adds autoformatting? It won't require you to lift a finger either way. –Henning Makholm (talk) 11:20, 31 March 2009 (UTC)
- This is an important point. Why cant autoformatting be implemented and later have a discussion about whether the MOS should encourage its use? The first section to this poll should have been left until after implementation when users can see what it actually involves. I seem to remember that {{cite...}} came about this way, and is probably more difficult for unregistered users to edit and breaks up the text far more than this proposal. |→ Spaully₪† 12:49, 31 March 2009 (GMT)
- Yes but {{cite...}} provides an important - one might even say essential - feature for an encyclopedia. This more than justifies the cost in terms of difficulty of use and so on that the markup introduces. Date autoformatting on the other hand provides at best an extremely minor cosmetic tweak. Most people don't care what formats their dates are written in. They will even use a variety of formats in their own writing or speech without a second thought. The number of people who are actually annoyed by reading a date in a format they don't like must be extraordinarily small. These pedants are considerably outnumbered by those like myself who find needlessly complicated markup annoying, or who are annoyed by the excessive linking and sometimes anomalous results created by poorly implemented date autoformatting. Hawthorn (talk) 23:38, 31 March 2009 (UTC)
- The number of people who are actually annoyed by other editors inserting date markeup must be even extraordinarily smaller. Even if you yourself find it annoying to enter such markup, I don't see how it can conceivably annoy you to let other people do it. Nobody is proposing to say that you should enter the formatting yourself if you don't feel like it, but how can it possibly disturb you if other people do so? Again, the proposal is about NON-LINKING autoformatting, so there will be no excessive linking to annoy you. –Henning Makholm (talk) 14:35, 1 April 2009 (UTC)
- Henning Makholm seems to be overlooking the fact that anyone who wants to edit an article has to deal with the markup other people have inserted - even if "deal with it" only means "figure out how to edit 'around' it". it can be extremely daunting. many people are saying that date-format preferences are not important enough to justify even the double-bracket markup it now uses, let alone adding even more markup (a la {{formatdate|12 December 1981}} to every date in the encyclopedia.) Sssoul (talk) 07:08, 2 April 2009 (UTC)
- Once the formatting is there, it is not daunting at all. On the contrary, it is extremely self-documenting how the markup you quote works, and I cannot imagine anybody using more than a quarter second thinking "hm, it says formatdate, wonder what that means ... ah, flash of brilliance, it probably is something about formatting the date that comes after.". Now, I can understand not wanting to go around remembering how the markup works, and thus not wanting to enter it from scratch, but really, anybody who have trouble noticing and editing around a markup as clear and unambiguous as the one you show probably wouldn't be able to stitch together a coherent sentence for the encyclopedia in plain prose.
- The double-bracketing is annoying because it creates needless bluelinks in the output that are hardly ever relevant in context. That is annoying. But a small bit of really, easily, self-documenting markup to make live easier for the (perhaps few) people who care about it? Not at all. –Henning Makholm (talk) 11:32, 2 April 2009 (UTC)
- okay, so you don't find it daunting or annoying to have to wade through markup that you perceive as unnecessary; other people do find it daunting and/or annoying. it's one of those "different opinion" things, and saying "no it isn't" doesn't change that. peace Sssoul (talk) 07:37, 3 April 2009 (UTC)
(In response to support vote # 82) Sorry UC_Bill, but you can't have it both ways. You removed the demo page that could have shown the community how date ranges (and hopefully date slashes) would have worked, and then you claim the programming is easy ("10 hours"?). Please re-establish the demo page if you really have got it going, otherwise you'll have to understand why I'm entitled to treat your claims above with a healthy dose of scepticism. To other voting parties, please note that what is proposed is far more than "four brackets ([[ ]]) around dates" (as even the post before UC_Bill's indicates). I'm not suggesting that you (are you back?) or perhaps other programmers couldn't get something going to handle individual cases, however as has been demonstrated over more than three years, it is increasingly difficult to not only get the basics of auto-formatting right, but seemingly impossible to specify how auto-formatting should work for the many ways that dates are represented on WP. HWV258 21:57, 31 March 2009 (UTC)
(In response to oppose vote # 138) "I have never come across a web information source or news provider that worries about this enough to give readers an option as to how to display the date"—well spotted; a very good point. In terms of "MMM DD, YYYY all round", there is no need to be so dictatorial, as very often an article will lend itself to "DD MMM YYYY", in which case, that can simply be the format of choice for the page in question. Is it that much of a worry to see dates in the "DD MMM YYYY" format in the England article and at the same time to see dates in the "MMM DD, YYYY" format in the USA article? HWV258 23:33, 31 March 2009 (UTC)
- It's no worry at all to me, as I am from a country (the UK) where both formats are used interchangeably. If you reread my post above (oppose 138) you'll see that The Times, regarded for a long time (at least before Rupert Murdoch bought it) as the UK's newspaper of record, uses MMM DD, YYYY, so why does the Wikipedia article about England lend itself any less to that format than the other? It's a myth that there is any consistent standard outside the United States, so why not just follow the standard that is used exclusively in civilian life there, and used interchangeably with DD MMM YYYY everywhere else where English is used? Phil Bridger (talk) 00:24, 1 April 2009 (UTC)
- I'm not sure if I'm following any more (my confusion above started with the word "there" in "so why not just follow the standard that is used exclusively in civilian life there"—where is "there"?). Are you suggesting that the England article should use "MMM DD, YYYY" date formatting? If you are, you should be aware that that's far more radical than anyone else has dared suggest in this debate? As an aside, if a global format is being suggested, I'd lean towards using the less syntactically complicated "DD MMM YYYY" (you know, the one that is used at the end of each of these posts). HWV258 01:04, 1 April 2009 (UTC)
- Sorry if I was unclear, by the word "there" I meant the United States. And once again, what is radical about using MMM DD, YYYY for the England article, when both formats are used interchangeably in England? Everyone seems to be assuming that there is a strong preference for DD MMM YYYY outside the United States, but, as I demonstrated with my links above (oppose 138), this is not true. Phil Bridger (talk) 08:54, 1 April 2009 (UTC)
- It's simply not true that they are used interchangeably. If you ever have to fill a form in the UK, it's quite possible that you have to enter a date into fields that look like this, __/__/____. In my experience they will always expect you to use DMY format, even though in some cases they may not mention it because it really goes without saying. This is because DMY format is much more common in general, including when the month is spelled out. --Hans Adler (talk) 15:13, 1 April 2009 (UTC)
- In forms the date is usually entered in all numeric format, so there obviously has to be a standard, and, yes, that standard is different in the UK and the US, but in other contexts where the month is spelt out both formats are common. Do you seriously think that people have difficulty understanding the date, or get any sense that there's something wrong with it, when thay read it in The Times or The Financial Times or The Daily Mail or The Sun], or The Daily Star, all of which say that it's "April 1, 2009", not "1 April 2009"? That's five of the ten national daily newspapers, so you couldn't get a much more even split. And I have given similar references above to show that both formats are used in several other countries where English is widely used. It's no good saying that it's "simply not true" that both formats are used when the evidence says that they are. And why is it, when we are supposed to base this encyclopedia on reliable sources, that I am the only one of the 248 people who have voted here so far who has bothered to provide any evidence? Phil Bridger (talk) 15:45, 1 April 2009 (UTC)
- Sorry for absurdly lecturing you without seeing that you are actually from the UK. I agree that standardising on MDY would be an option. However, this is very similar to standardising on Oxford style "ize" spellings for verbs in British English, and it seems that we haven't adopted this either. Therefore it seems unlikely that your argument will convince many in this poll. I think it should really be discussed separately. --Hans Adler (talk) 16:05, 1 April 2009 (UTC)
- I'd just like to concur with this point - whenever I see this debate crop up, I honestly find it hard to remember which variant I'm supposed to prefer, or which one I'm supposed to object to. Written numerically, there's a standard, but written textually then "1st February" or "February 1st" are pretty much interchangeable to a British reader, and I'll often use the two in the same piece of writing, or use one one day and the other the next. Shimgray | talk | 15:29, 3 April 2009 (UTC)
- A few of the above UK newspaper examples don't use American date format consistently when digging into the news articles. While the top banner of The Sun may show MDY, DMY-based content such as "Published: 04 Apr 2009" appear in articles before the current date ; here's a "13th April 2009" from The Daily Star, and another 13th April from The Daily Mail. Such commercial websites may not be reliable indicators for date format anyway as these can sometimes be skewed by software defaults, often biased towards US usage.
- Government standards and practice would seem more reliable to consider for this discussion. UK government usage consistently uses DMY and rejects US/MDY format. There are moot differences as to whether ordinal letters (12th vs 12) are used on day numbers, but there is no interchangeability of ordering on official UK levels e.g. TDA, Cabinet Office, Hants, Mansfield District Council.
- Furthermore, DMY is generally used by international bodies such as the UN , OECD , ITU , ICRC . That's compelling evidence of what date format is considered the norm on an international scale and where a global project such as WP should be headed. But if we can't get there in the short term, we should make good use of the autoformatting capability (already developed for the most part) for the benefit of the various audiences. Dl2000 (talk) 16:41, 13 April 2009 (UTC)
Silly autoformatting:
Uladzimir Katkouski, a well-known and award-winning Belarusian blogger, editor of several Belarusian websites and activist for the usage of the Belarusian language on the Internet. He made 1343 contributions to the English Wikipedia and 1286 contributions to nine other Wikimedia projects. Katkouski mostly edited articles about his country and his language. He was hit by a fire truck in 2006 and, after being in a coma for about a year, he died on May 25, 2007.
Context on link misconception
- To provide some context about the confusion, right now I count date linking arguments as a given justification on the following votes (assuming #s don't get switched by the time you read this): 12, 13, 22, 41, 42, 49, 73, 93, 101, 108, 109, 110, 114, 116, 134, 135, 137, 140, 145, 147. Currently that is 20 out of 159 votes (~12%). Also, neutral comments 3, and 8 share these same misconceptions. Who knows how many other votes, on both sides, share this misconception but do not mention it explicitly.
- Only a few of these comments make arguments against based on issues other than linking. The vast majority of the ones I've cited base their given explanation soley on the linking proposal. Shadowjams (talk) 01:20, 2 April 2009 (UTC)
- (1) Voters are under no obligation to provide all of their inner reasoning. Are you objecting to those who provided no comment? (2) It is no surprise that many people still refer to the concept of date autoformatting at "linking"—that has been the vernacular term for the concept for some five years. (2) There is reference to "the links" in the Statement for the concept of autoformatting, and to "new features" such as "database dumps" that would, of course, require a feature for editors to identify autoformatted dates as links. Why wouldn't some voters use the common synonym "linking" to refer to the concept of autoformatting. Tony (talk) 09:58, 2 April 2009 (UTC)
- Because it's a poll and not a vote, reasoning is instructive as to the actual proposal's support. Even the most adamant opponent should not want people to be misinformed about the nature of the proposal (or if they did, they could never vocalize it), and should be disturbed when there's evidence a number of users are confused.
- The "linking" shorthand is interesting and something I hadn't heard before. You could be right and some users are using the old terminology out of habit. But certainly some users, the one that mention "blue links", or describe it, believe the proposal will lead to date linking again.
- Of my original list (I'm working off the original numbering) numbers 13, 22, 42, 49, 73, 93, 101, 108, 109, 110, 116, 135, 137, 145, 147, either make the distinction themselves, or indicate it by referring to the particular properties of links (blue, they go somewhere, etc.).
- Your point also suggests that perhaps some users who call things autoformatting, without mentioning linking, are thinking of linking. Maybe my original census undercounted the number of misinformed editors. Shadowjams (talk) 18:05, 2 April 2009 (UTC)
Vote by User:Neurolysis
I note that User:neuro returned and added:
- Sleeping on this, I think I made this comment too quickly.
That looks to me like a retraction. Lightmouse (talk) 18:06, 1 April 2009 (UTC)
- Yes, it was. — neuro(talk)(review) 23:48, 1 April 2009 (UTC)
Make the server do it while assembling the page. Why are human editors concerned with this question? Why can't we make the server search for dates in the text and then reformat them for the viewer based on their preference? The server already changes <br>, </br> and <br/> to <br /> without anyone ever noticing. I think dates should be the same; that editors should just type a date and the server will handle how it's served up. There's no reason why bots or humans should have to go through articles and apply templates or formatting. Binksternet (talk) 15:39, 3 April 2009 (UTC)
- The problem here is that we don't want to reformat a lot of dates - in direct quotations, for example, or in the titles of cited works, both of which are places where we leave even broken spelling intact. Were we to let this sort of autoformatter loose, we'd need to have a simple way of marking those as not-to-be-changed... Shimgray | talk | 19:53, 3 April 2009 (UTC)
- There is a simple way to say to the server "don't touch this particular date"... we can use the nowiki tag. Binksternet (talk) 13:12, 4 April 2009 (UTC)
- Or indeed any other tag or template to be defined. But while I support autoformatting, it seems to me to be better that tags/templates are used to included it (i.e. an optional extra) rather than exclude it. That way, it's not a breaking change. Of course, editors or those who look after particular projects might choose to run a bot to automate "upgrades" to use the new autofomatting, but it should not be the default behavior of the server.
- Perhaps I have got Binksternet's comment wrong, but if the suggestion is that the server should be able to parse text and work out that a particular piece of text is a date, that is a very hard task. I'd suggest that some markup is inevitable. Also, I am not sure if the server understands different languages? Parsing language-specific text surely is the job of a particular Wiki's template not the server core. SimonTrew (talk) 16:14, 6 April 2009 (UTC)
- Perhaps I've also misunderstood the original comment, but just in case the suggestion is to parse dates that don't have some sort of mark-up wrapped around them, have a look at the following example: "On Jan 1, 2000 Wikipedians went crazy". Is that: all Wikipedians going crazy on 1 Jan 2000, or is that 2000 Wikipedians going crazy on 1 Jan? HWV258 22:42, 8 April 2009 (UTC)
- What about adding a freakin' comma after the year if the first of those two meaning is the intended one? --A. di M. (formerly Army1987) — Deeds, not words. 00:31, 9 April 2009 (UTC)
- Will you make a list of poorly formatted sentences on WP, or shall I? :-) HWV258 01:10, 9 April 2009 (UTC)
Autoformatting into ISO dates
One comment I see recurring among the opposes is that autoformatting is analogous to automatically handling US/Commonwealth spelling, and that we don't want that. Leaving aside whether we do want that (I don't), this argument misses another possibility - the ability to autoformat dates into ISO format. Very few human readers would want that (I would, but I'm a techy geek) but it could be extremely useful to automated parsers, particularly in, say, infoboxes. Cheers, This flag once was redpropagandadeeds 16:04, 3 April 2009 (UTC)
Autoformatting for non-registered users
The statement (Wikipedia:Date formatting and linking poll#Background statement) suggests that autoformatting is only useful to registered editors: "Autoformatting is a way of marking up dates to allow registered users to choose their preferred display format". However, is there any reason why autoformatting couldn't be performed for non-registered users? The user's locale could easily be inferred from the HTTP headers ("Accept-Language" in particular) and a suitable date-format used. Cheers, This flag once was redpropagandadeeds 18:17, 3 April 2009 (UTC)
- But how can we infer from the header what format the reader prefers? I have shown elsewhere in this discussion that "April 3, 2009" and "3 April 2009" are equally acceptable formats in the English language in the United Kingdom, India, Australia and Canada, and, if you really insist I'll link you to major English language media sources in Pakistan, Ireland, New Zealand, Bangladesh, South Africa, Sri Lanka, Nigeria and anywhere else that will show that it is a myth that "3 April 2009" is a nationally preferred format anywhere that English is widely used. The only country where there is a preferred format is the United States, so why not just knock this whole silly discussion on the head and use the format (April 3, 2009) that is acceptable everywhere that English is widely used. Phil Bridger (talk) 19:18, 3 April 2009 (UTC)
- With respect, you've shown elsewhere that both formats are used, not that both formats are equally acceptable. The Times may use MMM DD, YYYY but that doesn't mean that that format is equally acceptable within the Times' constituency.
- And of course we can't infer from the header the reader prefers - "The user's locale could easily be inferred from the HTTP headers ("Accept-Language" in particular) and a suitable date-format used." We can infer the locale; the date-format then used is based on the usual date-format for the locale - no second-guessing whether the unregistered editor is a Times reader or a regular en-gb DD MMM YYYY person.
- My point was that auto-formatting can - contrary to the suggestions above - be of benefit to unregistered editors: anonymous IPs with en-us in their HTTP header won't get DD MMM YYYY dates, for example, and only Times-readers are disappointed when their en-gb header results in them avoiding MMM DD, YYYY dates.
- Cheers, This flag once was redpropagandadeeds 19:33, 3 April 2009 (UTC)
- But it's not just The Times. In another post here I showed that five of the ten national daily newspapers in the United Kingdom use "April 3, 2009", rather than "3 April 2009". Where is the evidence that one format is preferred over the other in the UK? And the same goes for every other country where English is used to a significant extent except for the United States - I've given examples for India, Canada, and Australia, and would urge you and anyone else to provide such evidence if you are claiming that there is any country that uses English to a significant extent where "April 3, 2009" is not a perfectly acceptable format. Once again, why am I the only one of the hundreds of people commenting here to provide any evidence for their position? Phil Bridger (talk) 21:05, 3 April 2009 (UTC)
- In New Zealand (where I am) there is also no preferred format for dates. Individual publications may have a chosen style, but only someone writing for that publication would be aware of that. Nobody gives a toss what format you write dates in. Indeed it is hard for me to understand that this is even an issue. There is a DD/MM/YYYY convention for date abbreviations which differs from the US one, but non-abbreviated dates can be, and are, written and understood in any format. I don't mind writing dates in the US style. Indeed I do so anyway about half the time. However I would be opposed to wikipedia mandating that dates always be written in the US style. That sounds to me like a charter for date nazis. It is a big world with lots of cultures rubbing up against each other. Wikipedia is an international collaboration where people from all sorts of backgrounds contribute. Tolerance and understanding of differences is needed to make this work. Anyone who is so intolerant that they let the style that other people write their dates annoy them needs to be told to lighten up and get a life. Hawthorn (talk) 08:31, 4 April 2009 (UTC)
- My original point was that "auto-formatting can - contrary to the suggestions above - be of benefit to unregistered editors"; nothing said to date has changed that. It may well be that, as a community, we decide that unregistered readers see date as unformatted - that, then, will be our choice. My point is that autoformatting does give us that choice, despite claims to the contrary. Turning to your evidence, in Scotland one newspaper uses DD MMM YYYY and the other uses MMM DD YYYY (with no comma - first time I've seen that): I don't believe that this - or that the evidence you've presented - says anything other than newspapers use a variety of different styles. Surely the key thing here is what people are taught at school, use in their daily lives, and recognise in correspondence from government? If you wish, I'll try and dig our references supporting my belief that DD MMM YYYY is taught in schools in the UK, New Zealand and Singapore (countries in which I experienced education at some level) in preference over MMM DD, YYYY, that this format is used by the governments of those countries, and that people tend to use that format more as a result - and I dare-say that other editors can provide references for other countries. However you're asking for evidence that MMM DD, YYYY is not a perfectly acceptable format - which no one, so far as I can see, is suggesting. Of course people more familiar with DD MMM YYYY recognise and understand MMM DD, YYYY, and it's ludicrous to suggest otherwise. This is about preference, not about acceptability. My preference is for YYYY-MM-DD, and I'd like a system where that preference can be catered for. My preference in no way affects my ability to accept or understand MMM DD, YYYY or DD MMM YYYY. Your evidence shows that people can accept or understand MMM DD, YYYY simultaneously with DD MMM YYYY - which comes as no surprise to me - however I do not believe it shows anything beyond that. Cheers, This flag once was redpropagandadeeds 11:02, 4 April 2009 (UTC)
(In response to neutral vote # 14) Actually, there lies a significant problem: a not-inconsiderable proportion of dates are wrongly input (Sept 4, for example, or your example), and the proposed system would need to be programmed to fix each individual possibility. Another reason, I believe, that we should not mess with editors' control over simple fixed-text dates. Tony (talk) 06:43, 4 April 2009 (UTC)
- {{#formatdate:Sept 4, 2007|dmy}} produces "Sept 4, 2007". I don't see how that makes anything worse than it would be if an editor simply typed in an incorrectly written date without the autoformatting code. I voted "neutral" because it seems to me that the autoformatting question really contains two hidden subquestions: (1) should we permit use of #formatdate, and (2) should we require (or encourage or recommend, etc.) its use? I would vote "yes" to the first and "no" to the second; therefore neutral on the combined question. I do recognize that some people complain about having to wade through additional markup code when editing, but I personally don't see that as a big problem. And, on the other side, I recognize that some people seem to think it is important that every date in every article be viewable in the same format, but again I don't see that as a huge problem when we allow (and affirmatively endorse) such inconsistencies as American and British spellings and some other usages that vary from one article to the next. --R'n'B (call me Russ) 12:17, 4 April 2009 (UTC)
- R'n'B, I agree about your point with American regarding British spellings. As an American, encountering “I realise that…” interrupts my train of thought whereas encountering either “4 April” or April 4” causes me no interruption in my train of thought—and certainly no confusion.
The trouble is, “autoformatting” is nothing of the sort for 99.9% of our readership. For I.P. users, all it would do in any given article is default to one format or another. And that is no better than just writing out the date in fixed text. The real *benefit* of autormatting (presenting a date in accordance with a preferences setting) works only for registered editors. It’s not worth so much fuss to benefit so few users to address a purely stylistic issue over which no confusion arrises.
For some editors, this is just a big fuss to never see a date style that isn’t their way. Indeed, I just don’t “get it” because I can handle seeing either date format and have no patience for those who insist that everyone else jump through hoops for their viewing pleasure. For the programmers, its about cool code. But they’re trying to sell refrigerators to Eskimos. Greg L (talk) 02:23, 5 April 2009 (UTC)
Frankly, the whole thing seems foolish to me: All autoformatting does is hide inconsistencies, make atrocities like "Sir William Schwenck Gilbert (1836-11-18–1911-05-29)" invisible - for those with it turned on, that read: (1836-11-18–1911-05-29), and mean that, in situations where there's good reason to use a non-standard format, such as "Accessed 2009-05-04" in a footnote, where compactness is advantageous, it'll be replaced with a less appropriate one. What's the point? What next? Will we make it so that readers in Greece have Macedonia direct to the Greek region, but readers in the country of Macedonia have it redirect to the country? It's just not worth it: A solution searching for a problem.
Furthermore, autoformatting dates must require some sort of markup, because of the problem of dates in exact quotes, and other issues. A quote from a ship's log might begin "13-01-1897: Land spotted. Have adjusted course..." If text was automatically changed, and the user didn't know the obscure methods for turning off date formatting, then, for the convenience of a few pedants who insist on consistent formatting when they browse, we've managed to hurt usability of our editing interface in a non-trivial way. So we have to mark up dates. Can dates be marked up autonmatically? No, that hits direct quotes too, in a hard to spot way. So we're now lookign at manual markup of dates on millions of articles. Bugger that! You'll learn to live with occasionally being faced with the dreaded "March 3" or "3 March". Like normal people. Shoemaker's Holiday (talk) 06:31, 5 April 2009 (UTC)
- That about Macedonia would be a great idea, provided that Italian readers have it direct to the dessert. :-) A. di M. (formerly Army1987) — Deeds, not words. 19:03, 5 April 2009 (UTC)
Statement for is biased and fallacious: creative content is not a “user interface”
#Statement for betrays the wrong attitude that readers and editors are “users”, and that my creative contribution, protected by the GPL GFDL, is a “user interface” which should be rewritten by a machine to support “user preferences”. “Personalized date formats in operating systems” don't rewrite the books you are reading or correct the language in music you listen to. This explanation is biased and fallacious, and is misleading editors who read it and vote. —Michael Z. 2009-04-11 16:41 z
- "If you don't want your writing to be edited mercilessly or redistributed for profit by others, do not submit it." No mention of "machines" being excluded from the aforementioned editing is there? --WebHamster 17:11, 11 April 2009 (UTC)
- Also note content here is under the GFDL, not the GPL. Fletcher (talk) 17:14, 11 April 2009 (UTC)
- You're wrong. The GFDL requires anyone who alters my text to do it under the GFDL. There's no GFDL attached to preferences which rewrite page content, and there's no credit for the alteration placed into the article's edit history. The site can change the style using a skin, but if it edits my text then it is infringing on my copyright, according to my GFDL release terms. —Michael Z. 2009-04-11 17:45 z
- See WP:GFDL#4._MODIFICATIONS for the obligations of distributing modified versions of documents. —Michael Z. 2009-04-11 17:54 z
- IANAL, but alteration of mere date strings is likely de minimus. Your phrasing that it "rewrite[s] page content" is vastly weaker when one keeps in mind it is just date strings, which are not copyrightable works by themselves. One should also consider the implementation: if you use a template or other wiki-markup, you have consented to the autoformatting (else you would not have used that syntax); alternatively if you don't use any special markup, it's likely a bot or human will have to go make a change to the article which will be reflected in the edit history. Fletcher (talk) 19:58, 11 April 2009 (UTC)
- Templates may present letter-of-the-law licensing problems, but at least they are placed by humans and altered by humans, whose actions all appear in edit summaries, and thus are governed by the GFDL.
- But allowing a preference for registered or non-registered users to start rewriting the text opens a different can of worms – next it will be blind spell-checking, pirate-talk filters, and who knows what next? Shall we auto-load up machine-translated versions of foreign-language articles, without any local editor in the credits? Do you want your name to be the last one in the credits of an article machine-translated into Chinese? No, thanks.
- De minimus is a legal concept, but we are talking about the substance of people's writing; remember the substance of eats, shoots & leaves is a lowly comma. You can put all of the human editors and proofreaders you want on my writing, but I won't entrust it to a machine which is not being supervised by a responsible editor under GFDL. A spell-checker won't replace an editor, neither to produce a professional-quality document, nor to take legal responsibility for modifying and republishing my work. —Michael Z. 2009-04-11 23:28 z
- I'm concerned that you are claiming WP:Ownership over articles, or parts thereof, to which you contribute on WP. Is it OK if I read them in a different font from you? Or perhaps I prefer to use a speech reader, or read them in French using machine translation. One of us is missing the point; if you think you have ultimate control over the form in which "your" work, creative contribution, text is "republished", don't put it on WP; if not, what's the difference between the examples I gave and changing (manually or automatically) a date format? I can't see any voting stance that implies, as you seem to, that nobody should ever touch "your" text the way you have written it. Have I completely misunderstood? Best wishes SimonTrew (talk) 04:26, 12 April 2009 (UTC)
- I'd say you have. WP:OWN is about editing disputes and has nothing to do with this.
- You are welcome to transform the article any way you like for your private use. But when Wikipedia distributes articles by publicly displaying them on its website, it is bound by the licence. Read WP:COPYRIGHT, which says “the text of the Wikipedia is copyrighted . . . by Wikipedia editors and contributors and is formally licensed to the public under the GNU Free Documentation License”. Under the GFDL, if the article has been modified from the editors' version, then Wikipedia must not display it without taking credit for modifications. Wikipedia:GFDL:
- “§ 0. PREAMBLE: . . . this License preserves for the author and publisher a way to get credit for their work, while not being considered responsible for modifications made by others. . . .
- § “1. APPLICABILITY AND DEFINITIONS: . . . A "Modified Version" of the Document means any work containing the Document or a portion of it, either copied verbatim, or with modifications and/or translated into another language. . . .
- § “2. VERBATIM COPYING: . . . you may publicly display copies. . . .
- § “4. MODIFICATIONS: . . . you must do these things in the Modified Version:
- “A. Use in the Title Page (and on the covers, if any) a title distinct from that of the Document, and from those of previous versions . . .
- “B. List on the Title Page, as authors, one or more persons or entities responsible for authorship of the modifications . . .
- “E. Add an appropriate copyright notice for your modifications adjacent to the other copyright notices. . . .
- “I. Preserve the section Entitled "History", . . . and add to it an item stating at least the title, year, new authors, and publisher of the Modified Version . . .”
- I didn't make this up. You and I are bound by the licence, and so is the Wikimedia Foundation. The licence says that any modifications made must be acknowledged in distributed copies. —Michael Z. 2009-04-12 16:04 z
- You didn't have to quote chapter and verse, I am capable of looking up a reference. Anyway, I really think this is a dead end to this particular discussion. Nothing says the republishing agent has to be a natural person, the Wikipedia page rendering engine could just thwack a copyright notice at the bottom of the page saying "go see the original text"-- in fact, since the rendered version is ipso facto different from the edited text (assuming even the most minimal markup) I could argue it already should-- if it reduces a picture to a thumbnail, for example. These clauses are intended to stop people not crediting Wikipedia and its contributors a whole, not to stop minor changes for rendering purposes. I have started doing some translation and have to credit the original under GFDL, but that doesn't mean I can't change the article, in fact it's encouraged where appropriate The aim of the GFDL is to protect the Commons and Wikipedia etc and to ensure fair use etc. It does not mean, however you would like it to mean, that pages cannot be rendered in a different way by different engines, be they the server or client, or my own blurry eyes when I remove my glasses. I am not going to quote all kinds of references here but the whole Look and Feel argument of the early 80's (Lotus 1-2-3 vs Borlland Quattro) established that, in law in the US, but in practice everywhere. SimonTrew (talk) 22:08, 12 April 2009 (UTC)
- A school of red herrings (my favourite is “the discussion is now over, and here's 200 words explaining why...”).
- Yes, editing is encouraged, but only allowed under the GFDL. I'm not claiming copyright over my “look and feel,” but over the words I've written. What do your blurry eyes have to do with rewriting some of these words without respecting the explicit terms of the licence? Nothing.
- The GFDL does say that whoever the republishing agent is, individual or corporation, they have to take credit for their modification. And you're quite right, thwacking a copyright notice, along with the required change of title and the acceptance of credit could well be enough – but the foundation doesn't. The copyright notice linked from every article says explicitly that copyright is held by the editors, and not by the foundation. —Michael Z. 2009-04-13 20:25 z
- So argue for them to put a copyright notice on. That's irrelevant to the discussion. It's also slightly ingenuous of you to start with that quote when I never said anything of the sort; I know you will say neither did you, but to head the reply that way implies that. I know... let's suggest that every phrase, word or quote used in wikipedia should have markup and copyright references. So if I quote Churchill I better make sure I have a good source (History of the Engish-Speaking Peoples, perhaps) and make sure that I check whether it is in or out of copyright. What nonsense. The copyright issue is completely irrelevant, and whatever red herrings there were, you didn't point them out. Discussion is over, in fewer than 200 words. SimonTrew (talk) 23:14, 13 April 2009 (UTC)
- To answer Martindelaware's question (support 189)
No date autoformatting (DA) system will cope with reformatting these fragments (from Charles Darwin) "On 28 September 1838 he noted this insight." to "On September 28, 1838, he noted this insight." as well as reformatting "He died on 19 April 1882." to "He died on April 19, 1882." without complicating the syntax of the markup or making mistakes when the year is followed by a period (or other punctuation characters like parentheses or ndashes). The present, deprecated, system of DA fails to supply the required second comma to produce apposition. I have seen no suggestion that any "Son of DA" will fare any better. In fact, any proposed DA will increase errors by hiding the missing second comma from many editors who do not have that preference set. --RexxS (talk) 00:25, 12 April 2009 (UTC)