Wikipedia talk:Manual of Style/Linking

Questions re delinking

Further to the most recent blow up of this on the Queen (band) page, I thought I’d spend a bit of time putting together some questions and actually get some answers. The general ones are mostly - with some additions - on points that I and others have raised on many occasions, but that I have never seen properly addressed. The more specific ones are just a couple of examples I’ve dug out from a couple of scans of recent changes, which either highlight some of general issues, or which raise their own, more specific questions. Without wishing to dictate how others format their responses on a talk page, perhaps it would be easiest for people to add any comments under each individual question/point -

General points

• wp:link explicitly says: “Provide links that aid navigation and understanding”. That is, aiding navigation is a defined aim of linking, as well as it having some kind of pedagogic function.

• It also explicitly says: “Think carefully before you remove a link altogether—what may seem like an irrelevant link to you may be useful to other readers”. That is, a simple assertion by a couple of editors that a link is “not useful”, whether generally, on any page, or in a specific context, should not by itself suffice to sanction delinking the term.

• It also qualifies the suggestion to “avoid linking terms whose meaning can be understood by most readers of the English Wikipedia, including plain English words, the names of major geographic features and locations, religions, languages, common professions”, with the qualification/exception: “Unless they are particularly relevant to the topic of the article”. Yes, linking “France” every time it appears in every article would lead to thousands of somewhat pointless links. But does that mean it should never be linked, even in articles about things from France?

• Given the above, where is the consensus that terms such as “France”, “sitcom” etc need to be stripped from every article, even from infoboxes, whether by script or other automated process, or by manual intervention? Where, furthermore, is the consensus as to which of these terms are to be stripped? It seems to me that people running these removals are applying different criteria. Hence the process is rigid, arbitrary and inconsistent all at the same time.

• Why does the presence of a link to say “London” then require the removal of the separate link to “England”, where the text includes both terms, as in “London, England”. They direct to different pages. Yes, London is in England, and people could go to the London page, and then on to England, but why reduce navigability and push people through this convoluted loop when there’s no need to?

• Why remove links on the basis that “people know what/where” something is? First, at what point and on what basis do you decide when something is well-known enough – Sydney, but not Brisbane? Secondly, even if there could be some form of objective standard on this, so what anyway – I know what/where Darwin is, but that doesn’t mean I don’t want to read the WP page on it, getting there from a related page.

• What exactly is the “dilution” or “distraction” point? People can see the links, and decide whether they wish to go from the WP page covering the current topic to the WP page covering the related topic or not. How are they being distracted from looking at “better” links? And on what basis are we asserting that some links are uniformly – in all circumstances and for all readers – “better” anyway?

• How does the “readability” point – such as it applies at all – apply to infoboxes and lists? They are not prose.

Specific recent examples

• This edit to a page about a (rather tired) BBC sitcom removed, for example, the links to “sitcom” and “BBC One” - but not for example the links to pages about the other channels that the programme has been repeated on. These links are surely relevant, and in the case of the link to the page on the original broadcaster, BBC One, surely more relevant than the links to other channels. Isn't their removal somewhat arbitrary?

• This edit, on a “List” page, removed the links to some cities – eg Paris, Moscow – but not others, such as Brussels. What benefit is there here? Again, on what basis is this arbitrary distinction being drawn?

• Here, for example, the links to “pop music” have been removed, but the links to “country music” retained. She is best known as a singer – but we’ve also lost the link to that generic page, while keeping the link to the “political activist” page, something that is of secondary relevance. This makes no sense, surely, as well as again being totally arbitrary?

• Here, where is the rule that says we should not link nationality, as asserted in the edit summary? Personally, I’m unsure whether such links are needed, but there is always arguably the option of linking to the relevant “people” article, eg "British People" rather than simply UK, or even to a more focused link on say, in this case, “Politics of the United Kingdom”.

OK, essay/questionnaire done. N-HH talk/edits 11:01, 13 May 2010 (UTC)

Follow-up: more recent examples of poor delinking

• This manual edit on the Rugby football page removed links to the page that covers Football (in the widest sense, not "soccer") and to the United Kingdom, the country where the game was invented. When discussing where the game is played, it also removed the link to Portugal, while keeping the adjacent link to Romania (I later replaced all these and other links with links to the pages on rugby in each of these places); the same for New South Wales and Queensland - all completely arbitrary and inconsistent and not seemingly based on any fair reading of any style guidelines. Equally it left in place multiple repeat links to Rugby school and Rugby Sevens.
  • I can live with most of the relinkings, except for United Kingdom (what in that article aids readers' understanding of rugby? All they need to know is that the sport originated there). I also removed links to basic grade-school level terms such as "alcoholic beverage" (deceptively piped to "drinking") and "social class" (pointless when you have more specific terms, "upper class" and "middle class" linked in the same section). Dabomb87 (talk) 19:42, 21 June 2010 (UTC)

Actually, the country-name links have been vastly improved by N-HH in his piping them to "Rugby union in ..." pages (although will the reader know this? I'm trying to think of a way to make it more obvious). I have no objection to links to relevant things in a country. The link to the UK is an unfortunate example of the old way of doing it, though. Why not the same piping idea there? Tony (talk) 09:02, 23 June 2010 (UTC)

• This script-based edit to the Dubai page removes links in an infobox to India, the country from where a vast number of expat labourers come from.
  • It is true that many expat Indians go to Dubai (I have several such friends and relatives). However, a reader directed to the article Non-resident Indian and Person of Indian Origin will find on a quick skim of the lead the basic definition of an NRI, which is probably not what they were expecting. I fixed the piped link to lead to the "Middle East" section of the article, which actually mentions NRIs in UAE and may actually be useful to the reader. Lesson: sometimes, improving link specificity can be a viable alternative to delinking. Dabomb87 (talk) 19:52, 21 June 2010 (UTC)

• This version of a recent main article, Sydney Newman, had been left with no link in the lead to the BBC, but did have links to other broadcasters where he had worked. Again, arbitrary and inconsistent. There were also no links in the infobox for his profession, even though "TV producer" can hardly be said to be a common occupation.
  • This one I admit is a toss-up; if I had happened upon the article I probably would not have changed a thing. On one hand, the BBC is one of the most well known broadcasters in the world, and "film producer" is a pretty well-known and -understood occupation, so I see why Tony delinked those terms. On the other hand, Newman spent a large chunk of his career at the BBC, and while TV producer is pretty common, it is not as self-explanatory or universally known as author or athlete. While we're here though, I don't think that occupations are automatically qualified for links. As always, whether or not a career should be linked depends on whether readers would actually need to click on them; common terms such as writer or dancer rarely require links (unless you link to more specific career links, such as "romance novelist" or "salsa dancer"), but I would fully expect to see gastroenterologist linked. Dabomb87 (talk) 20:01, 21 June 2010 (UTC)

I don't think this is a toss-up. The BBC is a household name all over the world, not just in the English-speaking world. There's even "BBC America". "TV producer" is too much like a dictionary word: what English-speaker doesn't know its meaning? I agree with Dabomb about "gastroenterologist". For those who are interested, I do routinely add links to Signpost articles, and even FA nominations, when I copy-edit them. It goes both ways. Tony (talk) 09:06, 23 June 2010 (UTC)

Yeah, but again I don't see what that has to do with anything, whether common sense or wp:link. Yes most people probably know what the BBC is, but the BBC page doesn't simply say "national and international UK public broadcaster". It's a long page full of detailed information about the organisation, including, presumably, some detail about the place when Newman was working there. I'm British and know quite a lot about the media, yet if I were to sit down and read it, 90% of it would probably be news to me, and not just because it might be nonsense. Page stats suggest it's been viewed 130,000 times last month. Are these people all stupid? And why would we want to make it more difficult for those people and others like them to reach the page from ones that are clearly relevant/related? Similar points apply - in principle - to TV producer. No WP is not a dictionary - that's precisely the point. And, again, how would we draw the line as to what is a "household name" anyway for the millions of people who come here, from millions of different places and cultures, for millions of different reasons? N-HH talk/edits 12:14, 12 July 2010 (UTC)

I'll preface my comments by saying I was extremely annoyed by the patronising tone of the post above. I would merely point out that trotting out page stats like 'the BBC article has been viewed 130,000 times last month' is all good and true, but is pretty meaningless without a detailed analysis of the origin of the hits in an attempt to understand why people are landing there. It is undoubtedly a popular topic in its own right, but it is also widely linked to. Any typical reference section will generally contain at least one link to the BBC. So no, people who click on it are not necessarily stoopid ;-) The plain simple fact is we don't know why; we also don't know how long they stayed, or if they found what they were looking for. Ohconfucius ^¡digame! 01:45, 13 July 2010 (UTC)

• This edit to the Indigenous Australians page removed links to two out of the six Australian states (plus the ACT) in an infobox, as well as taking out a link to Australia in the first sentence. Why remove Queensland and Western Australia, but keep New South Wales and Victoria? Why take out a link to Australia on a page that even has the word in the title (it has since been replaced, with a slightly better link). That's possibly the most random one I've seen yet. Is anyone going to stand up for this one? N-HH talk/edits 12:21, 12 July 2010 (UTC)

Again, only a sample. N-HH talk/edits 19:10, 21 June 2010 (UTC)

This is a rare instance where one link to the whole "Australia" article might be reasonable; however, the part you point to has been relinked as "Australian continent", which is probably less offensive to Indigenous people, who might be forgiven for seeing the Australian nation as a cruel, oppressive, destructive phenomenon. The cultural sensitivities are a minefield. As for linking the names of the colonial states, they created borders in the continent that are at loggerheads with traditional Indigenous relationships with the landscape, and perpetrated and supported such evils as the murder, displacement, and cultural destruction of the original owners, starting in the 19th century when there was no Australian nation. The articles on the states (and territories)—which are increasingly regarded as silly and expensive relics of a poorly conceived federation—are written more from the invaders' point of view than is the Australia article. If readers want to see the boundaries to see where these states and territories lie, why not one section-link to the good map of this, in the "States and territories" section of Australia? Tony (talk) 12:55, 12 July 2010 (UTC)

I acknowledged the continent link was better (and also am aware of the broad political and historical issues you have highlighted). And as for the individual states, for better or worse, they are the current administrative/political divisions. Nor am I sure why you assume that readers only might want to see their boundaries, or why a link to a map would be more helpful. And anyway my point was, why unlink only two? You haven't addressed that, and it is one of the key points in the whole debate, of which this simply serves as but one example - the random consequences, especially in lists/infoboxes, of rigidly removing certain terms without looking at the individual context of each article. N-HH talk/edits 13:04, 12 July 2010 (UTC)

Thanks for responding btw. I think the general points that come up, from where I'm sitting, are that relevance and navigability are key issues as well, as the guideline states, and can justify linking, even to things that are well known or to pages that might not necessarily directly assist understanding as such; plus that it's difficult to draw lines according to the latter two criteria, given the variety of people who are likely to look at pages here, and that trying to edit links according to them often leads to fairly arbitrary inconsistencies. My preference is to err on the side of keeping a link when there's a close call, unless it's manifestly redundant or trivial. Or, as you suggest, find a better and more direct link. I think it's too easy, as people who might edit here to a lesser or greater extent, to forget that a lot of people who use the site probably come here very infrequently and that navigability does matter. We might use links less, and hence think them less useful, because we've already seen the more obvious pages. Not everyone has. N-HH talk/edits 08:31, 23 June 2010 (UTC)

Discussion

• The more professional approach being adopted is to provide the minimum number of links that give the best experience for the average reader of an article. Overlinking is rife at WP, and the great work being undertaken now by many editors is to reduce that overlinking. It is a skill to bring out the high-value links that add value to an article, and linking dates and other low value targets diminish the value of links to the reader. If you want to see where we don't want to go, have a look at a few pages on the French Wikipedia.  HWV258.  11:53, 13 May 2010 (UTC)

Par exemple? - Pointillist (talk) 11:57, 13 May 2010 (UTC)

And how exactly do you propose objectively defining what would constitute the "professional approach", or the "best experience" for the "average reader", and what "add[s] value"? And where did I talk about linking dates anywhere above? I note as well that you have not addressed in any detail any of the substantive points I spent probably far too long posting about above. Making vague noises about "overlinking" - which undoubtedly exists, and should be dealt with - does not give a small self-appointed group the right to impose an incredibly rigid stripping of specific links from hundreds of articles, where those links are often arguably very relevant to the articles in question. N-HH talk/edits 12:14, 13 May 2010 (UTC)

Your pejorative language ("impose", "self-appointed", "incredibly rigid", "stripping") is unnecessarily combative. I have spent a lot of time conversing with you, so I thought it prudent to give others a chance. I'll leave you with one thought: the foundations laid by the current work will benefit the long-termed perception and value of WP—and that is the only outcome desired by the editors involved.  HWV258.  22:20, 13 May 2010 (UTC)

It is simply descriptive, and that is what it is intended to be. The only word that could possibly be seen, in addition, as mildly pejorative is the term "self-appointed". Someone who writes in these combative terms about vanquishing their rivals is probably in more trouble when it comes to their language. Anyway, I would rather others commented as well, but I also want answers to simple questions from those who keep avoiding answering them in any detailed way. We know we disagree - you think these changes are of benefit, I and others query whether that is always the case. With the examples I cited above, those questions seem very reasonable. N-HH talk/edits 22:42, 13 May 2010 (UTC)

• Nick HH, you say "Why does the presence of a link to say “London” then require the removal of the separate link to “England”". Try Carnaby Street, Soho, London, United Kingdom, Western Europe, Europe, Northern Hemisphere, Earth, Solar system, Milky Way, Universe. I think the first one as a link is quite enough, don't you? That target includes ample geographical links itself ("chain" links).
• Yes, they do "aid navigation", but they come at a cost that needs to be balanced. Just as editors are charged with making calls on matters of prose and style (within the style-guide frameworks), so they are expected to think about the readership when they link. Prose is rarely perfect on WP, and sometimes the boundary between linked and not-linked cities isn't either. You could make that issue go away by linking no cities at all, or every city. The trend is definitely away from carpet linking, because people realise wikilinking needs to be rationed intelligently to be optimised.
• "I know what/where Darwin is, but that doesn’t mean I don’t want to read the WP page on it, getting there from a related page." Type it into the search box, then. Do you really think our readers click on many links at all? Naaaaah. Let's get real about it: webmasters and psychologists tell us that surprisingly little link-clicking goes on in contexts such as WP articles. Ask User:Piano non troppo, an ex-professional webmaster himself, who dealt with such issues in the corporate sector. It was his job to know. Ask a fundamental perceptual psychologist such as User:Holcombea, who knows about signal-to-noise ratio. Look at the robust research findings by marketing academics over the past six to eight years that indicate the reduction in functional behaviour by consumers when they are presented with more than a certain amount of choice (surprisingly little, actually). And then ask a reading psychologist like me whether the linking density you encounter in many articles on en.WP—and in most of the other WPs—is helping or hindering our readers.
• If you want everything linked, go lobby developers at WikiMedia to introduce total linking of every item, highlighted only when the cursor hovers over it. This is not a bad idea: the latest version of the Encarta desktop dictionary—the one that comes on every Mac—does this. But the big loss would be our ability to use our knowledge and skills to show readers which items are important, relevant, and focused sufficiently for them to bother even contemplating a click. It's all about rationing—logically, intelligently—just as we ration words when we eliminate redundancy in prose. Tony (talk) 13:12, 13 May 2010 (UTC)

Come on, point one is a little silly, when is something like that ever going to come up? We're talking about a chain of two or three words maximum, often in an infobox anyway, not in prose text. On point two, I agree - editors need to make calls. To me, that means not necessarily stripping all links to Germany from every single article, but deciding when and where such a link is relevant, and should, per wp:overlink, be included. As for carpet-linking, I think in a list it's less of an issue. The arbitrariness of removing some cities while keeping others, however, as noted in my example, is a problem. You seem to accept at least that such a middle way is not the best option, even if you'd presumably prefer outright removal of all of the links. As for point three, I'm not sure of the direct relevance of such research to what is actually quite a simple and trivial issue (and also, with no disrespect, I do not take as read the claimed expertise of anonymous wikipedia editors). It seems to be a fairly simple and uncontroversial point to argue that a link to a relevant/related topic is at least likely to be helpful at some point. It's less clear how it hinders people in any way - people have the choice to click on it or not, instead of or as well as other links. Why make them go through the search function, when there's a much easier way of connecting to directly relevant pages? And I have no stats to hand, but the idea that people rarely use wikilinks seems unlikely. I know I do. On point four, I really would be grateful if you and others would stop this "if you want everything linked ..." nonsense. I and others who have queried aspects of this campaign have been very explicit about supporting the removal of repetitive and trivial links, and of those whose relevance or significance is limited given the context. What we are asking for is a little more discretion over the removal of links to common terms, such as professions and countries/cities, where those terms are relevant to the topic at hand.

Amid all my own verbiage above, and all the theoretical debate about the fundamental purpose of wikilinks, that last comment reflects the basic point from where I'm sitting. I still fail to see what advantage is to be had, for example by removing the links to pages on Germany and World War Two from the page about a German General who commanded armies in World War Two; or by removing links to London and England/English people in the infobox for an English person who was born and lived in London. We are only talking about a couple of links on each page, and wp:overlink pretty clearly supports the view that those terms are fine to link on such pages. If you disagree with that, explain how your view fits with the wording of overlink. If you say that there is a separate, agreed consensus to override wp:overlink, point us to the discussion that came to that conclusion. These points have been evaded for long enough now, and batted off with vague assertions about "dilution", language that does not appear in wp:link. On the more theoretical points, as I've noted before, I also find the suggestion that some of us know better than others what pages those others should be looking at a little bizarre and elitist. You make broad assertions such as that wikilinks "need to be rationed", and suggest that people should turn to you if they wish to find out whether excessive links help or hinder. Your opinions are as valid as other people's of course, and you may even be right in some respects. But they remain your opinions. And, even if correct, they do not justify the removal of every single link to specific terms in every article. N-HH talk/edits 14:18, 13 May 2010 (UTC)

There's no advantage whatsoever, other than to meet a specific personal preference. For all the talk of a supposed "sea of blue", there has been nothing said that justifies the mass deletion of thousands of useful links, the "private" declaration of what is and what is not "common knowledge", and the condescending attitude that they know best what readers do and do not wish to read. It is very easy, of course, to then label anyone who even thinks of questioning this campaign as wanting to "link everything", and thus avoid addressing the actual concerns raised. For example, where is the research regarding link density in an encyclopedic context? Why are we being parroted data about consumers when that concept has little or no connection to a non-commercial project like Wikipedia? Why should we have to tolerate this secretive, we-decide-what-should-be-delinked script-based stripping of links when the list of so-called "common terms" is not easily accessible to Wikipedia's editors and has never been discussed? The scriptwork is even presented complete with edit summaries that lead the average user to assume it must be "official", consensus-driven cleanup when it is anything but. This is all based on a narrow, overly rigid interpretation of a guideline that was also purposefully rewritten to achieve this same goal. It would be nice if we could get some honest answers to these questions, rather than the avoidance that has occurred to date. --Ckatz^chat_spy 16:25, 13 May 2010 (UTC)

Ckatz -- I just tried to make a series of edits. Half an hour of work. And the new wikipedia system destroyed it. Poof. I'm thinking, it didn't like the overlinking either. Seriously, though, at the margins there will always be some who disagree. I would hope that editors can cut out the wasted time of fighting, and figure out how best to address overlinking. Clearly, we could link every word. Probably could have a bot do it. Would save enormous man hours, since you would no long have to hit that square bracket key. Would you be in favor of that. Possibly. I don't know. But I'm guessing most people would not like an all-blue page. Why? For the very reasons I and some others have mentioned. If most of you are with me till this point, we then come to the question of where to draw that line, for the benefits that we all agree are engendered by not having an entirely blue page. That's all this is about, at its core. Let's start pulling in the same direction. Just my two cents. Epeefleche (my sign button is now not working either).

Eppefleche's repsonse was to my "Comment" post below; I had originally posted it directly following the above but have now moved it below to avoid it being interpreted strictly as a reply. --Ckatz^chat_spy 22:39, 13 May 2010 (UTC)

Ummmm ... either I wrote my post even more poorly than usual, or you misread it, or you've just spilled coffee on your new slacks.--Epeefleche (talk) 19:05, 13 May 2010 (UTC)

I think you've misinterpreted my comment as being directed solely at you; it is not. (I'll clarify this on your talk page so as to keep this discussion focussed on linking.) --Ckatz^chat_spy 22:17, 13 May 2010 (UTC)

Comment I find it very frustrating to repeatedly see claims tossed about that myself, N-HH and others want to "link everything" (or words to that effect). The notion that I, or N-HH, or probably anyone else who's objected to the hard-line delinking is in favour of "linking everything" is nothing more than fiction, pure and simple. I've made that clear, and so has N-HH; it is simply a diversionary tactic from a small group who are pushing the delinking effort far beyond what many average users would consider a reasonable point. That group has drawn the line for what should and should not be linked based on their personal preferences, and it has begun to detract from the core functionality of the site. If you review the discussion to date, it becomes clear that we are not disagreeing about basic concepts (such as not linking simple words), nor the idea that we don't need to link every single time the word "Canada" appears in an article. The critical difference lies in the attitude that we should almost never link terms such as "Canada" or "New York", even in articles directly related to those terms. I've seen the delinking script used to strip away all links to the US in an article about its closest neighbour and largest trading partner, links to WWII in articles about battles in that same war, and so on. That is unreasonable, and - more to the point - there has never been a consensus to do such work. Why won't Tony and the others address repeated questions regarding why the list of "common terms" is hidden away in the depths of a script, rather than out in the open for debate and change by all Wikipedians? Why are personal opinions regarding the process - "sea of blue" and "smart linking" to name a few - being presented as if they were policy in explanations to editors who are unfamiliar with our guidleines? For that matter, why do we abide the continuous use of a loaded term as "sea of blue" when (often as not) we are really disagreeing over a handful of links amidst hundreds or thousands of words? --Ckatz^chat_spy 18:57, 13 May 2010 (UTC)

comment. I checked random edits by Tony1 and cannot justify most of them. What's curious, he arbitrary delinks some "common" names and leaves others, no less "common", linked which results in a particularly sloppy look [1]. Is there any reason to treat Hong Kong (delinked) and Thailand (left linked) differently? Whether it was a random slip, or some private judgement over who are "common" and who are not, is irrelevant; this arbitrary mosaic delinking must stop. East of Borschov (talk) 05:12, 16 June 2010 (UTC)

First, N-HH, could you please write less: people will be more likely to read it. I have a sore finger scrolling down and I find myself just skimming your text it is so long. Seriously, I'm doing you a favour in providing this advice. State the punch-lines at the start.

Second, this is just a re-run of the same old issues that have been aired here many times by the same people. Do I have to recycle the rejoinders? WP:LINK's guidelines on minimising links to common geographical terms (not to mention bunching them together) are well-established and well-supported by the community. The boundary between whether Thailand or Hong Kong should be linked is more up to individual editors, and if they want to go in and unlink "Thailand", sure, I have no objection—it has little value as a wikilink, and dilutes the important links surrounding it. I would have problems in re-linking "Hong Kong", just as I have problems linking "France" at the top of the article on "Champagne"; and "New York City". A boundary will always need to be applied to these items, just as we do WRT the linking of non-geographical items. This is nothing new: just a re-rurn of the same old dialogue. Tony (talk) 08:50, 16 June 2010 (UTC)

Damned if we do, damned if we don't, eh, Tony? You give those who disagree with you grief if they don't explain their position, then you chastize them when they do. Frankly, it would be a lot better for everyone involved in this sorry mess if you would just please address the points raised above. --Ckatz^chat_spy 09:03, 16 June 2010 (UTC)

I don't comprehend your first point here. There is a balance that must be struck in what is worth wikilinking and what is not. We long ago dispensed with the initial "link whatever you like" practice. Just as in prose style, editors have their own ideas: the larger picture (and some details) are set out in the style guides. You have breached this style guide on purpose in the article Squamish. I assume good faith that it wasn't to bait other editors. I ask you to assume good faith on my part. And I ask you to discuss substantive issues rather than personal ones, which has been almost the entire thrust of your posts on the matter of linking to date. Your next post will also avoid the substantive issue. Tony (talk) 09:10, 16 June 2010 (UTC)

Umm... last I checked, I wasn't the one accusing you of being a malcontent, or of trying to link everything, nor am I misrepresenting your statements, etc, etc. If you really want to keep this on a professional rather than a personal level, then you will have to make an honest effort to behave professionally. Look, Tony, this all began when you rewrote the guidelines to match your personal preferences. Frankly, I think you took unfair advantage of your reputation on Wikipedia to make changes without proper discussion. However, now we're two years down the road and we have to work with what we have. At this point in time, it would really help if - instead of relentlessly trying to push on with your vision of delinking - you would simply consider compromising in order to accomodate others who do not share the same dream. --Ckatz^chat_spy 09:21, 16 June 2010 (UTC)

"Your next post will also avoid the substantive issue." Whadd'I tell you? Tony (talk) 09:27, 16 June 2010 (UTC)

And once again, you're glossing over any and all concerns raised in favour of more cheap shots. Seems we're in a bit of a rut, doesn't it? --Ckatz^chat_spy 09:33, 16 June 2010 (UTC)

I am aware I can be a little verbose, but, you know, reading something that isn't total drivel isn't really that hard, and sometimes there are quite a few points to be made. Especially when most of them are never addressed in any detail, or are even deleted without your reading them - which is why we ended up at ANI at all - and all that we get in response are accusations of being "malcontents" et al or of "wanting to link everything". To be quite frank Tony, most of your comments at ANI and the raids you have launched on to people's talk pages who have dared to query your behaviour, or make simple factual points about the date case, reflects very badly on you. Seriously, I'm doing you a favour with that advice. When you find that most people who comment on a topic have issues with your actions and arguments, perhaps you should stop railing against them, and maybe just ask for a second whether they might have some valid points and concerns. In a collaborative environment, whether you like it or not, if you can't persaude those people that you are "right", especially about fairly minor style points, then by definition you are not right. N-HH talk/edits 14:35, 16 June 2010 (UTC)

Substantive point - Tony, you have at least above been a bit more specific in responding to specific issues, eg by talking about common terms such as Hong Kong, France etc above. However, I continually point out to you that there is an exception even in wp:link as written, which suggests linking them when they are relevant to the main topic. France in Champagne seems relevant to me. Canada in an article about an area in Canada seems relevant to me. On what basis do you disagree? Would you keep reverting even where consensus is against you on such points in individual cases? N-HH talk/edits 14:43, 16 June 2010 (UTC)

←"The" is relevant to the article on Champagne, too. Why not link it? "France" is sooooo general, and sooooo well-known, why don't you find a section link or a more focused daughter article ("Agriculture in France", if it exists?)—then your only problem would be that "France" as a pipe would be deceptive, and still no one would bother clicking it. "France" is adjacent to the more specific location, which links to France itself, if anyone would need to know even from that article. You are caught up in this concept of linking as auto-browsing: a magic blue carpet to anywhere vaguely relevant to the topic, just in case someone wanted to click it. They almost never do, I'm afraid, and your constant pushing for the linking of common geographical terms is further diluting the likelihood that readers will use the system. You think you're improving it, but you're degrading it. Good faith, but faulty reasoning, IMO. But taking this to ANI was in extremely bad faith—a political stunt to smear me. It is disreputable. Similarly, your use of political language is transparent—now my posts on users' talk pages are "raids"—oh give me a break. And as someone else pointed out above, the use of emotive words such as "stripping" rather than "unlinking" does your case no good at all. People see through the spin. You are not speech-writing for a politician or inventing language for TV ads.

And no, you don't write "total drivel"—you write well (just too much). This is why it's such a pity you've chosen to fight tooth-and-nail efforts to improve the readability and appearance of our text, not to mention the dilution of high-value wikilinks, and thus the utility of the whole wikilinking system. That is what we are trying to protect. Linking needs to be selective to be of real assistance to the readers. Linking "France" is not only useless: it insults the intelligence of the readers, even the eight-year-old. Tony (talk) 04:12, 17 June 2010 (UTC)

Tony, as I pointed out from the outset, I went to ANI because you ignored and deleted a clear note from me alerting you to the errors. And my use of terms such as "stripping" and "raids" is fair enough when referring to scripts being used on multiple articles, and your posting of notes on several editors' talk pages making wild accusations against them (eg telling a non-admin that they're not fit to be an admin, and leaving shouty messages in block capitals). On linking, as ever, much of your argument is about assertion and guesswork, about what every other user might be doing, or ought to be doing when reading articles. Editing involves judgment in the light of context, and linking is about navigability and options. And, for exmaple, in what way exactly is it "insulting the intelligence" of readers to offer even one link to France in an article about a French thing? It's not of course suggesting they don't know what/where France is. It's just saying - here's the option of going to that related/relevant page and reading it, if you wish to. It's far more insulting to people's capabilities to suggest that if we don't take out the link, they might be misled somehow into clicking it when they really didn't want to. I mean, that's just bizarre, as I've pointed out before. The dilution/distraction argument applies when a list of random ELs are dumped at the bottom of a page. There, filtering can work and help the reader. Wikilinks are much more transparent. N-HH talk/edits 14:56, 17 June 2010 (UTC)

What? Tony (talk) 15:22, 17 June 2010 (UTC)

What part of It's far more insulting to people's capabilities to suggest that if we don't take out the link, they might be misled somehow into clicking it when they really didn't want to. don't you understand? --Michael C. Price ^talk 07:51, 26 September 2010 (UTC)

Current consensus

OK, since we don't get answers above, and we don't get answers at the (second) ANI thread, let's start again, and keep it simple - 1) where is the consensus to delink common terms/countries in every instance, regardless of context; and 2) where is the consensus as to which terms/countries fall within the definition of "common"? Links to those discussions please. Then we can perhaps look at how to maybe move forward with an RfC. N-HH talk/edits 19:23, 21 June 2010 (UTC)

Nearly a month later, and nearly three months after I first asked them, and still no answer? N-HH talk/edits 12:15, 12 July 2010 (UTC)

Yes, quite telling, isn't it? --Michael C. Price ^talk 07:49, 26 September 2010 (UTC)

RfC?

I think it'd be helpful if we had an RfC on delinking, so we can trash out an actual consensus on what the principles are, and if/how any auto/semi-auto delinking should proceed. One thing that's needed is for those who are delinking to present what exact scripts and rules they are using in deciding what links to remove, and on what basis that decision was made. I think we all agree on the general idea that there is overlinking, and nobody wants to "link everything", but there is a grey area involved certain "common terms". A clear example is Champagne (wine): Tony wants to remove France from the lead arguing that it adds no value; I suspect many editors would think a link to France in the lead of a drink that is so strongly associated with France would be a good idea. The same edit did remove extraneous links like Prime Minister next to Tony Blair, though I'd argue that Co-operative is handy when mentioning wine co-ops. So the delinking here is open to debate. I think the RfC needs to decide 1. What is actually is considered to be overlinking. 2. Whether automated processes are the right way to approach delinking. Can we write an RfC wording that says this neutrally and succinctly? Fences&Windows 21:25, 17 June 2010 (UTC)

It is very hard to seek opinions simultaneously on (1) the boundaries between common items that should be, can be, and should not generally be linked, and (2) automation to unlink such items. For example, the result of (1) might well render a positive result on (2) unachievable technically. Automation requires its own set of proposals based on a result in (1). In any case, automation can never play a major role in reducing overlinking, since human oversight is required constantly, and the scope of items included in automation has to be only a small proportion of those that, in their context, need to be unlinked. It is essentially a manual task. (1) would need to be dealt with first.

The whole article on "France" is of little use to anyone but a reader who wants to wander through the most general links; "France" is now relinked right after "the Champagne region". The guideline says, "Always link to the article on the most specific topic appropriate to the context from which you link: it will generally contain more focused information, as well as links to more general topics." It also says, "Provide links that aid navigation and understanding, but avoid cluttering the page with obvious, redundant and useless links." and "can make it harder for the reader to identify and follow those links which are likely to be of value". and "Ask yourself, "How likely is it that the reader will also want to read that other article?" There are more than 100 links in the article text, and nine in the vicinity of the reinserted link to "France". So the guideline is telling us to be selective, in my reading of it. What is the value of the whole article on France at that particular point? Who doesn't know what France is? Nothing is stopping someone going to the article; but it seems just to dilute as an additional link at the top. Wouldn't a more specific link than the whole of France be more useful to the reader? ("Agriculture in France"?), wound into the article smoothly?

The same applies to "Prime Minister", which is already prominently linked from the top of the far more specific "Tony Blair" article: here, it bunches together with "Prime Minister", against this guideline: "When possible, avoid placing links next to each other so that they look like a single link".

These guidelines have evolved over years, and everyone was here when User:Kotniski led a conflation of two other pages into this one. Which parts of the guideline are at issue? Tony (talk) 03:03, 18 June 2010 (UTC)

Agreed, there are two threads to the issue - 1) the delinking of common terms in principle, and 2) the use of scripts or other systems to bring that about. Not all the unlinking is done by scripts. In principle I don't object to the use of scripts, and think they probably do a good job in terms of removing redundant, repetitive and irrelevant links to common terms that are linked too often, usually when the terms crop up in passing in unrelated articles, as they often will. However, I do think that a) the terms they are stripping need to be clear and open and agreed, and b) there needs to be manual review of their effects in each instance, not only to avoid obvious mistakes per the ANI thread, but maybe sometimes to restore one or two links that are going to be relevant in the context of that individual page, or that restore consistency to lists or infoboxes. My personal preference - and there is nothing in wp:overlink to say this is a "wrong" view - is to err on the side of inclusion in those cases. For those who edit here, it's too easy to get sucked into a narrow vision and also to overanalyse everything. I'm for example never going to move through to the France page from the French wine page in the future, and we can all theorise about what might subjectively be "better" or "high impact" links, but - forgive me for the cliche and the hypothetical - who's to say that a 15 year old from Hawaii who's looking at the latter page for the first time, to research a school project, wouldn't find it useful? Even if they broadly know what/where France is already.

I don't think it's impossible to draft an RfC that would at least help establish a broader consensus one way or the other on these issues, and that would agree how scripts should be compiled and used and their effects monitored/reviewed, and yes, sometimes mitigated. Also I do think that even without an RfC, those running scripts should be checking/proofing their effects in each case anyway, and also should learn to shrug their shoulders a little if another editor comes by subsequently to reinsert maybe a couple of links, if there's plausible justification for that. N-HH talk/edits 16:51, 18 June 2010 (UTC)

I do believe discussion on this page should be conducted without mentioning other users, unless in positive terms. I think the use of items such as "stripping", "campaign", "point-scoring" and "raids" should be avoided by all. I think other people would readily agree to this. Is that the case? It is most important to be collaborative. Tony (talk) 04:36, 19 June 2010 (UTC)

Let's add "sea of blue" to the list, OK? That would be of great benefit; what would also help would be ensuring that we all take great pains to represent each other's position accurately and with a complete absence of any hyperbole. Is that also acceptable? --Ckatz^chat_spy 05:03, 19 June 2010 (UTC)

And, if we could, accusations of "political stunts", that others have been "venting" or have "extreme views" or "do not accept the rules" or "complain loudly" etc etc, as well an end to answering others' posts with the response "What?", or simply not reading them altogether. Also - from others - accusations of "forum shopping" of "haranguing" etc etc. As my post above shows, I am more than capable of explaining the queries and issues calmly and reasonably. Reciprocation would indeed be welcome. N-HH talk/edits 05:11, 19 June 2010 (UTC)

ps: note the word "raids" was a one-off, used to refer to a very specific set of comments made on other's talk pages. "Stripping" I use regularly, and I use it with a purpose, since it accurately describes the act of rigidly removing all links, via a script. That's at the core of the issue, and the word is not intended to be inflammatory or to denigrate. To me it's simply accurate, just as you would "strip" wallpaper from a room. I don't simply say "taking out" because that can imply a more selective removal. And I'm not against that, and never have been, as of course you know. N-HH talk/edits 05:24, 19 June 2010 (UTC)

"sea of blue" is a descriptive means of addressing a substantive issue in the debate. "campaign" targets editors.  HWV258.  21:25, 19 June 2010 (UTC)

It's not a neutral term by any stretch of the imagination, especially given that it has been used in situations where said use is utterly ludicrous. --Ckatz^chat_spy 23:14, 19 June 2010 (UTC)

• "unlinking" or "delinking" are neutral; as opposed to linking and stripping. As an act of good will, we could all neutralise our language as perceived by others, couldn't we? Ckatz takes issue with some usage; I will endeavour to persuade people not to use it, where I can identify it. The important thing, to me, is that the focus be taken away from people. Tony (talk) 09:09, 19 June 2010 (UTC)
  • Yes, all of those items mentioned above, and an avoidance of anything negative about anyone. This would be a welcome move towards a more harmonious and productive environment here. I ask all regular editors to agree to this. Tony (talk) 13:33, 19 June 2010 (UTC)
    • All of these words could be substituted with more neutral items, couldn't they? Instead of "sea of blue", talk in terms of link/non-link ratio, or signal-to-noise ratio in a technical sense, or dilution. This is not hard if it keeps the peace. "Stripping" is an emotive word, to me. Tony (talk) 03:44, 20 June 2010 (UTC)

Let me make two suggestions to start this:

• One: It is very useful to note that wikilinks that start with : (which we use for linking to image files without showing the image, for example) will still show up and be linked. Any bot/script that is dealing with removal of links should ignore such links, as this gives a natural way for editors to tag a link as "opt out" from the delinking process.
• Two: We should have a discussion/RFC to determine the appropriateness and removal of "common geographic names in the English language" including what terms consistent this. This seems like the only assured area that a bot/script can perform delinking within prose due to the obviousness of it, and also appears to be an area where there is mutual agreement for this. If there are clearly other classes of words that can be considered at the same time, we should do that too, but we need very clear understanding of what these classes or lists will be and how they will be applied. Let's not try to solve all of the best types of linking/delinking in one go, but instead what can easily be done via bots/scripts. Assuredly any other types of links are going to require more editorial monitoring and the like. --MASEM (t) 14:11, 19 June 2010 (UTC)

Having read this entire thread, can I make a few points to narrow things down? (They're not really related)

• Scope. Can we narrow the discussion explicitly to geographic names, since that seems to be the main focus?
• Presence indicates relevance. The fact that a term is mentioned at all in an article gives weight to its being linked. Having translated a number of French Wikipedia articles, I have some sympathy with the idea above that they tend to link more over here than we do at EN:WP, even excluding the date linking and other kinda mechanical linking that they do more than we do. However, you will find that when places are referred to "France" is hardly ever mentioned, let alone linked. This makes sense: an editor has decided that it's not worth saying that Rouen is in France, so there's no need to mention France. That is, the presence at all of a theoretically redundant term nudges it in favour of being linked at first use (it does not make it a black-and-white decision by any means, of course). If we write "Rouen, France" we should either pipe the whole phrase (which is hardly WP:EGG) or also link France. And on the flip side of that coin, since we don't then say that France is in Europe The World The Milky Way The Universe we neither can nor should link those terms.
• Accessibility. There are two practical points that seems to have been missed about "you can get to France from Rouen" and so on. Those on who have limited ability to interact with the browser (e.g. because of some disability or because they are have limited technology such as a mobile device) might not find it particularly easy to jump to via a search box where they have to type it. On the other hand, those that use assistive technologies such can find links a hindrance since they might be spoken or displayed in a way that disrupts the article's flow. Also, not everyone has a 300Tb/s connection, and having to request the Rouen article JUST to get the France article up wastes bandwidth, even if only a small proportion of the article need be downloaded to be able to link through.
• Sea of blue. Can we just dismiss this since nobody (that I can see) is suggesting (for the purposes of this discussion) that links should be increased (beyond restoring links that were once present), over whether they should be automatically or semi-automatically reduced. In any case, presumably the WikiMedia engine, a skin, a client-side script or other intermediate technology could in theory be fashioned to link every word or phrase that had a corresponding Wikipedia article, on the fly rather than in the stored version. This would in any case fail to be as helpful as where editors manually put in piped links for things like natural reading order, specificity, disambiguation, and so on.

Just a few suggestions on how we might limit the discussion to what I see, as an interested but somewhat passive observer, to what are the main points. Si Trew (talk) 14:10, 12 July 2010 (UTC)

Past RfCs

As an aside, looking back over the archives, I see mention of previous RfCs. Can anyone provide pointers for them - just so we don't reinvent the wheel? --Michael C. Price ^talk 03:02, 28 September 2010 (UTC)

The new exception for repeated links for "See also" and "External links"

After having seen many helpful "See also" and "External links" sections deleted by well intended editors, attempting to "tidy up" articles, I personally have often been frustrated by this.  Often in lengthy articles, these sections are life-savers (well actually huge time-savers) for me and I don't understand the need to delete them in so many articles, or to bar links found above from being repeated in them.  To me, the logic of keeping these sections and allowing them to have doubled links in them seems to weigh heavily in favor of this, rather than the apparent logic that we might possibly consume too many gigabytes for this, or possibly make an article 'look' overly long.

As such, I've gone ahead and added this exception for "doubled links" to the list of "Repeated links exceptions".  Any comments on this new exception would be much appreciated.   Thanks,

Scott P. (talk) 12:59, 23 August 2010 (UTC)

• I've seen what you've done at Gary Renard. It seems to me that in this relatively short article, you'd be pushing the boat out saying that such links are 'life-savers'. These links are but repeat links of a secondary subject already very prominently linked in the lead, and I feel are strictly unnecessary. More specifically, I feel that the choice of links to be included in the 'See also' section ought to be those which are relevant but cannot be worked into the article, or which are inadequately piped, but whose inclusion would add value to the subject --Ohconfucius ^¡digame! 14:27, 23 August 2010 (UTC)

  Navboxes would be a great way to be used as time-savers for navigation, but hiding them below references, "further reading", and external links makes them nearly useless for this purpose. But putting them before references is something which the fellows at WP:LAYOUT will never allow you to do. A. di M. (talk) 15:54, 23 August 2010 (UTC)

• I also tend to disagree with Scott.  I am one of those editors who reduces or removes See Also sections when they do nothing but repeat links already included in body of article.  I have gone ahead and reverted your additions to the MOS until we reach consensus.TheRingess (talk) 17:08, 23 August 2010 (UTC)
  • I also tend to remove or reduce See Also sections when they simply repeat links unless they are important major topics directly related to the primary subject but which have not been directly mentioned in the article.  It's also important to keep in mind that external links are governed by WP:EL and WP:SPAM. Buddhipriya (talk) 17:22, 23 August 2010 (UTC)

So far this looks like I lose.  One for, one neutral, and three opposed.  Argh!  Apparently I'm in a small minority on this.  My wife says I must have OCD, because I always like to have things in the same places where I can easily find them.  Perhaps she is right.  Now where were those slippers again???? Scott P. (talk) 23:00, 23 August 2010 (UTC)

I'm all for minimising redundancy (in links and in words). I am very demanding of readers, expecting that they actually read an article, not just flash through it or treat it like bumping against a goal post. Tony (talk) 23:04, 23 August 2010 (UTC)

I too like to read an article thoroughly once.  But when I remember that there was a certain link in an article I read yesterday, or last month, I feel a bit cheated when some well-intended cleaner-uper has removed or pared down a previously good "See also" section, just to assure that I must once again re-read the entire article each time I refer back when all I wanted was the link.  So much for the wonders of the speedy internet. Hah!  Scott P. (talk) 23:11, 23 August 2010 (UTC)

You could just type it into the search box? :-) Tony (talk) 23:20, 23 August 2010 (UTC)

• . I don't know how far I should go in Scott's edit history to find the sort of concrete example used to justify adding repeat links to the 'See also' section – a large article with a link so deeply embedded it would be difficult to find again. What he did at Georgina Lightning was IMHO another set of redundant linking of easy-to-find information which might make me more likely to ignore 'See also' sections in future when I see them. Ohconfucius ^¡digame! 01:53, 24 August 2010 (UTC)

This is nuts....why are you making Wikipedia harder to use? If someone can't find a link because it's been tidied up, you've done a real disservice. Not all out readers are familiar with how best to use Wikipedia, and to be very demanding is nuts, flat out bonkers. Wikipedia should be easy to use, for everyone...and that includes easy to find links. Anyone reducing see also lists overly aggressively is imposing their personal reading style and usage habits on everyone reading Wikipedia...it's not up to you to train our readers, it's up to you to make Wikipedia easy for everyone to use regardless of their level of proficiency, good grief. RxS (talk) 04:49, 24 August 2010 (UTC)

Yes, I agree,  And so obvious.  But still the insane crusade of stripping links continues, making Wikipedia harder to use. --Michael C. Price ^talk 06:29, 24 August 2010 (UTC)

Its not as insane as you might care to suggest. People now (as opposed to two decades ago) realise that there can be noise pollution, light pollution, and there is such a thing as link pollution for web pages. Fortunately, we don't actually see much of a problem elsewhere in cyberspace, as the commercial world is well aware of potential problems caused by excessive linking. It's one important function of webmasters, who optimise linking among other things. Only here, at 'the encyclopaedia anyone can edit'©™ are there still a significant minority which appears to be ignorant of that fact. Ohconfucius ^¡digame! 06:45, 24 August 2010 (UTC)

And that is the problem: that you, and Tony, see this is a "fact", and that anyone who disagrees is merely ignorant and needs "educating". --Michael C. Price ^talk 07:56, 24 August 2010 (UTC)

Peas porridge hot, peas porridge cold, peas porridge in the pot 9 days old.  Well some like it hot, and obviously others like it cold, and since four like it cold and only three like it hot that apparently makes us three a bunch of orangutans, at least in one person's view.  Let them have their cold porridge, I've got other trees to swing from. Scott P. (talk) 08:35, 24 August 2010 (UTC)

Can I have some room-temperature peas porridge? A. di M. (talk) 16:35, 24 August 2010 (UTC)

• Could you two avoid emotive, exaggerated statements and analogies? Accusing other editors of being "nuts, flat out bonkers", is not helping calm, rational debate. And Ohconfucius, I wouldn't use the word "ignorant" for a second. People have their own point of view. Please tone down the language. Tony (talk) 08:59, 24 August 2010 (UTC)

• • Could you, Tony, avoid making statements such as You could just type it into the search box? :-) since, by that logic, we should remove all links? Of course you were told this sometime ago, but to no effect. :-) --Michael C. Price ^talk 10:25, 24 August 2010 (UTC)

My statement was not a breach of WP:CIVIL. It is not much to ask that people treat each other politely. Suggesting that the search box is an alternative to inserting low-value links is a perfectly valid argument; I'd like you to consider it. There is the matter of diluting high-value links, which should be of concern to all of us. It is very difficult to get readers to click on any links at all. Part of the service we offer to readers is to select the most useful links. Otherwise they'll just ignore all of them. Tony (talk) 12:55, 24 August 2010 (UTC)

Who said it was a breach of civility? You're getting paranoid. --Michael C. Price ^talk 09:21, 27 August 2010 (UTC)

Back to the topic, I oppose the suggested (and reverted change), being one of those who find huge chunks of repeated links unhelpful, unnecessary, etc. Dougweller (talk) 19:02, 27 August 2010 (UTC)

Pros and cons

I would like to see a clear listing of pros and cons of repeated links in see also sections, since the discussion above got a bit astray. I'll try to provide some below from the statements above and my own opinions. Please add more if I missed anything.

for

1. Readers don't have to actually read the article to find them

   Some above say we should demand effort from readers; some say we should make it easy. Personally, I don't think we should be making judgments on the amount of effort readers should have, but rather on the practical usefulness of being able to skip reading the article. A previous proposal included this, stated as: The "See also" section may repeat very important links from the article's body, but only if it is judged likely that a reader may use the page solely to find the other link (and hence, may not read the entire article in order to find it).

   I personally don't think that, with the increasing problem of information overload (of which Wikipedia is part!), we shouldn't stand in readers' way to other articles if that's what they are seeking.

2. It's what a random reader people might expect of a See also section

   The very expression "see also" conveys "relevant related links", not "relevant links that weren't used in the text" (especially since most of them will be included in the text).

against

1. If a reader is searching for a different term, he might as well use the search box

   This was argued against, above, as being a rationale that would remove all links. Besides agreeing with that, I add that sometimes the reader might not remember (or even know) the precise title of the article they seek. A navbox would serve this purpose, but what about when there's only a handful (say, 5 or less) tightly related topics?

   "the search box is an alternative to inserting low-value links", Tony says. But the point is to add high-value ones (if they're important enough to be mentioned in the article text, they certainly don't lose their importance by being repeated. They might become redundant, but not uninmportant -- and arguably still remain useful)

2. It might amount to link pollution

   I'd say that this is another issue, one of too many links in the see also section, rather than redundant links. The amount of links ultimately boils down to common sense. There's even a {{too many see alsos}} already in use.

Also, I would like to point out some arguments from past discussions which I think describe well the need for discussing this exception:

• (...) by looking at your counter-proposal, with its detailed exceptions and exceptions within the exceptions, I cannot help but think that what you wrote there is a bot. In other words, it's simply calling for somebody to say, "Hey, that's a nice functional specification, I'll write an AWB after it." Which is exactly what you don't want to happen. The counter-proposal boils down to: "Avoid linking to the same article twice within a single article and use common sense." Since the latter part is implied in any WP guideline, the present version is the one I support. (...) PizzaMargherita
• the very fact that we haven't yet seen AWB-enabled changes tells me that in large part, the existing wording is working as designed and that readers do understand that this is a question of balance and judgment. Rossami
• my formal training was in computational linguistics, so I've thought about these issues a lot—and that background makes me very conscious of the limitations of technology when dealing with natural language. I don't think my counterproposal is very amenable to a bot. You're right that it's a careful specification—I think we should strive for that when possible in style rules, as it cuts down on disputes. TreyHarris

These points, and the fact that the current rules are indeed being applied my many as hard-and-fast principles (and yes, made their way to AWB), make it clear, imo, that a specific exception is in order for cases where the benefit exceeds the drawbacks. --Waldir ^talk 13:41, 1 October 2010 (UTC)

I like see also sections. They are useful, irrespective of what links appear in the rest of the article and should be independent of it. --Michael C. Price ^talk 19:18, 1 October 2010 (UTC)

I'm sorry, but even though we agree on the issue, I must say that "I like X" is precisely the kind of language I was hoping to avoid with this subsection. You say that they are useful (which I agree with) but don't back your claim with specific examples/use cases. That unfortunately makes it easy for other people to dismiss your comments... --Waldir ^talk 15:35, 2 October 2010 (UTC)

Do we have to cite specific example to make a generic point? I doubt citing specific examples would help, given how this article is just (widely divergent) opinions, not facts. I could have explained my reason in more detail, though. So.....

.... I like see also sections. It is useful to have the most pertinent or interesting links collected together in one place, without having to hunt through an article for links that only appear once, disparately.

--Michael C. Price ^talk 15:47, 2 October 2010 (UTC)

• The name says it all. It implies that other subjects which may be relevant to the article's subject, but are not dealt with or linked to, may be linked there. Using this as another place to bury repeat links is really stretching tde definition. --Ohconfucius ^¡digame! 14:04, 3 October 2010 (UTC)
  • Aside from the name, what is your objection?--Michael C. Price ^talk 14:59, 3 October 2010 (UTC)
    • The "See also" is of particular use where it is grammatically hard to insert a focused link in the main text without a deceptive pipe (Agriculture in France piped to "France", which I assure you no one will click on). I don't quite see your difficulty in "hunting down" links. Do you mean you've thought of the link target before you've come across it? My advice would be to read the article as a whole, or use the search box instantly if you wish. Tony (talk) 15:42, 3 October 2010 (UTC)
      • You're welcome to your opinions, just as long as you realise they are just that, opinions. --Michael C. Price ^talk 15:48, 3 October 2010 (UTC)
        • And you yours. Thank you very much for expressing a view. ;-) --Ohconfucius ^¡digame! 15:57, 3 October 2010 (UTC)
          • Yes, my views are just opinions. Are you able to say the same, or are you going to claim that they are backed up by "studies" or some such pseudoscience? --Michael C. Price ^talk 16:38, 3 October 2010 (UTC)
            • Oh, well… since you don’t seem to be in a constructive mood today, Michael and apparently seem anxious to come here, make oodles of waves, insist on having things your way, and flippantly dismiss others’ opinions as being “just opinions”; you, sir, are welcome to be tendentious, as long as you realize that Wikipedia’s policies do not forbid us from just completely ignoring you. Greg L (talk) 16:17, 3 October 2010 (UTC)
          • Your "departure" didn't last long, did it? --Michael C. Price ^talk 16:38, 3 October 2010 (UTC)
            • My “departure” was in regard to the threads at the bottom of the page, on which no progress was being made because no one wanted to do what you insisted on: begin linking the holy crap out of articles again. I see that your nonsense springs eternal at old threads on this page.

              The metrics make sense to me too, Michael. There is a clear trend among all the different-language Wikipedias to reduce the links in articles to just those required to enhance the readers’ understanding of the subject matter. After rising to absurd proportions, the proportion of words linked in articles has declined on en.Wikipedia by one-third in the last four years. The downward trend has been consistent as editors all across the project realized we had created articles that looked seas of blue turds. These statistics apparently don’t please you. As they say in the military: “So sad – Too bad.” Greg L (talk) 16:55, 3 October 2010 (UTC)

              • Those are not the pseudoscientific claims I was referring to. If you're going to play dumb over claims made in the past here, that's your business. As for the claims you are now making, so what? If link density has gone down that is no argument that it should continue, or even that it was wise in the first place. --Michael C. Price ^talk 17:33, 3 October 2010 (UTC)
                • Playing dumb? Well, you sure are obnoxious, aren’t you? The link density is going down because the community wants it that way and doesn’t see things your way. This place is ruled by consensus. Period. Get that through your head. And “consensus” does not equate to “Whatever Michael wants ‘cause he hangs onto issues like a pit bull on someone’s throat.” You don’t get your way be being tendentious beyond all comprehension and flogging a dead horse until the heat death of the universe. You think the current practices on Wikipedia are unwise. Well, that’s all *extra special* but doesn’t matter since precious too-few people share your view that everyone should start re-linking the living snot out of articles. And why do they feel that way? Because doing so would be an unwise, lousy technical writing practice. I can’t help it if you can’t see that. Greg L (talk) 18:15, 3 October 2010 (UTC)
                  • Your carefully reasoned response did at least give me a good laugh. Thanks.
                  • But if you're so sure consensus is with you, why the hysterical reaction and violation of AGF to the suggestion of an RfC? I'm not the only one to have pointed out that a small cabal of 4-5 editors here does not amount to a consensus. We shall see. --Michael C. Price ^talk 18:25, 3 October 2010 (UTC)
                    • I don’t know to whom you think is in violation of WP:AGF, but I hope you aren’t suggesting it is me. I know you mean well. But like any club with an open membership (Wikipedia so qualifies), it gets a small but regular stream of people convinced they have stumbled upon a better way to spin gold yarn and who can’t fathom why the natives here don’t start hoping up on down with our spears exclaiming how “The dude in the pith helmet has shown us the light.” All the other-language Wikipedias are experiencing a drop in the link percentage and en.Wikipedia’s rate of decline is right smack in the middle of the pack. You wanna change the world and are acting all frustrated. I can’t help that. A handful of us bother to give you the time of day; I guarantee you there is an ocean of others out there who quietly embrace the reasoning and aren’t suckers for coming to this talk page to get a belly full of witty banter with you. Greg L (talk) 19:39, 3 October 2010 (UTC)
                      • GregL has a point although I don't like the way he expresses it. If you want an RfC so much, you don't need anybody's permission, especially after this much discussion. Whether anyone wants to comment further at the RfC, is something else. Art LaPella (talk) 20:01, 3 October 2010 (UTC)
                        • The problem is that GregL (and others) regard linking as "spinning gold yarn". Until that attitude changes how can we make progress? --Michael C. Price ^talk 03:36, 6 October 2010 (UTC)
                          • I suppose by gathering a consensus to overrule GregL (although that consensus might agree with him instead), in addition to dealing with the civility problem. What would you suggest? Art LaPella (talk) 04:30, 6 October 2010 (UTC)
        • Michael, everything that drops from my lips is "my opinion". Tony (talk) 16:03, 3 October 2010 (UTC)
          • yes, but in the past your opinions were claimed to be supported by "research" and "metrics", which we now know to be groundless. --Michael C. Price ^talk 16:38, 3 October 2010 (UTC)

Linking words in quotations

Are there any special rules for linking in quotations (quoted passages)? Should they be avoided altogether? I give here a random example. In an article about, let's say, "Christian morals" would the links to sin and death in the following quotation be OK? "For everyone belongs to me, the parent as well as the child—both alike belong to me. The one who sins is the one who will die.—Ezek. 18:1-4, TNIV" -Mycomp (talk) 23:34, 8 September 2010 (UTC)

I never use links in quotations. Instead, I may add what seems to be the relevant link in the "See also" section. Sometimes the link is to a disambiguation page. 69.3.72.9 (talk) 16:16, 12 September 2010 (UTC)

This has been discussed more broadly at MT:MOS in the last year, with an increasing number of editors leaning toward permitting links inside quotations when they will be very helpful for reader comprehension, and removing links where they don't add much. E.g. a link to Lilliput is generally reader-helpful in a quotation that include the word "lilliputian" because most readers have no idea what it means, while a link to "mathematics" is not likely to aid readers and will just distract from parsing the quotation in context. It's important that when such links are used they do not introduce any kind of bias or PoV pushing, per WP:NPOV, nor push readers toward a novel interpretation or synthesis per WP:NOR, but such problems are not very common and easy to spot and fix. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 01:54, 17 September 2010 (UTC)

Yes, I agree. Links inside quotations are okay. Especially where the speaker is using allusion - links there can be very helpful. --Michael C. Price ^talk 03:39, 6 October 2010 (UTC)

• In the case cited, 'die' or 'death' should not need linking, unless you were contemplating a pipe to Death_and_culture#Philosophy.2C_religion_and_mythology. --Ohconfucius ^¡digame! 03:53, 6 October 2010 (UTC)

Major announcement: the Silliest Wikilink of the Month awards

Resolved

 – Just an FYI; no MOS issues to address.

Our judge, Ceoil, will soon announce the winners of the awards for August, July and May 2010, and at the end of this month will announce the winner for September.

He has agreed that we might then change the focus of the competition from individual wikilinks and small groups of wikilinks to whole articles that are badly overlinked. Inevitably, those valuable editors who perform gnoming services are confronted with overlinking throughout whole articles (particularly of "dictionary" items). In almost all cases, this has arisen earlier in WP's history, when there was no coherent strategy for maximising the utility of the wikilinking system. It's a lot of work to clean it up, and the Silliwilli awards was set up to encourage this work.

Therefore, we have decided that from October 2010 onwards the awards should be judged in terms of whole articles. Competitors will still be asked to list individual links (but expanded to six of the funniest, most useless, most inexplicable individual links in the article, as an example of the entry); however, the removal of overlinking from the whole article will be the sole determinant in the award. We expect this will reduce the number of entrants each month, which will compensate for the extra work by the judge in analysing the entries. Tony (talk) 04:07, 10 September 2010 (UTC)

Seeing green

Resolved

 – Question answered.

Why is some of the text in this article showing up as green? In utter cluelessness, your friend, GeorgeLouis (talk) 00:24, 12 September 2010 (UTC)

{{xt}} Dabomb87 (talk) 00:24, 12 September 2010 (UTC)

I pose a politely worded query in English and am answered with some kind of odd symbol? Curiouser and curiouser. GeorgeLouis (talk) 03:51, 12 September 2010 (UTC)

Apologies. The above template that I linked makes the text display in green font. It is used to highlight examples, I believe. Dabomb87 (talk) 04:04, 12 September 2010 (UTC)

The main MoS has long used green text in Georgia to identify example text. Using italics or bold or whatever to identify examples doesn't work because sometimes the italics or bold is what's being demonstrated. Wikipedia_talk:Manual_of_Style/Archive_110#An_idea:_markup_for_bad_examples was the first discussion about also using red text to distinguish wrong examples; there were others shortly after. Adrian J. Hunter^{(talk•contribs)} 04:11, 12 September 2010 (UTC)

Dispute

Whoa, someone's mucked this up really badly. It's completely absurd to suggest that people should not link languages and geographical names, etc., within reason. Per WP:POLICY, Wikipedia guidelines describe actual, current, observable, consensus-based best practices on Wikipedia; they do not advance one person's or one camp's opinion of what WP best practices "should" be by proscribing common, or prescribing unusual, editor behaviors.

It is very clear, simply from looking at articles and infoboxes, that the vast majority of WP editors believe, and our readers expect, country, city and other geographical names, language names, and other proper nouns, to be linked at first occurrence the vast majority of the time. WP:MOSNUM, the controlling guideline on numerics, strongly suggests always linking first occurrences of currencies, units, etc., in any case where confusion could possibly occur at all, and does not suggest that more common usages shouldn't be linked (a common but not universal practice – i.e., one about which there is not solid consensus – and one that this guideline is not in a position to attack without a clear showing, e.g. in an RfC on the topic, that a preponderance of editors support a ban on such links). I am therefore making significant edits to the "Overlinking and underlinking" section to correct this problem and several others, including direct conflict with WP:MOSNUM and WP:SELFREF, among other issues, including bad list style, redundancy, etc., etc. MOS-watchers need to keep a much better eye on this page, as it has clearly been PoV-pushed in a reader-unhelpful, anti-linking direction.

— SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 01:37, 17 September 2010 (UTC)

I think you’ve been too bold with your edit. The community has been through this at great length with wide input and you just changed something all by your lonesome. The wording there was no accident. To avoid over-linking articles, we avoid linking common terms everyone is familiar with and focus only upon the links that will truly add value. For instance, if there is an article on “Scientific goofs”, there might be this sentence: American scientists in the 1960s thought they had invented polywater. Given the nature of the article, only one link has value: polywater; all the others are superfluous and add no value. The rule is simple: link only those items that enhance the readers’ understanding of that subject. If it is an article on occupations, one would link the first occurrence of “scientists.” If it is an article on countries, one would link the first occurrence of “American.” But just linking the first occurrence of pretty much everything regardless of context results in articles that read like The sewer cover in front of Greg L’s house. Greg L (talk) 02:36, 17 September 2010 (UTC)

Without delving too deeply into SMcCandlish's argument, I must simply agree with Greg L that there is simply too much linking going on. Sincerely, GeorgeLouis (talk) 02:48, 17 September 2010 (UTC)

If the rule is going to be "the names of major geographic features and ... should not be repeatedly linked", this either duplicates WP:REPEATLINK, or comes close enough to duplicating to be confusing. It's true that words like "American" are often linked, but I don't know why anyone would click them. If we're trying to make the guideline fit the usual practice, the usual practice is to link major countries throughout the article, not just once. When my AWB software unlinks such words, I often leave the first link to a country, religion etc. depending on how closely related it is to the article's subject; for instance, George Washington links to United States in the infobox, though not in the first paragraph. The Main Page is another undocumented exception; see Wikipedia talk:Selected anniversaries#Country links. Art LaPella (talk) 05:56, 17 September 2010 (UTC)

I agree with Greg L's summary.  HWV258.  07:25, 17 September 2010 (UTC)

[outdent] To respond to all of the above at once (using "you" generically):

1. WP:BRD exists for a reason. There really is no such thing as "too bold", per WP:BOLD and more importantly WP:IAR. This is a wiki. Being bold does not do damage, and criticizing editors for being "too bold" is rather nonsensical.

2. As I've mentioned on my own talk page in response to Tony1, for the first time in several years I am invoking IAR, and ignoring the dafter parts of this section of the guideline, because they do not represent consensus at all, only what Tony describes as an "uneasy balance" between argumentative factions on this talk page. I will continue to link, sensibly, as I have done during my entire editing history and as most other experienced editors do, regardless what the perennially disputed section says, because that's actual de facto standard WP practice and has been for years, since before I was even an editor here. And I see precisely zero evidence that consensus has changed, only that certain parties here are tenacious, and through long, bitter dispute have worn down more sensible stances to agree to a compromise position that doesn't actually satisfy anyone. The section gets in the way of my ability to improve the encyclopedia, and that is precisely the circumstance for which the IAR policy was codified.

3. I'm not going to pitch a fit about being reverted (I expected it) so long as the "D" in WP:BRD is engaged in here. You can't have the "R" without the "D". My edits were well-explained, based in policy and much more widely accepted guidelines than this one. The responses thus far have been a) personal opinion not rooted in any such bases, b) just "don't rock the boat" conservatism, and c) musings and statements that are non-responsive to most if not all of the issues I raised. None of these reactions makes for a strong position, alone or combined.

4. I don't care if the wording there was no accident. It is still awful. The fact that something boneheaded, confusing and user-hostile was the product of a sausage-making committee full of people who don't agree with each other on much of anything is neither surprising nor any excuse. The section conflicts with WP:POLICY, WP:MOS, WP:MOSNUM, WP:BIAS, WP:CONSENSUS and WP:SELFREF (at least – I'm probably forgetting some) on a large number of points, is longwinded and in parts redundant, is invasive of MOSNUM's scope, has poor grammar in parts, is palimpsestuous and hard to parse, is clearly biased and prescriptive/proscriptive, and does not represent the actual practice of the majority of experienced, good faith, intelligent editors, among other issues.

5. I agree that there is too much linking going on. I unlink stuff all the time. It's one of my most common types of mainspace edit, in fact. Believing that overlinking is at play on WP does not equate to a mandate to introduce ridiculously vague and over-broad advice on (against, really) linking that contravenes very long-standing and very well accepted actual practice by the vast majority of editors. It would be much better to properly identify (WP:RFC, anyone?) what most editors do actually feel should not be linked. With that in mind, I introduced a handful of clear examples, each of which was selected because I have actually found them linked for no reason and delinked them, and had a good laugh; they were ones that stuck in my memory.

6. The "linking first occurrence of everything = a sewer" argument is a total straw man fallacy, since neither I nor any one else since the pre-Web days of the Xanadu Project has suggested doing so, that I know of. So, please don't be exaggeratory and melodramatic. In actual fact, I was quite explicit about what should be linked at first occurrence and gave, also, clear examples of what not to link at all. A reasonable person could argue that my take was perfect or too exclusive or too broad. Worth discussion. The interesting part here to me is that the language as it stood, and stands again since I was reverted, actually commits the "sin" you point out, in the inverse: It effectively suggests that nothing should ever be linked at first or any occurrence except under very restrictive conditions, that do not at all match how Wikipedia has operated from day one to the present (nor virtually all wikis of this "informational, researched articles" format, such as Memory Alpha, Battlestar Wiki, etc., etc.).

7. If you won't "delv[e]...deeply into [my] argument", and we don't disagree that there is too much linking going on, your comment about not delving but feeling there's overlinking going on is basically meaningless, since my edits do not suggest that too much linking isn't going on (in fact, I clarified how it is going on), and you don't present any argument against the specifics I changed. Unexplicated "me too" one-liners like that do not help build consensus, one way or the other, be they in formal !votes like WP:AFD, where they are mostly ignored by closing admins, or in informal discussions like this one.

8. If my changes resulted in a redundancy with other wording in the guideline (which I did not set out to edit from top to bottom, only uni-sectionally), this is not an indication that the change was bad, only that further editing needs to happen outside and/or inside this section, since the change simply made the document agree with itself better, instead of going off on a wild tangent trying to ban links that most editors consider completely normal. This is actually precisely what I would expect, given that much of this document has been stable for years, but this section is a frakfest of agenda-pushing and emotive argumentation, with layers upon layers of barely- to totally-incompatible edits, with greatly varying degrees of common sense applied. If moderating the extreme pre- and proscriptions in the contentious section makes it come more into line with the stable rest of the document, this is a very, very strong indication that the section in question has been badly off-kilter and getting worse.

9. It's not important whether or not you find a link to "American" useful or not, understand why it is there or why someone else might appreciate it, or would ever click it yourself. This encyclopedia isn't written specially for you. It's written for everyone who can understand some English, anywhere in the world. This includes people in censorious and propagandistic countries like China (P.R. of), where actual facts about the United States and Americans are often generally unavailable or distorted, except for those who have figured out ways around the censorship and gotten here to get sourced facts. Another way of looking at it: A lead section intro like "Ndele P. Mbebe is a Botswanan professor of physics teaching at Rutgers University since 1998..." is what I would guesstimate 95% of experienced WP editors would write. Someone from Botswana might not see any point to that first link and would never click it, since they already know all about Botswana. A physicist might feel likewise about the 2nd link, and so on. Most other people would not have such an "I don't give a hoot" reaction to such links, and see WP:BIAS for why making the US some magically special case is not acceptable. Taken to its logical conclusion, some might use this position to suggest that linking to Iraq shouldn't be done except where Iraq is really, really important to the topic at hand, because Iraq is all over the news all day long, and we all already know about Iraq. But I saw poll results in an Associated Press story about a year or so ago that reported that only a tiny percentage (less than 10%, maybe less that 5%, I forget) of Americans could even correctly identify Iraq on a map that showed borders but no country names. The point being, there are also sorts of reasons in favor of linking to articles on significant topics that provide context to an article subject - being a Russian and being Maori produce completely different worldviews; A Galician writer and a Castillian ("Spanish language") writer, both from Spain, will produce literature with a different "flavor", audience, social impact, etc. And so on.

10. The usual practice is certainly not to link every occurrence of countries and such. It's a common new editor mistake that experienced editors like you and me correct on sight. Please do not confuse experienced Wikipedian standard practice (codified or otherwise) with unhelpful noob editing behavior that happens to be frequent (and frequently undone). It takes virtually all editors (including me back in my wide-eyed time, and surely you in yours) some time to fully grok when and when not to link and how to do it in ways that aren't misleading, confusing, distracting, leading or otherwise unhelpful. When I refer to common, consensus-accepted, observable best practices I'm obviously referring to the former kind of editing practice, not the latter, and actually added noob-helpful information on what not to link.

11. PS: A side point I must stress, and I have to do this in multiple forums from time to time: Infoboxes are entirely dependent upon and subordinate to articles' main prose but (important here!) they are severable, distinct entities. All information that appears in an infobox should also appear in the prose, in one form or another (usually more developed, instead of summary form), and be sourced in the prose, even if sources are also in the infobox (which is usually unnecessary). I realize that, especially in biographies, this ideal is often not attainted, at least not until GA or FA review. But it must be goal for a very simple reason: WP content can be repurposed in any way (within the bounds of the license) via any means for any purpose. This includes recycling but filtering the wikicode to re-present the prose in other non-MediaWiki-wiki, blog, e-book, etc. marked up formats that may preserve styles and links, but strip all infoboxes, navboxes and other adjunct template content! D'oh! The upshot of this is, of course, that we cannot at all guarantee that a link in an infobox will remain "the first occurrence" on the page as it is reused elsewhere. Because users (here) often read only the infobox when skimming for quick information, first occurrence in an infobox of something worth linking should be linked. But because many users ignore infoboxes as "noise", and are here to read an in-depth article and already know the summary details of the topic they are researching, the same is true of the main prose - link the first occurrence in the article proper. I've been writing/editing articles this way for 5 years (linking in main prose and in infoboxes as if they were separate entities with a parent-child relationship, which <gasp!> is actually the case). I cannot remember one single case of anyone reverting me on this practice. Not one. It's self-evidently the only sensible way to do it, if and when one understands and considers WP's broader, off-site userbase, many of whom don't even know they are looking at content ultimately from WP – and without infoboxes and the like – unless they read the fine-print attribution. Infoboxes and navboxes (which also link to things that may already be linked in main prose) are tools – mini-pages, if you will – that are separate from, even if subordinate to, the main article prose. Personally, I think that the "no subpages" policy should be modified to always put infoboxes in a /infbox subpage. This would help remind people of the nature of an infobox and its relationship to the parent article, but I guess that's neither here nor there.

That's all I can think of for now. Apologies for the length but I want to be as plain and explicit on all aspects of this as possible, to minimize the amount of time people argue past one another and don't understand where the other side(s) is/are coming from. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 08:44, 17 September 2010 (UTC)

I still agree with Greg L's summary.  HWV258.  09:28, 17 September 2010 (UTC)

Moreover, attempting to justify your actions by citing WP:BRD, WP:BOLD, WP:POLICY, WP:MOS, WP:MOSNUM, WP:BIAS, WP:CONSENSUS and WP:SELFREF in an enormous smoke screen amounts to nothing more than WP:I MADE IT BLUE SO IT MUST BE TRUE.

There is no escaping the fact that the tide has been turning against overlinking the last few years, the wording you made a colossal change to was the product of vigorous debate and compromise by many editors over a protracted period of time, and that wording truly represents the best consensus to date (notwithstanding that you oddly cited WP:CONSENSUS in an attempt to justify your unilateral, undiscussed, colossal change to the guideline). It doesn’t matter if you think “it is still awful.” It appears that four of us (GeorgeLouis, Greg L, HWV258, and Tony1) are in one camp and only three are in the other camp (you, your links, and your flotilla-like posts). Stealing a stunt of yours, that is WP:CONSENSUS. And that consensus is to keep the existing wording for the moment.

Now, you may keep discussing the matter here if you like. I suggest you keep your arguments shorter because we are all volunteers here and time is limited. If you have an idea that makes sense and gains traction with others, great. If not, then as they say in the military: “So sad – too bad.” There is no politically correct requirement that others admire your ideas and edits as much as you do, nor should eventually tuning you out be construed as an invitation to you to wade back in and do as you please with a flame thrower. Greg L (talk) 15:59, 17 September 2010 (UTC)

You contradict yourself. You claim a 4 vs. 1 situation^[#foobar|*], yet simultaneously claim that there was a "vigorous debate by many editors over a protracted period of time", resulting in the current version (a claim echoed, at my talk page, by Tony1). Can't have it both ways. If you really prefer to use numbers less than 24 hours old, well okay: A paltry Alamo of 4 editors in support of the current language, versus thousands upon thousands of experienced editors continuing to link exactly as they did a year ago and five years ago, indicates three, and only three, things with undeniable clarity: 1) Few Wikipedians give a damn about what this document says, indicating that either this guideline and the issues it attempts to address aren't very important at all, or that editors have lost faith with this page because it doesn't match actual practice (I strongly suspect the latter, and if anyone tried to MfD this as unimportant, you know it would be a speedy keep, and proof that the "unimportant" theory is false). 2) Too few editors have had any input into any of this (which I already knew and is why I've suggested, twice, an RfC) for you to declare that any kind of meaningful compromise has been reached (it's too thin a base of support to even have changed the guideline from how it read 2 years ago, much less to declare a new consensus on what it currently says). 3) Sparsity of participants aside, there clearly is no consensus anyway, by definition, since opinions are near evenly divided (see other critics of the current wording, above and in recent archives), thus resulting in the long, heated debate resulting in a messy compromise. It is not at all clear that there has been a tide against current, experienced-editor linking levels (I agree that there has been a tide against genuine overlinking, but you, Greg, are using that as a smokescreen to cover the fact that you are trying to redefine much of, if not most, normal linking as "overlinking"). All that's really in evidence, after factoring in all of this, is that a small group editors have used this page, which was stable and well-accepted for years, as a place to push their personal anti-linking agenda. Your Johnnie Cochran jab is a non-starter, since I've already laid out in detail both above and on a point by point basis in edit summaries how all of those policies and guidelines apply and to what. If I'd simply linked them, with no explanations, I would understand your criticism, but if falls flat here. Likewise, simply rejecting, without substantive response, my criticism that the section "is still awful" is a head-in-the-sand move, since I explained in detail how it is awful, and you have failed to address even a single one of those points. Just shouting "no!" at people doesn't make a rational argument. You and I are frequently at loggerheads, and I'm resigned to that, but I believe that we both have WP's best interests in mind, and could probably work together more productively if you'd take less auto-defensive, must-fight-at-all-costs positions and actually address the points I raise (and those of others - I'm not the only one you argue to the death with). That said, I realize that it takes two to argue, and I'm known for being argumentative myself. I will endeavor to be open-minded with you and the others who you say support your views, Greg (and with HWV, who explicitly states such support).

[^* And it's not 4 vs. 1 anyway; I'm not the only recent commentator who has a problem with this section. See for example the criticism by Michael C. Price above. (If that link doesn't work later, try /Archive 9#Re: "aid understanding, avoid obvious"). There are plenty more such views expressed in archived threads.]

— SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 18:40, 17 September 2010 (UTC)

Responding to points directed at myself:

8. You agreed that "further editing needs to happen", so no significant disagreement here.

9. If my feelings for my native U.S. are a problem, then let's change the example to China, which is one of several major countries on my list for unlinking. Botswanans are unlikely to need to click an explanation of what China or the U.S. is. If they wanted to read all about China, they would have typed China into the search box, rather than read an article that happens to mention China. (Exception: If they're reading about Shanghai, then maybe they also want to read about China. If they're reading about a scientist who happens to be Chinese, then no, they don't want to read all about China.) Chinese may or may not know what Botswana is. I didn't find anything at WP:BIAS asserting that all countries are equally well known.

10. An experienced editor like me, rightly or wrongly unlinks more countries than you imply I do. The distinction between a noob mistake, and a practice opposed by a 4:1 majority (while recognizing how you challenged that statistic), is not a distinction I would want to explain during a flaming edit war of a kind that periodically afflicts the Manual of Style. Each side often considers the other to be making a noob mistake. A major purpose of having a Manual of Style is to arbitrate such dialogues of the deaf, in a way that is less likely to be interpreted as hostility.

11. The George Washington article doesn't link United States in the first paragraph. I was just describing the handiest example, not presenting it as my ideal. Art LaPella (talk) 21:36, 17 September 2010 (UTC)

8. Are you sure you want to agree with point 8? I'll take you at your word. It brings up quite a bit more than that the page needs further editing.

9. You and others in favor of massive de-linking are laboring under the very false assumption that the point of a wikilink is to "find out what something is" (someone else above said something to this effect as well). Links are to "provide all sorts of encyclopedic information about something", a much, much broader mandate. A large percentage of the time, someone clicking a link in an sentence in an article knows very well what that something "is", and are following the link because they want to look for something non-obvious about that "something" that is relevant to the context in which the link appears or (even more often) to something in the readers' mind that has been triggered by the appearance of the "something" in the context in question. Some of the most fruitful Wikipedia reading is the following of these links at reader whim and under reader individual interest. Massive delinking as proposed by this version of the guideline text is effectively robbing readers of the opportunity to chase their own interests (or would be, if many editors actually did what this guideline presently tells them to do). Many readers reading about a Chinese scientist damned well do want to get to the article about China, because if it's not a total shite article, it probably has an entire section about science and technology in China, and academia in China, and so on, that may help the reader place the original article subject is a much broader context than his or her little bio article does. Please stop assuming that everyone uses Wikipedia the way you do, which is apparently in a very linear and narrowly focused manner. All you "it's a sea of blue! it's a sea of blue! aaaaa!" people need to do is quit fighting with everyone else and turning guidelines on their ear, and just go install a few lines in your default.css that make links do something less intrusive for you than be bright blue. What you've all done is akin to tearing down a bridge used by everyone because you find it too bumpy and loud, instead of fixing your own bad shocks and bald tires.

• Bang! Cavendish, you shot yerself in the foot. When you have a Chinese scientist's bio, this is the sort of link you need – not some semi-hemi-demi-pertinent link (China) ;-). Unless we change this blinkered mentality like yours of arguing for linking the most obvious, rather than the most relevant or germane, people will continue to do so without engaging their grey matter. --Ohconfucius ^¡digame! 07:50, 22 September 2010 (UTC)

10. I'm not the only one to have pointed out there is no 4:1 majority (see below). I'm sorry I even used that back at you guys as a joke, since it's now being bandied about as if it were a valid statistic. Let's not be silly. I get your point that this camp or that camp can accuse the other of noob mistakes. I'm not doing that. I'm drawing a distinction between overlinking errors committed by noobs, including linking of countries and such every time they appear, linking of dictionary words for no reason, linking of all dates, etc., on the one hand (linking simply because it can be linked), versus carefully thought-out link made by editors who know what they are doing, including linking of dictionary words when their context in the article is very special, dates that add important information or context (e.g. 2009 in film, in a movie article), and first occurrences of countries, currencies and various other things (linking because it adds something some non-trivial subset of the readership will find helpful). This distinction has nothing at all to do with whether some experienced editors have a deletionist bent, a more stringent idea of what constitutes "overlinking", while others are more inclusionist and permissive. Those are basically two different discussions entirely. The fact that the latter case is, clearly, an ongoing debate indicates beyond any shadow of a doubt that consensus has not been reached on a firm boundary for what constitutes "overlinking". Yet, we also know for a clear fact that many things are consensus-accepted as overlinking.

(See also, below again, for an a-b-c-d... layout of how two radically different issues are being conflated here, often intentionally by those who are trying to confuse broad agreement that blatant overlinking is a negative, with illusory claims of broad agreement that their personal interpretation that almost all links to countries qualify as overlinking. Different issues.)

What needs to happen, and what I did and was reflexively reverted on, without (to date) a single substantive response to even one point raised by my edits, edit summaries and follow-on talk page material about those edits, is that the over-breadth and vagueness of this section has to be pared down from prescriptive and proscriptive agenda-pushing positions that blatantly obviously do not have consensus (because multiple parties are, and for two years have been, disputing it), with a restoration of the wording to a) what does have consensus as "overlinking", and b) what can be clearly observed as normal non-noob practice (i.e., linking most "major topics" if you will at first occurrence, and forbidding redundant linkage).

I still stand by my edits. I'm not criticizing you in particular here, Art, but not one reply on this entire thread has provided a logical, well-reasoned, evidence-backed rationale for the revert that was applied to every single change I made, in knee-jerk, blanket fashion. "I think you went to far" (to paraphrase) is not an argument, it's simply an opinion like "I think chocolate tastes good". "This has consensus, so don't change it" (to paraphrase again) when there is overwhelming evidence of no consensus, is not an argument, it's simply disingenuous nonsense (or evidence of cognitive problems). I believe strongly in WP:BRD, but at some point the "D" in that has to actually happen, in a substantive way, about the changes that were "R"'ed, or the "B" is going to come back in play, with the burden of proof shifting to the reverter(s).

11. So what? The George Washington article isn't evidentiary of anything salient here. I'd bet you real cash money that I can find versions of that article that do link United States at the top of the prose (as should be the case). I can also find many, many bio articles that do likewise right now. GA and FA candidate are likely to not have them linked, because GA and FA reviewers almost slavishly follow what guidelines like this say, whether they agree with them or not, and whether they are stable or not, because that's just how WP:GAN and WP:FAC work, for better or worse. And I agree with that process. They have to use something as an arbitrary baseline, and our guidelines are that baseline, even when they are problematic. If GAN and FAC didn't do this, it would be a woefully biased popularity contest where whoever happened to be doing GAN/FAC work that week or day got to impose their highly personal preferences as reject a great article in favor of a crappy one for totally subjective reasons. So, an example like that really isn't germane to this debate at all. The funny thing about the G.W. article is it's actually a great example of where United States absolutely, positively should be linked at first occurrence, since the US and the very concept of the US is utterly central to the historical figure, and the figure is deeply bound up with the existence of the nation! Sheesh, that link should be there even under a version of this document that were more anti-link. Thanks for proving one of my major points for me. >;-) (And contrast this case with something like the country's appearance in a phrase like "mined in large quantities in Botswana, France, Indonesia, the United States and Venezuela" at Bauxite, to make up a counterexample. I can understand opposition to, and might even be convinced to lean toward opposing, links to those countries there because they are not intimately tied to the subject. However, there is no doubt that a well-developed country article would provide sections on industries and natural resources, and that these would provide context to and information about mining in those countries, if not bauxite mining in particular, and that this information would be helpful to some subset of readers of the Bauxite article. My personal jury is still out on links of grey-area utility like that.)

[Outdent] There has not been "wide input", nor is there any "4:1 majority"; the more controversial changes are largely the result of the personal preferences of a handful of editors who doggedly pursue this goal. Furthermore, the statement "you just changed something all by your lonesome" is particularly troublesome given that the reason we are here today is due to someone completely changing the focus of this part of the guideline two years ago, "all by their lonesome". There has to be room for compromise, as requested numerous times in discussions on this page and elsewhere. Guidelines cannot be used as a means of imposing an individual vision on the project. --Ckatz^chat_spy 09:00, 18 September 2010 (UTC)

8 and 11 are about things I never said. 9 has been debated to death; I hadn't seen the previous debate before. As for 10, if there is such a broad consensus for linking everything you want to link, then I'm glad you're here trying to change the guideline; how else would we know? Anybody can say the other guy's opinions and edits don't count, because we have it thought out better. Maybe we need statistics correlating number of links with number of edits or something, although that implies that noobs conform because we know better, not because of "When in Rome". More on this at User:Art LaPella/Because the guideline says so. Art LaPella (talk) 06:59, 19 September 2010 (UTC)

Whilst you are correct that this remains a very localised discussion, where the numbers are hardly significantly, you seem to be implying that some editor unilaterally changed the guideline two years ago. Yet you fail to explain how somehow it has remained magically stable and, strangely to you, inexplicably enjoys widespread though not unanimous support. Guidelines exist to put everyone on the same footing, and I would say this one does its job quite well. Caricaturing modestly, I would say your idea of 'compromise' seems to be being allowed to ensure that Canada, well known country though it is, to remain linked throughout the project. I guess I ought to be grateful that you don't go around systematically reverting my unlinking edits. Please demonstrate, other than by paying lip service, that you accept linking to the extent we have 'enjoyed' (sic) in the past may indeed be detrimental and that the acceptance of this new reality is more widespread than you would care to openly admit. --Ohconfucius ^¡digame! 13:02, 18 September 2010 (UTC)

It has remained in this form because some of the handful of major supporters of the major wording change against linking are browbeating, incessantly repetitive attack dogs who will not let their position go nor compromise, but simply always insist, without evidence, that their position is correct, and that everyone else is wrong about everything, again without any evidence that this is true, and will never actually address any criticisms or concerns, no matter how many times it is demanded that they do so. It is impossible to have a rational discussion/debate in the face of such illogical and fallacious tactics. What usually happens is that such nonsense is ignored by everyone else, the changes are reverted, and things go on normally, as they should. But in a few cases here and there (I can name some others, though I'd kinda rather not get into it), such parties are so loud and act so effectively in small-numbered but exceedingly vocal and lock-step concert, probably coordinated in e-mail so there's no evidence of canvassing, that more rationale discourse is drowned out, and everyone else just gives up and goes away with massive headaches and dangerously elevated blood pressure. What happens after that is that the hijacked guideline is largely ignored by everyone except the special interest who have usurped it, and editors continue doing what they have been doing for years. The bad part, for day-to-day editing, is that this inevitably leads to edit-warring between established editors who know how to edit, regardless what an pseudo-consensus has to say on some screwed-beyond-recognition guideline page that's been radically altered away from actual practice, versus newer users who treat all guidelines as gospel and don't understand that wikipolitics is always at play in them. Another downside is that WP:GAC and WP:FAN will force the undoing of good editing that reflects actual de facto standard, consensus practice as evidenced by actual experienced editor behavior in the aggregate, in favor of something stupid that someone has inserted into a guideline, because both process are (necessarily) bound to follow the guidelines as a baseline of neutral arbitration and fairness. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 02:37, 19 September 2010 (UTC)

PS: Please stop conflating the two very, very different issues/arguments at play here. Virtually everyone agrees that overlinking does happen, and that it is a Bad Thing. And we all even agree on some of what constitutes overlinking, like links to dictionary words or dates, unless there's a very special reason for doing so in some limited context. The other and completely severable issue is that a small but incredibly vocal and tenacious minority want to define as overlinking links to countries and such, that the vast majority of editors consider perfectly valid links – as clearly evidenced by the fact that most experienced editors link them at first occurrence despite two years of this increasingly disputed guideline saying not to (that last part is very important)). There is no connection between these points. Yet this is the second if not third time in two days that a supporter of greatly increased delinking, in this thread, has tried to imply that they are the same in a vain hope of strengthening his/her argument. I'll spell it out as clearly as possible so no one misses it the next time it happens: Just because a) overlinking does happen, and b) overlinking is detrimental, and c) there is no substantial disagreement about these points, does not mean d) links to countries or some other categorization being attacked by certain parties to the debate actually constitute overlinking (no evidence of broad consensus on this idea, and strong evidence that the idea is controversial, thus the historically raging and now renewed disputed on this page, and widespread disregarding of this part of the guideline), e) nor that links to such topics are actually detrimental (zero evidence of this whatsoever), meanwhile f) the vast majority of experienced editors continue to edit and link the way this guideline advised before it was heavily modified by "authoritative and deletionist" (see original quotation above) anti-linkers. This "well, we both agree that there's overlinking and it's bad, ergo we have to stop linking countries like the current text says; I win, argument over" nonsense is not a valid argument. It's fallacious from top to bottom, and it's not fooling anyone. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 02:58, 19 September 2010 (UTC)

Ckatz, that's kinda what I thought. See just above on this very page here (or there if the first link doesn't work due to archving) for an outright demand for evidence of consensus for the current version, with supporters' abject failure to provide any at all. That right there is enough to revert this to what it was before the major changes were pushed, or (as I did in my BRD move, just rewrite to be like that but better). — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 02:37, 19 September 2010 (UTC)

NB: I'm actually going to bow out of this for a while, since I've said my piece. I know GregL doesn't like long posts, and I'm not here to antagonize him. I also know he is prone to protracting arguments if he feels antagonized, and that I am more likely to argue for longer and with more heat if I get irritated by people who launch into to me because they feel antagonized (based on direct experience with GregL and some others here), so I don't see the point in any more back-'n'-forth right now. I've made all the points that I feel I need make, and have backed them up with clear and sound rationales. Meanwhile, I've gotten absolutely jack in response that substantively addresses the issues that I raised in my "too bold" edits. I'll let others who also care about this dispute take up the torch for a while. Drop me a {{wb}} at my talk page if my attention is needed. Closing (for now) summary: It's plain as day that there is no consensus for the current wording of this section, and there never was, otherwise it would be naturally impossible for there to be long, bitter debate about it, a debate that was self-defeatingly pointed out by supporters of the disputed language. The way forward is to remove pre- and proscriptions that do not have consensus, clarify the parts that do, and either just have the document STFU at that point, or (as I did) spell out what actual practice is among the majority of experienced editors (acknowledging that some experienced editors like Art take a more link-restrictive viewpoint). — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 04:20, 19 September 2010 (UTC)

I'd just like to add my name here as someone who is relieved that the overlinking of the past is finally being reduced, and I'd really hate to see any attempt to increase it again. As I see it, if someone wants to read about France they'll type that word into the search box. In addition to that point, the more links an article as, the less noticeable each of them becomes, so for the pro-linkers too, less is surely more. SlimVirgin ^{talk|contribs} 14:31, 19 September 2010 (UTC)

Oh, my goodness, what a lot of words! From what I can make out, somebody wants to simplify the MOS. So, why don't you post your suggested MOS change in your own sandbox, with a link here — then people who agree with you can have a look at it to get it ready for posting as a change to the MOS? In short, I have no idea what is proposed, and I don't intend to wade through all the above to find out. Sincerely, your very good friend, GeorgeLouis (talk) 21:46, 19 September 2010 (UTC)

• WP:TLDR. This section indicates to me that there may be an innate inability of the proponent to cut to the chase. I see no substantive arguments from him to justify calling for linking of words such as English, English language, United States of America, US dollar, when these are clearly words that >99% of people reading it know the meaning of, other than being what the proposer believes is the objective truth©™. This guideline has been arrived at by consensus through thorough discussion, and the onus is on any proponent of such a major change to prove that consensus has evolved to justify such a radical change in the wording. --Ohconfucius ^¡digame! 02:04, 20 September 2010 (UTC)
  • Absolutely! Links such as those mentioned almost never add value to the articles in which they are being placed. They simply dumb-down WP. I have trouble believing we are still discussing this.  HWV258.  04:09, 20 September 2010 (UTC)

{{ec)) If you are going to start off by saying you didn't read it, you probably shouldn't go out on a limb to say something isn't included. In fact, SMcLandish does make a reasoned point about why those terms should be linked. While apparently you only ever click on article links if you aren't familiar with the subject, not everyone browses that way. Someone reading a passage that mentions the US dollar may want to go to United States dollar to see what it includes about the topic. Someone reading "US$200" may want to click to see whether the article mentions a rough conversion between her currency and the dollar. That sort of use of Wikipedia is equally as valid as yours.

Your mentioning consensus also ignores what was said. The point is that while the current language may have been arrived through comprise among people at this talk page, it does not reflect the actual practice of a significant portion of, if not most, experienced editors. What exactly "consensus" means is much more woolly with Manual of Style and other guidelines than with articles. While consensus at an article rightly reflects the views of the editors of that article, consensus for a guideline should generally reflect actual editor behaviour rather than the views of the editors who happen to be interested in the guideline.

Finally, your point about there being a burden on the proponent is true, but you get what needs proving badly wrong. It is not his or her burden to establish consensus has changed; rather, the burden is to establish why it should change. Consensus then moves or doesn't. From the few replies that are actually responsive or show that the party replying has bothered to read what he is arguing against, it seems (though I may be wrong) editors think that at least this part of the Manual of Style should be prescriptivist rather than descriptivist. I don't see why. Actual practice among experienced editors at articles across the project is to link many of these terms. -Rrius (talk) 04:31, 20 September 2010 (UTC)

An article might mention the cost of a bridge built in Gdańsk in the 1970, and then give a US dollar conversion at that point in time. The USD is the currency of the largest economy of the English-speaking world, and also one of the 4(?) most important reserve currencies in the world. What's more, such a link would not be germane; even it's relevance is questionable because it is only a reference point. If someone wanted the current conversion rate, we would point tot he fact that WP is not the news. Anyway, most readers would know they will not find it here, and that they should to go to yahoo finance for today's exchange rate to the zloty. --Ohconfucius ^¡digame! 04:41, 20 September 2010 (UTC)

Ummm, no. Knowing about the articles you object to is almost as important as knowing the arguments you are countering. United States dollar#Exchange rates provides historical conversions to several currencies. -Rrius (talk) 04:50, 20 September 2010 (UTC)

I'm sorry for shaking you out of your apparent smugness, but you've entirely missed the point. A typical construction, in your case, goes like this: 'The Blah-blah Tower was built in 1904, at a cost of INR4 million (US$985,000)^[12]', where citation #12 is the source which says it was $985,000. There is no value to the reader of enticing him/her to visit the United States dollar article just to look up what another editor has put down in the table at United States dollar#Exchange rates – that information will be in the source. The INR (or PLZ or name your currency) isn't even in your table, so it would be entirely moot in your case. Mention of the US currency is just for reference purposes, and not at all germane nor relevant, so there is no earthly reason to link to USD. --Ohconfucius ^¡digame! 06:58, 20 September 2010 (UTC)

"Actual practice among experienced editors at articles across the project is to link many of these terms"—well, once upon a time, the actual practice at WP was to link dates and date fragments in regular articles; however that practice is now deader than the dodo.  HWV258.  04:48, 20 September 2010 (UTC)

Assuming people are stupid is not a great way to win arguments. As you know well, date linking was a part of date autoformatting, which was deprecated following a discussion with wide community involvement. That has never been the case here, and certainly wasn't for the existing language. I'm a persuadable editor, but it will take actual arguments rather than what we we've seen in support of the current language to this point. Just because I can restate SMcLansish's argument and defend it against weak arguments doesn't mean I agree with it. The fact is, I'm inclined to believe the best and most realistic language would be somewhere between the two passages in the summary section below. -Rrius (talk) 05:02, 20 September 2010 (UTC)

Oh, and tangentially, why did you type [[Dodo|dodo]] instead of [[dodo]] to produce dodo? I see this around from time to time, but I haven't figured out why. -Rrius (talk) 05:07, 20 September 2010 (UTC)

Nah, lots of editors liked underlining dates as a means of getting to a page of nebulous information (and one or two still try to flog that particular dead horse). That's why the linking aspect of the debate attracted separate RfC questions (here and here). The point being of course, that what was once established practice is no longer desirable (be in linking or auto-formatting).  HWV258.  05:24, 20 September 2010 (UTC)

I disagree. Underlining just years was done only by a small minority long before date autoformatting. I simply don't believe there was ever a point where most experienced editors thought linking dates was a good idea. Linking month-day dates was about autoformatting, and when it went out, there was, at best, a de minimus minority of editors who still thought linking them a good idea. Once again, I doubt you could show that there was ever a majority of editors who linked months and days for a reason other than autoformatting. Point being, of course, that, as I said, established practice was only changed after a very long discussion among a wide swathe of members of the community. -Rrius (talk) 05:34, 20 September 2010 (UTC)

You are entitled to your opinion, but those of us who went blow-for-blow through the entire miserable experience are also entitled to retain our memories of the events. I'll try one last time: at one point, being able to link/auto-format dates was widespread; now it isn't. Therefore the argument that "actual practice is..." is not always a good reason for maintaining that practice.  HWV258.  05:49, 20 September 2010 (UTC)

Except that that your example doesn't make sense. Linking of February 12 was never organically the actual practice of editors. At point A, there was a guideline saying dates should be linked for the benefit of logged-in users. At point B, there was a debate involving a large segment of the community, with notice given prominently to all editors. That led us to where we are now, point C, where the guideline now says not to autoformat. The situation is completely different from what I'm talking about, which is the actual behaviour of editors. Can it change? Of course. But the guideline here should follow any such change, not seek to bring it about. The question is then, what do people actually do? Do they link units of measure and currency, United States and United Kingdom? In my experience, most people do, and a few people sweep through and delink, only to have many of the terms relinked down the line. -Rrius (talk) 03:35, 21 September 2010 (UTC)

You have forgotten about the point between A and B when there was widespread delinking of dates (before the various RfCs). Remember that monkey-see; monkey-do often applies, and linking US and UK simply because they exist in an article is a brain-dead activity that appeals to the link-lunies. Delinking them because they don't deepen the understanding of an article is a thought-out and intelligent activity that strives toward one goal: to make WP easier-to-use for all readers. In terms of "...a few people sweep through and delink, only to have many of the terms relinked down the line", that's obviously a statistical guess; but I will say that it is necessary for a few bold editors to show the way (and that is their right at WP). That is exactly what happened with date linking and formatting—to the current advantage of WP. Enough users have now given cogent reasons for not changing the text of this MOS.  HWV258.  04:44, 21 September 2010 (UTC)

That's an excellent example of the arrogant attitude that causes so much friction around MoS issues. This "we're right, anyone who objects is just too stupid to get it" mentality needs to change. --Ckatz^chat_spy 09:50, 21 September 2010 (UTC)

Your post is unfair as it doesn't address the issues raised (and is a disappointing effort from an editor whom I generally respect). I didn't use the word (or imply) stupidity. You neglect to mention that the RfCs overwhelmingly supported the instinct of the editors who began the date-delinking work (the same editors who are now working to improve WP by removing links that do not deepen the understanding of articles). So let's be clear: it's not just "we're".  HWV258.  10:26, 21 September 2010 (UTC)

I agree insofar that there is a lot of arrogance and friction around MoS issues. That's all the more reason to quarantine the arguments here, not re-argue them at every article — either that, or don't let the MoS call itself a guideline at all. Or if we're going to try to make the rules match the consensus of experienced editors, we can still have plenty of arguing and arrogance over defining that criterion. Art LaPella (talk) 06:09, 22 September 2010 (UTC)

Re: "Dodo": to match the title of the article exactly, but to get a lower case "d" in the text.  HWV258.  05:24, 20 September 2010 (UTC)

The first letter of an article title is not case sensitive, so [[Dodo|dodo]] and [[dodo]] both point to the same place without redirect. -Rrius (talk) 05:34, 20 September 2010 (UTC)

Simply a precision born of too long a time spent in the computer industry. Thank you for your observation.  HWV258.  05:49, 20 September 2010 (UTC)

• Enough already. I’ve got a medical experiment starting in a few days and don’t need to be wasting my time on this tired old issue. We’ve had more than enough words here. This dispute will not be won by the editor who has the most time to waste and is most willing to pound his or her keyboard to death; it borders on tendentiousness. Our policies on linking are clear and are designed to ensure that the only words that are linked are those that will actually enhance a reader’s understanding of a particular article. This principle applies everywhere on a page, including infoboxes. Badgering the community to death and putting poor HWV258 in a position where he feels he is the only one standing sentry on this issue is poor form. There is no consensus to change our policies to avoid overlinking. Drop it please. Greg L (talk) 14:57, 21 September 2010 (UTC)

Summing up the linking dispute

The suggested MOS change is the last 4 edits by SMcCandlish, which were largely reverted by Tony1. The main controversy is whether to scale back this:

"Unless they are particularly relevant to the topic of the article, avoid linking terms whose meaning can be understood by most readers of the English Wikipedia, including plain English words, the names of major geographic features and locations, religions, languages, common professions, common units of measurement, ..."

down to this:

"Unless they are particularly important to the topic of the article, avoid linking plain English words whose meaning can be understood by most readers of the English Wikipedia (e.g. dog, breakfast, river, elbow). While the names of major geographic features and locations, historical figures, religions, languages, and other proper names, as well as currencies and units of measurement, are often but not always linked at first occurrence in an article's prose or infobox, they should not be repeatedly linked." Art LaPella (talk) 22:59, 19 September 2010 (UTC)

Thank you, Art LaPella. Is the first version you cited now to be found in WP:MOS? I couldn't find it. Or is this simply a suggested change? Your friend, GeorgeLouis (talk) 23:50, 19 September 2010 (UTC)

WP:MOS has subpages such as WP:Manual of Style (linking), and you are now reading its talk page. The text I quoted is at WP:Manual of Style (linking)#What generally should not be linked. Art LaPella (talk) 00:20, 20 September 2010 (UTC)

I would not like to see the proposed change implemented as it would be a diluting mish-mash. It is important that all links exist only to deepen the understanding of the topic of the article in which they have been placed. (I find it ironic that "dog" and "breakfast" have been placed in such close proximity in the example.)  HWV258.  04:20, 20 September 2010 (UTC)

I think he may have intended also to juxtapose 'river' and 'blue' ;-) --Ohconfucius ^¡digame! 04:31, 20 September 2010 (UTC)

SMcCandish's proposed change would eviscerate our clear, prudent guidance. It is entirely unacceptable.—DCGeist (talk) 04:40, 20 September 2010 (UTC)

I am fine with the current wording of the guideline. I am not fine with people who believe that unless they are particularly relevant to the topic of the article, avoid linking means 'never, ever link anywhere'. A. di M. (talk) 09:10, 20 September 2010 (UTC)

That is indeed a major issue; the wording is being used by a few individuals to justify stripping out links en masse, often with no apparent regard for appropriateness. If more discretion were to be demonstrated, especially with regard to geographical links, we might not have this problem. --Ckatz^chat_spy 09:20, 20 September 2010 (UTC)

Perhaps you were thinking of this? If you meant my own unlinking, then naturally I think I use "discretion" consistent with the "particularly relevant to the topic" clause. Art LaPella (talk) 20:29, 20 September 2010 (UTC)

• The current wording is fine. The community has had a belly-full of this issue and some of the above posts drone on like a filibuster from a Southern senator opposing a civil rights bill.

  The proposal is a bad idea because it outright invites linking totally extraneous garbage, such as if there was an article on “Scientific goofs”, this sentence: American scientists in the 1960s thought they had invented polywater. …would have three needless things linked when all it needs is one link: polywater.

  There is always a losing side to these contentious issues. However, “contentious” does not equate to “chronically reoccurs like genital herpes.” The current wording is simple and couldn’t be clearer. The tip-off that it is a thoughtful guideline is it begins with these, oh-so-logical words: Unless they are particularly relevant to the topic of the article… and then isn’t immediately followed up with caveats designed to undermine that very principle.

  There is no stomach for revisiting the issue, let alone actually changing the current guideline. Greg L (talk) 23:51, 21 September 2010 (UTC)

Ha. Deja vu. An editor comes along to query the wording of the linking guidelines, and/or the very rigid interpretation of the "unless relevant .." qualification - while making very clear their absolute agreement with the removal of redundant, repetitive and trivial links - and gets shot down by the same four-five person clique who by virtue of their presence on this page believe that they represent the consensus across Wikipedia, told that they're on the "losing side" and should head off to the dustbin of history; that they don't understand what linking is for, and that they're duping and confusing other readers about what other pages they should be able to link to easily (as if there is a definitive right and wrong about those last two, uniform for every single reader); that they want "seas of blue" etc. You know, when this happens this regularly, both here and on the talk pages of those removing vast numbers of links from pages, you might stop to consider that there's a problem here, and that your consensus isn't quite as strong outside of this bubble. People are just asking for flexibility, and an understanding that just because most people (not all, note) know pretty much what and where for example Italy is, or what a lawyer is, we don't need to remove every single link to such pages from related ones. Those pages, after all, say much, much more than simply what and/or where the thing is. Even Italians and lawyers might be grateful for a link on occasion. And, of course, people don't have to click on it if they don't want to. Choice, it's an amazing thing. People are capable of exercising it when confronted with options. And if someone could give a clear and simple answer to the question posed here, that might help. N-HH talk/edits 17:06, 23 September 2010 (UTC)

The "four-five person clique", though undercounted, is something to think about. Last year I think there were 3. But 2 or 3 people telling us what the consensus is, and that anyone outside that consensus doesn't count, present a similar problem. Either way you look at it, keeping the hostility level down will encourage more participation here and a better mandate. To answer the 2 questions: the current discussion is probably as good an indicator of consensus as we're going to get, and nobody says "regardless of context". Of course there is no standard list of common countries, but mine is no secret; search User:Art LaPella/AWB list for the word "Pacific". Art LaPella (talk) 02:01, 24 September 2010 (UTC)

I agree about hostility, but my experience when I entered this debate after seeing a link to "France" chopped out of an article on my watchlist, was a slew of people - names that are now very familiar - telling me I was wrong to even raise the issue, as well as totally and utterly misrepresenting - continually and repeatedly - my position. That just sets the tone from the outset. And, like I say, I noted above after popping by this page again, the same thing happening to someone else. Plus people posting links to pictures of crying babies. The other problem is that this is a pretty closed, backwater forum. Consensus here can in no way be taken as consensus across WP for something that is a little more fundamental than whether to use Oxford commas, or en or em dashes. Anyway, thanks for that link - but come on, it's utterly unclear what that is all about to the average reader or editor, myself included. Plus if everyone has different ones, where is the consistency? N-HH talk/edits 15:05, 24 September 2010 (UTC)

If your issue is Champagne (wine), I probably would neither add nor delete a link to France from the lede of that article. But most country links I find are less relevant than that one. So if you think Champagne needs a link to France, that is an argument for refining OVERLINK, not for making OVERLINK pretty much irrelevant. Crying babies: yeah, I almost posted an objection to that. Backwater forum: sure, but once again, what's the alternative? Appoint you as the judge of what the consensus is? That link: I didn't expect you to try to read the AWB regex code, just the list of countries and related words before and after "Pacific"; was that your problem? Sure everyone has different lists, or no explicit list at all if they don't use software, but why do you want consistency? On the one hand you think OVERLINK is overregulating from a backwater forum, and on the other hand you think it's underregulating because it doesn't list exactly what should be unlinked. If we don't really have authority for OVERLINK, then where would we get authority for agreeing on a specific list? Other Wikipedia rules such as WP:CIVIL don't list a specific list of forbidden swear words; the context matters too, not just the words. Art LaPella (talk) 20:38, 24 September 2010 (UTC)

• Current consensus status quo is fine and dandy. We have WP:IAR if there is an exceptional case we need to cover. Flexibility and choice are fine, but the point of having a manual of style is to have some degree of consistency. Too much flexibility and choice will defeat this goal. Keep the guideline as it is, please. --John (talk) 17:13, 23 September 2010 (UTC)

I'm actually broadly fine with the guideline as is too, as I would be equally be with the proposed change (which possibly I might slightly prefer). However, as noted above, people need to read the qualifications in them, eg where they say common terms should not be linked, unless relevant. Also where they state that links are about general navigability, as well as about simply directing people to where they "should" go or about what will "help understanding" in some strictly defined sense for some ideal, standardised reader. That's where these problems ultimately all stem from. In an article about an Italian lawyer, links to both Italy and lawyer on one occasion each do not seem to be overlinking, by common sense, standard editing practice or an accurate reading of the current guidelines. But a small number of editors regularly remove them, even from infoboxes, often using automated tools, citing "overlink". That just seems unnecessary, and not especially helpful in any obvious way. N-HH talk/edits 18:55, 23 September 2010 (UTC)

• Choice is all well and good, but when you go to a concert, you'd expect there to be a certain amount of quiet for you to enjoy the music. Extraneous noise detracts from the listening experience. You are of course free to plug into your iPod, but please switch off yer bleedin' phone. As to 'lawyer' above, I'd say it was a word which should almost certainly be unlinked in a biography - unless the subject was a specialist and there was a more specific branch of applicable law, such as land law or employment law that can be linked to. Linking to 'law' or 'lawyer' brings 'sweet FA' to the party. Ohconfucius ^¡digame! 02:38, 24 September 2010 (UTC)

I'm sorry, but I don't understand what the music point has to do with anything. It's a very odd analogy. And a link to lawyer brings a link to lawyer for any reader who might happen on any occasion to want to use it and read in some detail about the profession that this person was a member of. That page says a lot more - I assume - than "lawyer: a person who does law". The more important question is what having a link to a significant term takes away, when it's clearly relevant to the topic of that page. How actually does it detract or distract? Does the word being a bit blue magnetically pull people towards clicking on it, when they might not otherwise? You personally might not want it there, but with the greatest respect, so what? There are millions of other readers and editors. N-HH talk/edits 15:05, 24 September 2010 (UTC)

I don't understand why notions of "hostility" and "friction" have arisen. Another way of looking at it is: healthy debate. I have tried to respond to issues, and I believe most other editors have done the same. No harm; no foul.
Here's an example of what a healthy application of the guideline produces. Four months ago, with that edit, I removed links to six common terms and a country name—links that did nothing to deepen the understanding of the topic. There have been no reversions of my edit (indicating community support), and the article scans better without the links. It's not just the delinking of words such as "lipstick" and "piano" in that article for which the hard-working editors involved in this issue strive, it is also a change of culture that will hopefully deter editors from wanting to make such nebulous links (or at least to have editors think carefully about the relevance of links, as opposed to the scatter-gun approach to linking).
 HWV258.  07:05, 24 September 2010 (UTC)

I agree with that edit in its entirety (I would say though, that lack of reversions is as likely to indicate indifference as much as "support", when it comes to other people's views). As I agree with the idea that links should not be scattered thoughtlessly. As I agree with the idea that people should think before adding them. But I also agree that people should think before removing a link. Which is, of course, something the guidelines explicitly require, when they ask people to bear in mind that a link they might want to remove "may be useful to other readers." There is a such a thing as a good idea pushed too far on occasion. N-HH talk/edits 15:05, 24 September 2010 (UTC)

Amen to that. The delinking jihad has continued for too long. Time for a bit of common sense, and having terms linked more than once per article or section would be a good start. --Michael C. Price ^talk 23:20, 24 September 2010 (UTC)

"...delinking jihad...". Just when you think some progress is being made, along comes the usual childish pejorative nonsense. Sigh.  HWV258.  00:09, 25 September 2010 (UTC)

Cf your own language, HWV258: "link-lunies". --Michael C. Price ^talk 07:49, 25 September 2010 (UTC)

Ah, but Michael, my comment wasn't directed at you (you're not one are you?). But let me humour your point for a moment: are you saying that you wait for someone to make a comment that you don't think is appropriate to justify returning an inappropriate comment?  HWV258.  08:02, 25 September 2010 (UTC)

Ah, but HWV258, my comment wasn't directed at you (you're not part of the jihad, are you?). --Michael C. Price ^talk 10:25, 25 September 2010 (UTC)

You knew very well my inclination towards delinking when you posted, so you are being disingenuous. For the record, "Jihad" refers to a religious duty by Muslims and is therefore a poorly selected word in this context. My other point remains unanswered.  HWV258.  10:36, 25 September 2010 (UTC)

"...delinking jihad...". I would like to believe that he is merely trying to wind you up, so I would react accordingly. --Ohconfucius ^¡digame! 08:10, 25 September 2010 (UTC)

"childish pejorative nonsense ..." Sigh. Art LaPella (talk) 01:42, 25 September 2010 (UTC)

• Isn't it? I think the adjectives were more or less spot on, but I would have just used the noun 'rhetoric'. --Ohconfucius ^¡digame! 01:50, 25 September 2010 (UTC)

mein gott! "having terms linked more than once per article or section would be a good start" – need that like a proverbial hole in the head. I think only a very small number of the more vocal linking advocates, if any, would support that view. --Ohconfucius ^¡digame! 01:50, 25 September 2010 (UTC)

Articles and sections can be very long. The context of words and phrases changes; sometimes a term, in context here elicits inquiry, whereas there will not and for another reader it will be reversed, hence repeated linking is sometimes advisable for ease of use. Why is that deemed such a radical opinion? I've been following this linking fiasco for years and never seen a rational counter-argument that held water. --Michael C. Price ^talk 07:49, 25 September 2010 (UTC)

Gentlemen, we agreed a few months ago that on this page, the temperature would be kept cool, with no emotive or personalised language. HWV, "childish" is unhelpful here; and Michael, labelling editorial work as a "jihad" is just as unhelpful. Could we please engage on the issue rather than commenting on editors and behaviour? Tony (talk) 08:14, 25 September 2010 (UTC)

You are quite correct. "Childish" wasn't anywhere near the appropriate adjective in this case. I will try harder in future.  HWV258.  08:26, 25 September 2010 (UTC)

Tony has asked me a question, which I have responded to which is relevant to this discussion. Must dash. --Michael C. Price ^talk 10:25, 25 September 2010 (UTC)

Query regarding wikilinking at Village Pump Misc.

I have posted a query at the village pump, on the question of what is the best (long-term) way to handle redundancies among related articles. Since this pages discusses wikilinking, I thought some of you might have something to say. Cheers! AGradman /

/ talk 03:40, 23 September 2010 (UTC)

Acronyms

Should acronyms in parentheses be included in links or not? For example, National Football League (NFL) or National Football League (NFL)? –CWenger (talk) 16:22, 23 September 2010 (UTC)

I'd go with the latter one per the KISS principle. The former uses more complicated markup with no discernible advantage. A. di M. (talk) 13:06, 24 September 2010 (UTC)

Looks neater in isolation like this, but in a sentence already crowded with high-value links, one might be inclined not to extend the blue in this way. In fact, if the very next item is a link, there's an advantage (as MOSLINK points out) in have a bit of black in between. So I'd say it depends on context: editorial judgement is called for. Tony (talk) 14:31, 24 September 2010 (UTC)

Link "utility"

Copied from my talk page:--Michael C. Price ^talk 21:12, 25 September 2010 (UTC)

Hi Michael, I'm posting here to inquire in greater depth into your views on the density of internal links in WP article text. Am I right in thinking that the differences in view might have something to do with the extent to which an individual (i) finds that every additional link signal is likely to dilute the effects of others in its vicinity, and (ii) regards densely linked text as more likely to suppress the likelihood that readers will click than to stimulate it?

I suppose it is with a balance-sheet mentality that I approach the decision as to whether to link or not to link, not unlike my attitude to the use of optional commas in prose. I'd be interested to discuss your perceptions of linking, visually and in terms of utility and "dilution". Perhaps I am overstating these issues? I've temporarily watchlisted your talk page. Tony (talk) 08:40, 25 September 2010 (UTC)

Time prohibits a full answer, but I basically think your commercial metrics are inappropriate for wikipedia. Commercially it is the absolute rate of click-thrus (=money) that are significant. Naturally such rates decrease as we progress through an article because most people stop reading at some point before the end. Link density at WP should be governed by the utility of a link = absolute rate / probability of reading the linked term - i.e. a conditional, not absolute rate, which will necessarily be higher than your commercial metrics indicate. --Michael C. Price ^talk 10:16, 25 September 2010 (UTC)

I'm not sure I really understand this. The commercial analogy eludes me. Do you mean the "absolute" rate of click-thrus for a single link in an article, or for the sum total of links in that article? You've introduced another issue, which is the probable fall-off in reading as you go through the article. What are the implications of this? When you say "link density at WP should be governed by the utility of a link, I mostly agree, although there may be a few other factors that are relevant. I'm trying to get my head around "the utility of a link = absolute rate / probability of reading the linked term".

Are you familiar with findings on signal-to-noise ratio (in the psychology of perception) and the amount of choice available (in supermarkets, I think)? Tony (talk) 15:05, 25 September 2010 (UTC)

You often mentioned the metrics had a commercial basis or validation ; if not, the point doesn't matter. The important point still stands, which is

link density ~ conditional click-thru rate = absolute click-thru rate / probability of reading the linked term

The denominator has hitherto been ignored. The other issues are red herrings.

--Michael C. Price ^talk 15:38, 25 September 2010 (UTC)

Correction: it was Ohconfucius, not Tony, who said "the commercial world is well aware of potential problems caused by excessive linking." But the point I'm making still stands. --Michael C. Price ^talk 08:26, 26 September 2010 (UTC)

Surely the utility of a link, if anything, should be the product of absolute rate and the probability of reading the linked term and not divided by the latter; the probability of reading the linked term is clearly related to how common-garden the linked term is, irrespective of how relevant. --Ohconfucius ^¡digame! 15:42, 25 September 2010 (UTC)

Not the utility, but the link density. A link at the end of an article may have little overall utility since few readers get there, but to those that do it may as useful as a more-read link at the beginning. --Michael C. Price ^talk 16:00, 25 September 2010 (UTC)

I don't think a non-germane link would have any greater 'utility' if it is the first word linked in an article or the last. Yes, a word placed and linked earlier on would be more likely to be clicked on than if that word appeared and was linked later. This has more to do with being first in line (or forst in the consumer's mind) – a concept better understood by marketeers. Any link which appears a second time would have the marginal utility which approximates zero, much like for a second copy of any given newspaper. You seem to be conflating mere propensity to click on a link to actual usefulness, as would be measured by 'utility'. If you are suggesting that we should have higher linking density in the earlier sections and lighten up towards the end, our current linking propensity already achieves that. --Ohconfucius ^¡digame! 04:17, 26 September 2010 (UTC)

Taking your last point first we should have higher linking density in the earlier sections and lighten up towards the end, I am suggesting the reverse - namely that link density should be more evenly distributed across an article - so there is obviously some major misunderstanding going on here. The metrics that Tony always cites skew the density towards the beginning of an article, because they don't take account of the fact (which no one disputes) that the beginning of an article is read more than the end. My point is that this is no reason to weight the link density towards the beginning of the article.

Any link which appears a second time would have the marginal utility which approximates zero ... rubbish. But I believe this has been rebutted often enough already.

You seem to be conflating mere propensity to click on a link to actual usefulness. No, my point is that this is the mistake that Tony's metric argument makes. I don't know how to make this clearer. Near the end of an article you are adrift in a monochrome sea - how can this be useful?

--Michael C. Price ^talk 06:56, 26 September 2010 (UTC)

You are now suggesting that we somehow weight links towards the end of the article – is this to encourage readers to get to the end? I have often found it difficult, with the 'link only the first occurrence' rule, to lighten the load on the lede. For me, it sort of defines the 'natural link density'. Also, as you have been putting great importance into your notion of 'utility', it seems counter-intuitive that your suggestion would actually reduce the utility of links in the later sections of the article, so what's the point? Your notion of utility appears to be from the viewpoint of commercial sponsors, who would naturally like their links to be put earlier in any given article. My notion of utility is from the reader's viewpoint. It places importance on whether the link adds value to the reader, and its frequency/scarcity, and ignores where in the article any given link is. --Ohconfucius ^¡digame! 10:03, 26 September 2010 (UTC)

You seem to have misunderstood everything. I'll try again later. But I agree with utility is from the reader's viewpoint. It places importance on whether the link adds value to the reader, and its frequency/scarcity, and ignores where in the article any given link is.--Michael C. Price ^talk 10:10, 26 September 2010 (UTC)

PS. And who might be a suitable/relevant authority on this utility of which you speak I can read up on? --Ohconfucius ^¡digame! 15:45, 25 September 2010 (UTC)

We are the authority. --Michael C. Price ^talk 16:00, 25 September 2010 (UTC)

• Michael, what bothers me is the (admittedly anecdotal) evidence that readers rarely click on links. User:Piano non troppo, with something like 50,000 edits here, is an ex-webmaster, and has interesting things to say about his professional experience of this matter. I can dig up some diffs to what he has said, if you like. (But off to bed now.) Tony (talk) 16:47, 25 September 2010 (UTC)
  • I simply don't believe that, based on my own usage - and it is irrelevant to my point about maintaining constant link density through out an article. But anyway, I was responding to your claims, not someone else's. Time to return to the appropriate venue. Any objections to me moving this all there?Done --Michael C. Price ^talk 16:52, 25 September 2010 (UTC)

Can we keep the discussion more concrete? Not having a degree in economics, I had to read it three times before getting everybody's points. :-) Anyway, we're not selling anything, so whether the readers just prints down a copy of the page and reads it without giving a damn about the links, or they open each and every linked article in a new browser tab until the browser crashes, that's their freakin' business. A. di M. (talk) 16:22, 26 September 2010 (UTC)

Okay, let's start again. The term utility is misleading here, so let's drop it. The delinking argument says, in part, links later in articles are clicked on less often and, therefore, they should be removed. My point is that, they may be clicked on less often, but that doesn't mean we should remove them because.... ...it is not the absolute rate of a link's selection that is relevant, but the selection rate / reading rate. Obviously if only one reader in a million reads as far as a link at the end of an article, but then always clicks on the link, then that link should be retained - even though it may be the least used link in the article. All the talk of metrics I've seen around here (including User:Piano non troppo's contribution) seems to talk about absolute click rates, which is simply not the relevant metric. --Michael C. Price ^talk 20:51, 26 September 2010 (UTC)

"The delinking argument says, in part, links later in articles are clicked on less often and, therefore, they should be removed." Sorry, whose argument is that? I'm not sure I agree with the notion that links are less likely to be clicked on the later they occur in an article. I suspect there are a number of factors at play: one of them is likely to be that readers are less willing to divert at the top of an article, just as they're getting into the topic (unless there's a technical term they really need to know more about, and which isn't briefly glossed in the text—I actually believe technical terms should be glossed briefly, unless it can't be done without clutter/awkwardness). Tony (talk) 09:40, 2 October 2010 (UTC)

Recognize this quote? "At any rate, some substantial number of readers do not scroll past the first screen, and most do not scroll to the bottom. This suggests: Links placed toward the bottom of an article are far less likely to be seen than ones at the top, and even more unlikely to be used." Note the non-sequitur at the end. --Michael C. Price ^talk 14:53, 2 October 2010 (UTC)

Are you aware that with several popular browser one can open a link in a new browser tab while continuing to read the page they are reading? A. di M. (talk) 11:35, 2 October 2010 (UTC)

Bot archiving

Can anyone tell me why is the bot not archiving? It seems it is set to archive at 20 days, but there were threads dated as old as 20 July which were still unarchived. --Ohconfucius ^¡digame! 06:35, 26 September 2010 (UTC)

• About this: I object to this undoing of my archiving. This page is now in excess of 200k. The threads I archived are as good as dead. Ys, perhaps some questions have remained unanswered, but that is the nature of threads... Filibustering isn't going to breathe life back into an expired equine. I vote we redid the archiving. --Ohconfucius ^¡digame! 10:21, 26 September 2010 (UTC)

I vote we keep it present. 20 days is far too short a period for archiving. And the content is pertinent. --Michael C. Price ^talk 16:03, 26 September 2010 (UTC)

I've changed it to 30 days. If no one has commented on a topic in a month, it should be archived. -Rrius (talk) 19:14, 28 September 2010 (UTC)

Link "dilution"

We need to examine another term we see bandied around, link "dilution". Assuming that it means that the occurence of neighbouring links reduces the incidence of clicking on an individual link, why do the delinkers think this means links should be removed? To use an analogy, the occurence of many flavours of ice-cream lowers the purchase of any individual flavour, but that is not a justification for only selling vanilla. Consumer choice is paramount here. Why does not the same principle apply to links? --Michael C. Price ^talk 06:59, 27 September 2010 (UTC)

Presumably because there are thousands of flavors, not just 31. If there are too many flavors like "Crawling Caterpillar Delight", you might not even notice the chocolate. Art LaPella (talk) 18:23, 27 September 2010 (UTC)

You'll notice it if you like it. Same with links. Choice is good, because it caters for variation in interests and background. --Michael C. Price ^talk 19:14, 27 September 2010 (UTC)

Everyone else is against a "sea of blue", but if you would always notice a link you want, then a sea of blue would be ideal. But I don't think you would notice. Thus linking should only be done when its value exceeds its dilution effect. Art LaPella (talk) 22:18, 27 September 2010 (UTC)

We are not talking about a "sea of blue". Thanks for wasting our time.

My question was serious: why is link dilution bad? Isn't it just indicative of readers making disparate choices?

--Michael C. Price ^talk 02:16, 28 September 2010 (UTC)

"You'll notice it if you like it? You don't go to an article with the aim of recognising your favourite link.

"Consumer choice is paramount here"? Which consumer? The ones who think it's worthwhile clicking United States, Japan, Marisa Tomei, drama, celebrities, or profit—which are all found (in increasing order of uselessness) in the current professional wrestling article. If we don't apply high standards to linking, how low do we go? No doubt there are readers (people with a foreign background, youngsters, poor readers, etc) who might not know the meaning of form, display, and strong (all of which are found in the first sentence in that article), so by the logic of "Consumer choice is paramount here", they would have to be linked. Well, I (and many others) say "no"—there's a search box for nebulous words and phrases.

Linking is now, and always has been, a question of limitation—to the ones that deepen the understanding of the article.

 HWV258.  04:18, 28 September 2010 (UTC)

The "form, display, and strong" point is silly, since we already know Wikipedia is not a dictionary. The "how low do we go" point is just sea of blue (=contiguous links) recycled. More time wasting. As for "which consumer", offering choice caters for most or all readers. Same answer for deepening the understanding of the article. Deepening is relative to a reader's current understanding, which varies from reader to reader, and from time to time. The one-size-fits-all approach that the delinkers take just doesn't work - it fails to recognize the diversity of readers. --Michael C. Price ^talk 05:26, 28 September 2010 (UTC)

Of course it's a silly point. It's a blatant straw man, a logical fallacy the deletionists in this discussion have been using as a crutch the entire time this "debate" has been ongoing (it's not really a debate, which requires logically sound arguments from both sides; it's just a bunch of chestbeating and pissing). This "form, display and strong" theme is a rationally bankrupt argument. No one has actually proposed dicdef linking like "strong", yet the deletionist argument, boiled down, is "we have to be more and more anti-linking otherwise everyone's going to link stuff like 'strong'". Well, the non-deletionists here agree that such linking is stupid. What they don't agree on is the slippery slope argument (another fallacy) that because noobs tend to make stupid links like this that more and more restrictive anti-linking language has to be adopted or the sky will fall down. The sky isn't falling down, and never will. Experienced editors undo the mistakes of noob editors, today just like last year just like in 2005, and life goes on. I'm amazed at the level of sheer frakking hysteria on this page. Everyone needs to calm down and go do something else for a while, like go outside or read a book, and quit obsessing over nitpicks in a guideline no one pays attention to anyway. People link, refrain from linking, and unlink based on common sense and experience, pure and simple. I doubt that even 1% of editors with over 5,000 edits have ever read this page. I did some WP:BRD editing on it to get the point across how screwed up this page is, and I hope that will conttribute in some way to this document eventually making sense, but my breath will not be held. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 19:18, 28 September 2010 (UTC)

The time wasting occurs because you haven't distinguished consumer choice from the ultimate choice offered by the sea of blue. Whatever semantics we use, some links are more useful than others, and the least useful links should be delinked. When we get past that point, we can discuss why linking to China is seldom very useful. Art LaPella (talk) 05:42, 28 September 2010 (UTC)

"Least useful" to who? We need to be careful about bandying such terms about, since usefulness is a relative concept. Do you understand that? --Michael C. Price ^talk 08:05, 28 September 2010 (UTC)

Least useful to whom?? why, the man on the Clapham omnibus, of course. ;-) --Ohconfucius ^¡digame! 00:38, 2 October 2010 (UTC)

But you'll notice that those "dictionary" words are not red-links, so it is possible for editors to spiral into the depths of linkitis (and many do). What's worse is that there are always the tinkerers—who believe that they can squeeze one more link out of the tube (well, an article isn't really finished while there's something else to link, is it?). On the other hand, the current policy of only linking terms that deepen the understanding is elegantly simple.

I've had a good go now, so if you believe that the arguments presented are "time wasting", then we'll have to leave it at: enough editors believe the current wording to be effective, so I guess it stays.

 HWV258.  06:20, 28 September 2010 (UTC)

Interesting. I'm asking why link dilution is bad, and not getting any answers. --Michael C. Price ^talk 07:52, 28 September 2010 (UTC)

I have no bloody idea 'bout who the hell Marisa Tomei is, so I wouldn't find a link to it useless. The only reason why I find that link unnecessary in the current professional wrestling article (but unnecessary ≠ harmful) is that I guess that The Wrestler will link to there anyway. IOW, never make too strong assumptions about the readers' knowledge: whereas we can assume that everyone knows (or knows where to look up) all these words and has heard about these languages, most of these cities, and the first few dozens of these countries, any further assumption is going to be wrong for a non-negligible fraction of readers. A. di M. (talk) 09:14, 28 September 2010 (UTC)

My incredulity about this debate has reached new heights with the thought that there is someone who has "no bloody idea 'bout who the hell Marisa Tomei is". :-)  HWV258.  09:49, 28 September 2010 (UTC)

Well, my mother didn't know what Guinness is until I told her a couple of years ago. A. di M. (talk) 15:56, 28 September 2010 (UTC)

Not everyone reads People magazine or watches E!, so be as incredulous as you like. Not everyone has eidetic memory for celebcruft. I'm even a movie buff (I have over 500 DVDs and Blu-rays, and probably at least as many films/shows on some 1.5TB hard drives reserved for media rips) and I cannot right now name a single Tomei film from memory, nor can I form a clear picture of her in my mind. I remember her being pretty and brunette and that's it. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 19:06, 28 September 2010 (UTC)

It's interesting that a sense of humour is the first casualty in this. I was joking (didn't the smiley help you to detect that?). The point being that Marisa (lovely, and apparently anonymous, as she is) does nothing to deepen the understanding of an article on wrestling.  HWV258.  22:48, 28 September 2010 (UTC)

And yet someone reading about wrestling might want to read about her. As per MOS, hence the link. --Michael C. Price ^talk 05:38, 29 September 2010 (UTC)

And someone reading that article might want to read about "profit"? How low do you go? Do you link "matches" in the second sentence just in case? She does not deepen the understanding of the wrestling article sufficiently to warrant a link. There's also the issue of chain-links (e.g. she would be found if someone follows The Wrestler movie link). Lastly, anyone who is really interested in her, can copy and paste her name into the search box. These are concepts well understood by most of the editors here, so we all need to move on with more concrete aspects of the debate.  HWV258.  06:11, 29 September 2010 (UTC)

The old "use the search box" line. So let's remove all the links, eh? Why have have links at all? --Michael C. Price ^talk 06:55, 29 September 2010 (UTC)

• I know it's actually a rhetorical question, the answer to which you know full well. Rather disingenuous, from a physicist. You have been trying to blind us all with science about utility, you then failed to explain the utility curve for any useful good or commodity starts at a non-zero value; as x tends to infinity, y tends to zero – only for links, x does not need to tend to infinity, and I suspect x only needs to reach two before utility starts tending to zero in the vast majority of cases. --Ohconfucius ^¡digame! 14:39, 1 October 2010 (UTC)

• Strange, I thought it was the delinkers who were trying to blind us with science, with all their talk of metrics, webmaster experience and degrees in psychology. Yet, upon examination, it is their arguments that are exposed as pitefully lacking. If my position is so obviously crap and bad science then it should be easy to rebut my last response in the Link "utility" section. Still waiting for any response there.--Michael C. Price ^talk 19:11, 1 October 2010 (UTC)

• Quoting you: Strange, I thought it was the delinkers who were trying to blind us with science, with all their talk of metrics… Saaay. Were you going to follow up on that one with “When the universe was made by Him 7,000 years ago”? Blinding people with science? (I mean, WTF???) Are you trying to win arguments here? Stay on point. You seem to want to link the living crap out of articles and that’s considered by the community to be three daddy steps backwards. Give it up; come back when your position is somewhere remotely close to “middle ground.” If you keep badgering us on this, I’m might make a volcano rumble. Greg L (talk) 01:07, 2 October 2010 (UTC)

• In this case, I saw someone who I thought was a scientist opposite me. I'm not thick, but were just not quite on the same wavelength – and I'm sure there are plenty others in a similar situation. You gave up trying to explain to me above, remember?? My repositioned argument is merely a response to you, as you appear to have started to make rather POINTY and sarcastic statements of positions you do not actually support or advocate – statements which are not conducive to a healthy debate. Please indulge me... how are my arguments "pitefully (sic) lacking"--Ohconfucius ^¡digame! 23:08, 1 October 2010 (UTC)

• The question has been answered, by ALP, amongst others, to the extent that to do any more would be considered ad nauseum by the small number here who are not yet suffering from the WP:TLDR syndrome. Or perhaps you just don't like or accept the response(s)? --Ohconfucius ^¡digame! 00:45, 2 October 2010 (UTC)

• MCP, did you come to the conclusion somewhere in your experience that not dropping an issue and setting your egg timer to remind you to pound away on your keyboard some more on this issue or that somehow pulls out a win? The most tendentious editor gets his way?? A lot of middle-of-the-roaders are just tuning this one out and letting you rant. I’m not sure there is a consensus to even discuss this matter any further, much less any stomach for making significant changes. Why don’t you just drop it? Greg L (talk) 23:53, 1 October 2010 (UTC)

• The whole approach adopted by Cavendish appears to be to discredit the guideline as one that nobody bothers to read, let alone follow. The assertion that the guideline somehow came into being through some equivalent of immaculate conception in a atheist world is making me laugh and cry at the same time. The tactic adopted by MCP seems to be to filibuster about his point(s) not being addressed, even when they patently have been, albeit not to his satisfaction. It seems his 'non-acceptance' more like denial. --Ohconfucius ^¡digame! 01:58, 2 October 2010 (UTC)

What a lot of sound and fury. And no response at the Link "utility" section. Still waiting for any response there....--Michael C. Price ^talk 06:00, 2 October 2010 (UTC)

In this particular case, it's obvious from the context that Marisa Tomei is an actress who performed in The Wrestler; but in general, if you use a proper name and the reader doesn't know it they won't get whatever point there was in using it, so they will want to know whom or what that proper name refers to.[2] A. di M. (talk) 18:17, 29 September 2010 (UTC)

Linking only the end of a chain is elegant (in this case the Tomei link is found via the movie link).  HWV258.  00:11, 2 October 2010 (UTC)

I agree that too many links make it harder to notice each individual link (see this), but: 1) the threshold before this happens is much higher than some people here claim (I think it's about one link (excluding superscripted links to footnotes) per hundred visible characters, paragraph-wise); 2) the solution is not going to Special:WhatLinksHere/China and removing 'em all; rather, it's looking for the paragraphs which have the greatest number of links per hundred visible characters and thinking about which links are the least useful in that context. A. di M. (talk) 11:11, 28 September 2010 (UTC)

With respect A. di M., the solution is to keep the links that deepen the understanding of the article in question. I do agree with you that finding and removing all WhatLinksHere is not correct, but don't agree that some sort of localized density is the criteria for inclusion or removal. Sometimes a high density of links may help to deepen the understanding of the article, and other times it might be fine to remove a single link in a long paragraph—if it is nebulous to the topic.  HWV258.  11:40, 28 September 2010 (UTC)

Actually the criterion is not just to "deepen understanding" but also link to "articles the reader might be interested in". The MOS says

Ask yourself, "How likely is it that the reader will also want to read that other article?"

--Michael C. Price ^talk 11:46, 28 September 2010 (UTC)

Absolutely. The guidelines do not reduce the purpose of linking to that single concept, which is then to be interpreted in the most limited way possible, so as to exclude *every* link to so-called "common terms" or things where "everyone knows what they are", however we would attempt to define those anyway. And they do not in fact mention the words "dilute" or "dilution" at all. They do however say, as common sense suggests they should -

• Internal links bind the project together into an interconnected whole
• When writing or editing an article, it is important to consider not only what to put in the article, but what links to include to help the reader find related information
• Provide links that aid navigation and understanding
• Think before removing a link—it may be useful to other readers

Excluding all links to some of the site's most commonly visited pages, even when they appear on pages for clearly related and directly relevant topics is not justified. Yet it happens. Nor is it clear that it helps otherwise inept readers focus on the "better" links - however, again, we would try to define what is better for everyone as a whole - and prevents them being distracted in some way that prevents them exercising their own choices. N-HH talk/edits 13:08, 28 September 2010 (UTC)

And it is just unbelievable chutzpah to assume that we, the superhumanly intelligent magically empowered editors of this largely ignored guideline of supreme universal importance are in any position to decide what is a "better" link for any given reader. Some readers want extremely narrow links, and others want broader ones. We already lean toward being narrow, and that's good enough. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 19:00, 28 September 2010 (UTC)

Pray tell: have you ever thought about just ignoring it?? --superhumanly intelligent magically empowered editor of this largely ignored guideline of supreme universal importance ^¡digame! 23:34, 1 October 2010 (UTC)

While link density does indeed vary a lot according to the type of text, it's still a quite useful indicator. If you have a very high link density (say, over 3 links per hundred characters) but you find that all the links are vital and none of them has already been linked before, odds are that you're writing too densely and you'd better revise WP:NOT#Wikipedia is not a manual, guidebook, textbook, or scientific journal, especially points 5 and 7. Conversely, "it dilutes other links in the vicinity" is not a valid rationale for removing a link if there are no other links in the vicinity (though there might be different reasons to remove it, e.g. "it redirects back on this same page, as it should" or "it is so irrelevant to the topic of this article that most readers following it would mutter 'WTF?' and hit 'Back'/close the new browser tab within 500 milliseconds"). A. di M. (talk) 15:56, 28 September 2010 (UTC)

This claim is far too over-generalized. Some text is link-dense on purpose, and has to be that way, such as glossary entries in stand-alone glossary lists, which are not usually read from top to bottom, but only referred to for clarification of a single term, linked to from its context in another article. Each entry in a glossary (or glossary style list of some other sort) is for practical purposes a stand-alone mini-article. I do get your point about link density in everyday WP prose, but part of this document's failure is that it is worded already in too blanket a fashion, such that too few editors know what sorts of exceptions can apply, or even believe that there are any exceptions.— SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 19:00, 28 September 2010 (UTC)

India is an example of a fairly densely linked article—much more than you'd normally want—but what overrides this is the fact that almost all of the links are highly specific targets; for example, "agriculture" goes not to the lame target you commonly see in country articles, but to "agriculture in India". "Corruption" is not the annoying dictionary link that is so irritating, where you really do mind that it dilutes high-value links; no, it's piped to "corruption in India". Nice one. The only problem with such piping is that the pipes are "deceptive", and likely to suppress clicks because readers are so used to the silly targets that WP can be littered with. You hope that if a reader bothers to try one of these links in this excellent article, they'll soon realise they're not being taken for a ride. Having said that, I just removed a few silly ones, such as "economic system" and "capitalism" and "private enterprise"—these are dictionary links, and spoil what is otherwise excellent practice. Tony (talk) 16:14, 28 September 2010 (UTC)

That's a self-fixing problem. As more and more articles are written (as they obviously are - WP is growing, not shrinking), link quality will improve (articles like Agriculture in India will increasingly exist, and even become more fine-tuned - Traditional agriculture in India, Factory farming in India, etc.) As link quality inmproves, readers will make more use of the links. So, there's really no issue here. Wikpedia's not broken; we don't need to try to "fix" it. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 19:00, 28 September 2010 (UTC)

Yes, I agree that the India article is a good example (and thanks Tony1 for taking the time to improve it). "Dictionary links" is a good term to use as we move forward with the common goal of improving the quality of articles.  HWV258.  23:02, 28 September 2010 (UTC)

[outdent] I have to comment that using the Professional wrestling article as exemplary of, well, jack is totally absurd. The entire field the article is about is fraudulent nonsense, pandering to the lowest common denominator fanbase. The article, like any article about a pop-culture topic, is frequently if not mostly edited by its random fans, not by neutral parties much less experienced editor. Given the nature of that fanbase, anyone intelligent enough to not be a part of that audience should automatically expect that article to engage in overlinking (as well as bad grammar, spelling errors, original research, fancruft, unsourced statements, vandalism, violations of WP:BLP, and every other article flaw, often many at once), except when periodically cleaned up by experienced, non-fan editors. Using that of all articles as some kind of test case makes this entire thread as absurd as something from Alice's Adventures Through the Looking Glass. — SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 19:32, 28 September 2010 (UTC)

If we can all agree that the issue with country links is whether they are often clicked, not whether they are ever clicked, then we can stop bringing back examples like linking dog, which one person in a billion (in my superhumanly magical opinion) would click if they suddenly had an urge to read about getting a dog. Art LaPella (talk) 20:48, 28 September 2010 (UTC)

We're not debating how to handle only the "good" articles, and all readers have the right to encounter well-constructed articles. (I don't agree with the "lowest common denominator" value judgement comment—lots of the fans are awfully nice chaps).  HWV258.  23:37, 28 September 2010 (UTC)

The real dispute: Relevance (explanatory vs. navigational linkage)

I think we should really get down to where the real dispute lives and resolve it: "Relevance" and whether a link to x "is" "helpful" to "readers". I "scare-quote" all of those because definitional devils are in the details of all four words. While I think it is very clear the more "deletionist" or "anti-linking" subset of editors of WP:LINKING have unreasonably and fallaciously suggested both that the noob practice of linking just about everything will spread like wildfire and that the more moderate or "inclusionist" or "pro-linking" editors of WP:LINKING have actually advocated such linking – both of these are abject nonsense, they actually do have a valid point, especially with regard to places. And it's not just noobs that overlink in this case. Here's a passage I pulled from a billiards article largely written by a good and experienced editor (and admin), and also worked on by me and various other people for several years now:

This marked the demise of professional straight rail in the U.S., which only had a six year run from 1873 to 1879. Meanwhile, straight rail professional play continued in Europe, with high run counts consistently climbing. Frenchman Maurice Vignaux posted a 1,531 count in Paris in 1880, while American George Spears had a high run of 5,041 in 1890. Later runs of over 10,000, in addition to the one previously noted, have been accurately reported.

I think it can be reasonably argued that these links to French people and United States (they're not even a consistent type of target!) are not particularly useful. I know for a fact that some editors find them a "sea of blue" distraction and a devaluation of linkage, since they've said so. What I'm not seeing is their recognition that - whether they personally feel this way about lead links or not - they're outvoted on lead links. The vast majority of editors and readers apparently do feel that an opening line in an article lead like:

Maurice Vignaux was a French professional player of carom billiards...

certainly is an appropriate use of a country link. The astronomically overwhelming majority of such cases are linked. This is solid proof of consensus for this practice, and we all know very well that a handful of holdouts against a landslide in favor of what amounts to a defacto standard on WP does not indicate a change in consensus. On the other side, the perhaps increasing (certainly existent and certainly alleged to be increasing) frequency with which the former, in-middle-of-random-sentence country link is deleted without dispute (or never put there to begin with) is strong evidence that there is not consensus for performing such secondary links and that they do constitute over-linking in most cases. (An example of a case where this wouldn't be the case might be "John Smith, UK ambassador" - his country is intimately connected to his relevance as an ambassador in whatever context he's likely to be appearing in, so identified.) I think a clear compromise is in order here. If I remember correctly, even Tony1, who feels strongly about this stuff, didn't say that a WP:LEAD link to a country was harmful, only that he'd never bother clicking it himself. I also can't remember anyone on "my" side of the debate suggesting that WP would fall apart if we didn't always, always, always provide links to a country just because it's mentioned in passing. From countries we can generalize to other things, like occupations (note in the quoted billiards passage no "player" link), and so on.

If we could get consensus on this here, I think a lot of the debate would be settled. And it would probably take very little additional wording in the guideline text to explain the difference between these two kinds of links – as someone said elsewhere, they are explanatory vs. navigational. It is the very job of the lead to explain, summarize and put into context the article's topic, while this is not at all the purpose of the average sentence in the average article paragraph, which is to elucidate some detail. The point of most of the prose is to present information as a stream, and excessive linking does impede the flow (not as much as come claim, in my opinion, but I do not deny that the effect exists). The lead is nothing like this at all, but is something of a one-stop shop for the gist. (The fact that infoboxes are like this on steroids is why stuff should be linked in them as well).

— SMcCandlish Talk⇒ ʕ(^Õل^ō)ˀ Contribs. 23:16, 28 September 2010 (UTC)

Better. I would have unlinked "Frenchman/French" in both those situations, but surely the lead is a more arguable case for linking it because it says the whole article is in some sense French, and we're no longer arguing in contradictory absolutes. To back up claims of consensus with some real (though quick) statistics: 7 out of 10 random articles with major countries in the lead linked them, and 4 out of 10 random articles with major countries further down linked them. Of course the statistics don't try to distinguish what data doesn't count because of alleged noob edits. Art LaPella (talk) 02:31, 29 September 2010 (UTC)

1. Who is going to click "France" and the "United States" here? Who? You seem to cast our readers as leisured browsers who want to wander through the site at random. That is patently unfair to those who want to be pointed towards valuable links. Have you read the huge, encompassing articles on France and the US? How do they help the reader, and again, do you really think readers will want to divert to them just as they're getting into a topic? For that matter, at any stage of their reading of the article.

   They don't have to divert to them immediately; they could open them in new browser tabs and read them after finishing reading the Vignaux article. (That's quite unlikely for such links as "France" and "United States" in a bio, but it's something to keep in mind in general.) Note that (at least with Firefox) you can't do the "Open Link in New Tab" thing with the Search button, so for readers who like to do that (e.g. me) being able to use the Search button doesn't help. A. di M. (talk) 09:28, 29 September 2010 (UTC)

   Seconded, as a fellow Firefox user. Also, "just use the search box" means navigating through any disambiguation page. All in all, a considerable downer on the Wikipedia experience. --Michael C. Price ^talk 19:16, 30 September 2010 (UTC)

2. The guideline resists the bunching of links that might well be seen as a single link. This is precisely what you've shown here. Why are the names red links? That seems deceptive, in this instance. Please replace them with blue-link names so we can see the effect.

   Because if the second link is red the "chain link" argument doesn't apply. In the Speed of Light article there's the phrase "gamma-ray burst GRB 090510"; right now the first link is vital, but when an article about GRB 090510 is created it'll have to be removed to avoid having two links looking like one and because the GRB 090510 article would link to gamma-ray burst in the first sentence anyway. A. di M. (talk) 09:28, 29 September 2010 (UTC)

3. "The vast majority of editors and readers apparently do feel that an opening line in an article lead [with a link to well-known countries] certainly is an appropriate use of a country link." You use the remnants of the pre-existing practice that grew for the first four years or so in WP's history—of linking without thinking about selectiveness—as evidence for your claim that editors, let alone readers, like them. Come now, you know this is not logical.
4. Could you write shorter posts, please? Tony (talk) 02:50, 29 September 2010 (UTC)

SMcCandlish, the problem with your second example is that it rests on the premise that editors linked nationalities because they thought it was a good idea, which may not be true. Based on what happened with date autoformatting (piggybacked on date linking), it is equally likely that editors just see that other articles were written that way, and emulated that style without thinking. Dabomb87 (talk) 03:14, 29 September 2010 (UTC)

I'm striking my words "back up claims of consensus" because that isn't what I meant. It has been argued that major country links throughout the article don't count because only noobs do that. But if that's so, then the same noobs presumably link major countries in the lede also. So if we ignore several statistical complications and simply subtract 4 out of 10 from 7 out of 10, we get 3 out of 6, which is no consensus at all. Art LaPella (talk) 04:12, 29 September 2010 (UTC)

Where is the consensus that "The vast majority of editors and readers apparently do feel that an opening line in an article lead like: Maurice Vignaux was a French professional player of carom billiards...certainly is an appropriate use of a country link."? It looks rather lame to me and I might take out the mention of the country from the lede altogether when copyediting. Make it something like:

Maurice Vignaux (born in Lyon in 1905) was a professional player of carom billiards...

Wouldn't that be more interesting for the reader?--John (talk) 06:17, 29 September 2010 (UTC)

Agree with John, although if there's no other context to indicate where Lyon is, I don't mind "(born in Lyon, France, in 1905)". Also, SMcCandlish, I find instances where "professional" is also linked. It all ends up a jumble of links. We owe it to our readers to be specific in both our wording (as John has exemplified) and in our linking. Tony (talk) 07:15, 29 September 2010 (UTC)

Dunno, the town of birth needn't be relevant to one's life. For example, after the age of a few months I've never lived in the city where I was born. (I think that's why the MoS discourages writing the place of birth in the first sentence, though I think that's too strong because there'll be many articles where it will be more relevant than the day and month of birth.) A. di M. (talk) 09:28, 29 September 2010 (UTC)

I've repeatedly argued, ever since I came across this issue, that relevance and navigability are the key issues for linking, rather than the idea of what is puportedly "useful". The guidelines say this. Common sense says this. Editor practice - even excluding the much-derided overenthusiastic linkers - says this. Yes there can be debates over how to apply this in each case, but there's a certain degree of objectivity there, and it's far easier than trying to assess what is "useful" or "deepens understanding" for people as a whole when we are dealing with millions of readers and editors, who have massively varying degrees of knowledge, who use WP for a hundred different reasons and in a hundred different ways, and who may be first time visitors to the site. With the examples raised, a link to France in an article about a French thing seems reasonable. A link to France when the country appears in passing mention, almost certainly not.

As for Tony's statement above - Who is going to click "France" and the "United States" here? Who? You seem to cast our readers as leisured browsers who want to wander through the site at random. That is patently unfair to those who want to be pointed towards valuable links. Have you read the huge, encompassing articles on France and the US? How do they help the reader, and again, do you really think readers will want to divert to them just as they're getting into a topic? - this also encapsulates the problems I've seen since I first looked at this. You know, to answer your question, people might. As noted, people come here from all sorts of different places and for all different sorts of reasons, and use the site in all sorts of different ways. I don't have to show that everyone will use a link to justify retention. I'm not sure how many would get there from any one specific link, but France and US are probably among the most viewed pages here. And I can pose some questions in return - why reduce the options for getting to them, from a related and relevant page? If the articles are so generally unhelpful, why have those pages at all? Where is your evidence that people want some higher minds to point them to certain pages and not others? N-HH talk/edits 11:01, 29 September 2010 (UTC)

"You know, to answer your question, people might [want to follow a link to France or the US]]. Then link every single word, because people might want to follow them. I'm interested in "the". There's an article on it. A link would aid navigability, and it's certainly relevant to a topic. If you want to link every word, get the developers to make a system that does link every word (without colouring it); this is what my desktop Encarta dictionary does. I don't mind that on Encarta, and it would be preferable to undisciplined linking on WP; except that in WP articles, it would disable us from pointing up the important, useful links. But total linking is clearly what you are striving for. Tony (talk) 12:59, 29 September 2010 (UTC)

Ugh. You're still doing this strawman crap and just making stuff up about what I have said or might even think, despite my being very clear about my view, and despite being asked not to about 1000 time. This is the other 50% of the problem. N-HH talk/edits 13:05, 29 September 2010 (UTC)

• It is best not to order other editors around and make personal attacks in edit summaries, and not to use aggressive language in posts. I stand by what I said; I'm surprised it seems to upset you—that was certainly not my intention. If you want to link France in such contexts, then why not link "the"? It is a perfectly valid question, and enquires into why you draw a boundary where you do. Please take it seriously, and do not personalise. Tony (talk) 14:05, 29 September 2010 (UTC)

Tony, it is best not to make things up and fling wild accusations about what other editors might want. You personalised this by suggesting, for oh, about the 98th time, in both your edit summary and your comment, that I "want to link every word", when you know full well that I'm not calling for that, and nor is anyone else. You've been asked politely to stop this and have ignored it again and again, so this time I was a bit more forceful. I don't claim that you want to "remove every link", because I read what you say, I understand what you say and I do not feel the need to misrepresent what you say. I explicitly said above that links should be to relevant and related pages. I said links to France where the country is mentioned in passing are, in my view, not needed. However, I've pointed out for example in the past articles about a BBC producer where every single link to "BBC" and "TV producer" were removed, even from the infobox. An article about a French Algerian, where the initial link to "France" was removed but that to "Algeria" retained. These are the things I am taking issue with, along with the exclusion of links to the article on the United States - a page that has 3m visits this month, which suggests at least that some readers here would like to see it - in, say pages about its neighbour Canada or things from the US, not efforts to remove supposed links to "the". Which, of course, in reality is virtually never linked, and which doesn't even have its own page anyway, which clarifies quite what a red herring it is anyway. Now, if you want to have a sensible discussion about the role that the general concept of relevance might play in the linking debate, and where the thresholds might lie for what is relevant and, indeed, for what constitute "common terms", feel free to enter that debate, although it will probably be with someone else. I've discovered again what a futile effort it is. As it is, oddly, to maintain any form of consistent talk page indenting with an MoS guru. N-HH talk/edits 14:32, 29 September 2010 (UTC)

How is the relevant to champagne? There are languages with no word for the and you can talk about champagne in those languages just fine. A. di M. (talk) 16:07, 29 September 2010 (UTC)

Why not link the? Answer: Editorial judgement - something that seems sadly lacking in some contributors here. --Michael C. Price ^talk 16:31, 29 September 2010 (UTC)

Linking France: "Where is your evidence that people want some higher minds to point them to certain pages and not others?" Linking "the": "Editorial judgment". See why we're going in circles? Of course some links shouldn't be made, and some should. Phrasing OVERLINK is a hard question, but getting past the irrelevant argument that some people will want to click a link should be easy. Art LaPella (talk) 20:24, 29 September 2010 (UTC)

But no one's making that argument, ie that we should link simply because someone might use it. When I made the point that some people might use a link to the US or France or whatever, I was simply responding directly to Tony's question - "who would clink on such a link?". That specific response was not intended to answer the more general question "what is the criteria for linking?" nor was it phrased as such. The rest of my post was very, very explicit about where I stand on that. Nonetheless, as ever, my answer was twisted and thrown back, and my own questions went unanswered. As for example these ones have for months, despite being raised on many, mnay occasions. The point about editorial judgment is, as noted, key. And it needs to be applied in the context of each page, not by blanket assertions that "everybody knows what/where X is", therefore the links to that term are going to removed from thousands of articles at a time, often by scripts and seemingly without review. There has never been any cross-WP consensus mandating a small number of editors to do that, and make it look like some semi-official clean-up.
Now there are going to be disagreements on links in each case of course, and as to whether they are relevant or not. But those debates are manageable and can be part of normal copy-editing (and probably won't include discussions about whether the word "the" is a relevant/related topic that is ever worth linking). Also, basing decisions on relevance is far easier than trying to work out en masse what terms and/or detailed information is definitively "known" or "common" when we are talking about millions of readers. However, a) we need to get to the point of accepting the principle at least; b) we need to read and accept where the guidelines call for the inclusion of links to other relevant pages, whether common terms or not; and c) I would argue we should err on the side of retaining a link if it's a 50-50 case, since more "harm" arises from removing the options for navigability than from having the odd extra blue link (and we are usually talking only about one or two links). N-HH talk/edits 11:54, 30 September 2010 (UTC)

Your explanation relieves my main objection, regardless of what you said or didn't say before. So if necessary, I will remind you of your own explanation in the future. Art LaPella (talk) 18:32, 30 September 2010 (UTC)

Sorry, but I don't quite understand what you mean there. N-HH talk/edits 19:14, 30 September 2010 (UTC)

OK, then you relieved my main objection. Did you get that part? Art LaPella (talk) 21:30, 30 September 2010 (UTC)

Well, I'm not sure what exactly your main objection was, or what exactly you mean by saying that I relieved it. Sorry, not being pedantic, just confused! Either way, we seem to be settled as far as I can tell, so probably no need to keep on. N-HH talk/edits 10:04, 1 October 2010 (UTC)

Break

---

• SMcCandlish wrote as follows:

Maurice Vignaux was a French professional player of carom billiards...

…and then wrote “The astronomically overwhelming majority of such cases [the “French” part] are linked.

Just because a lot of editors do something doesn’t mean it is desirable; often people do things because they are simply following the other soldiers to the chow line.

In the above example, I wouldn’t bat an eye if I encountered that “French” link in the first paragraph in the lede of a biography. But at all times, I try mightily to avoid linking words and terms—particularly in the main body text—if they are in any way tangential to the *real* subject matter.

Our practice towards linking has certainly changed over the last two years and the general principle nowadays is to link only those words and terms that help the reader to better understand the subject matter of that particular article. I might add my own flavor in all technical writing, which is as follows:

Thou shalt not cause needless confusion in thy target readership nor induce *!* brain-interrupts in thy readership by doing odd, unconventional, or unusual things.

So if the article was Notable mistakes in the history of science, I would be tempted to link as follows:

• In 1966, an American scientist presents a paper on what would later be known as polywater.

But in this one…

• In 1966, a Ni-Vanuatu scientist presents a paper on what would later be known as polywater.

…I’d link where the scientist came from because the demonym, which is for someone from Vanuatu, is oddly done and that country is much less familiar to the middle of the bell curve of our readership than is a country like the U.S.

And certainly, I would not link an entry in an article titled “Notable mistakes in the history of science” like this…

• In 1966, a American scientist presents a paper on what would later be known as polywater.

This latter example is rich with extraneous links that do not help our readership to understand the subject of “scientific goofs”. This assumes we define “the middle of the bell curve of our readership” with some common sense and that we are not writing for the lowest common denominator such as a 4th grader from somewhere in East Africa where they power the village computer from a solar cell. The vast majority of our readership that will be visiting this article already knows what “scientist,” “scientific paper,” and “U.S.A” are, and the “1966” article has precious little else to do with “scientific goofs.”

Three years ago, it used to be “I link, therefore I am.”

Today it is “Does this link really help the reader to understand this particular subject?”

Finally, if the article is short and/or is very linear in nature, I never re-link a word or term. But if the article is very long or has sections that tend to stand alone in their own right (like our Anwar al-Awlaki article, which tends to have a large portion of our readership doing parachute drops directly to certain sections), then I will often re-link key words.

Greg L (talk) 21:54, 29 September 2010 (UTC)

Broadly I would agree with most of that, subject to previous comments about the problems inherent in focusing on what we think might help the reader understand the subject (eg how do we know? which reader? what about other criteria? etc etc). We don't need, usually, to repeat links or link things that are mentioned in passing and/or are not directly related to the topic at hand. The problem I have is that I often see de-linking that is pushed much harder than that, leaving inconsistencies, or taking out every single link to pages that are pretty clearly, by any standards, relevant and probably - or maybe just arguably - worth including. There are some examples here. N-HH talk/edits 12:05, 30 September 2010 (UTC)

Guidelines are just that: guidelines. I didn’t see anything alarming in the first difference you provided. There was a lot of much-needed cleanup Ohconfucius addressed in that article and there is room for honest disagreement on some of them. Ohconfucius was educated in England so maybe he thought it blindingly obvious that the target readership ought to know what “BBC One” means. Or maybe he thought it is linked in the template and that’s enough. The article now has “BBC One” linked in the first occurrence in the lede of the body text and that makes the most sense to me.

More to the point, what I see in today’s version of the article seems to properly reflect the broad principle of “link to enhance the reader’s understanding of that particular subject.” The community managed to do the right thing using the existing guidelines for direction.

This seems more like a content dispute over different editors’ views on how best to adhere to the spirit of the guideline as it applies to a particular article than it is a dispute over how the guideline has any shortcomings. Am I missing something here? Is all of this an issue originating over how “BBC One” was linked in the template-based sidebar (which I seldom read) and that was the basis for his not having it linked in the body text? Do we really need to have atomic-level nuances spelled out in MOS to settle edit disputes? Common sense seems to have been sufficient in this case since the article seems to have come out all right after a consensus view on that article was achieved.

The distinction between sidebars and body text is precisely the sort of thing I was talking about when I wrote about readers parachuting directly to a section. IMO, even if it’s linked in a sidebar (or caption), the first occurrence of a key word or term in the body text should still be linked. And if the article is of a long, nonlinear nature with standalone sections to which readers often tend to directly skip (like Anwar al-Awlaki), important links to critical or unusual words or terms may be re-linked. Greg L (talk) 17:35, 30 September 2010 (UTC)

Well, yeah my position overall is that a lot of the mass de-linking is blatantly in breach of the guidelines, as currently written, quite apart from any debate about changing them. And just a bad edit, regardless of any guidelines (although the guidelines are, as it happens, usually invoked by those involved). Removal of text or functionality has to be justified in some way, whether by reference to guidelines, policy or whatever - the burden is that way round. And it is a mass de-linking - the example we're looking at here is just that, an example. Yes, I restored some of the links there because I happened to spot it. Those partial reverts stuck and so a resolution of some sort was reached. But we are talking about thousands of articles, and arguably problematic changes that are being made minute-by-minute across the site without consensus. I don't have the time or the inclination to go through them all and reinsert individual links or argue the toss over each of them. Why should I, or anyone else, have to? Wouldn't it be easier if people just held off making some of them in the first place? N-HH talk/edits 19:24, 30 September 2010 (UTC)

Should we raise an RfC to address the breach of guidelines? --Michael C. Price ^talk 19:47, 30 September 2010 (UTC)

Still more wikidrama is indicated over another dispute over linking? (*sigh*) Greg L (talk) 21:40, 30 September 2010 (UTC)

Well, I just don't think I would have anything to say to anyone who suggests that, where the BBC – the most well-known organisation in the world – is linked to many fold in an article about a sitcom, these multiple links should not be removed by "mass delinking" – such is this classic case or Overlinking, and the reason I see for the built-in search and replace tool. If anybody thinks that is a breach of the guidelines, I'd retort that you were talking out of your rear orifice (pardon my French). --Ohconfucius ^¡digame! 01:15, 1 October 2010 (UTC)

I'm American and I recognize the BBC, although it might be better to link it anyway (I'm an editor, not a reader). I hadn't heard of BBC One before Wikipedia, so that should surely be linked. But just once. Well, at least not repeatedly in the same paragraph. (Oh, and you were the first to mention anybody's orifice, so I consider that an unnecessary escalation.) Art LaPella (talk) 02:44, 1 October 2010 (UTC)

• Whilst I accept that I may err in unlinking some words, but when, as it appears, there are complaints about the entirety of an edit, then I would take issue that I was acting within the bounds of WP guidelines. If it's a case of someone believing one or two links were overzealously undone, I can also accept –depending on the word, of course– they would be replaced, per WP:BOLD, just as I removed them. --Ohconfucius ^¡digame! 09:37, 1 October 2010 (UTC)

• Not if "BBC One" is linke in the same sentence, or thereabouts, please. It is a breach of the guidelines to plaster the readers with less focused links in the same area of text as more focused links. The more focused always have a link to the less focused at their opening. How many times would a reader want to divert from an article directly to both broad and focused topic? It is a textbook example of diluting a system that needs to applied with some care if readers are to click at all, on anything, in the first place. And the notion that then "BBC" should be linked again and again through the article is going to turn back the clock to the days when wikilinking was blue paint spattered through our text, and rightly ignored with irritation by readers. Tony (talk) 03:15, 1 October 2010 (UTC)

But. I. Am. Not. Asking. For. Repeat. Links. How. Many. Times. Do. I. Have. To. Say. That. JFHC. I am taking issue with the removal of every single link to so-called common terms, on the basis that "well I know what it is, and I guess maybe most others probably do as well, so we shan't be linking to that article at all. Even though it is relevant, related, and the page carries more information on that second thing than simply a dictionary definition". Also with the assertion that editors have the right to do this as of right, and that it's up to others to go around after them tidying up by occasional reinsertion. N-HH talk/edits 10:04, 1 October 2010 (UTC)

You raise the "why not link France at the top of the article on Champagne (that is, the wine)". Because "France" is just to broad to assume that it is useful. I keep asking why you wouldn't link to a more specific article (or even section), than "France". Like "Wine production in France", or less satisfactorily, "Agriculture in France". The guideline requires specificity. Tony (talk) 10:36, 1 October 2010 (UTC)

Well, we're still going round in circles. Useful in what sense? Useful to whom? Who are we as one or two individuals to assume anything about how to define those concepts anyway, uniformly for millions of users here? As it happens, I wouldn't object to more specific links where appropriate. Having said all that, the guidelines a) discuss navigability as well as usefulness, narrowly defined; b) say more specific links should be "considered", not that they are "required"; and c) do not rail against "less focused" links being in the same area of text as "more focused" links, again, however we would define those anyway, (to respond to a point from your previous comment). I fully support people going through articles removing redundant, trivial or repetitive links, or links to things mentioned in passing, and improving links where necessary so they point to a more focused target. I occasionally do it myself. But I do not, for the fortieth time, support the removal of every single link to so-called "common terms", however defined, when they are relevant or related to the main topic. Please, remove 20 repeat links to "France", "BBC" or even "Hastings-Thompson syndrome", as part of a normal copy-editing process. But please, also leave just one of the former, where the term is plausibly relevant or related. They're likely to be useful to some people, and are not that distracting really to people who don't wish to use them. N-HH talk/edits 11:06, 1 October 2010 (UTC)

MoS update

It is clear that the current version of the MoS does not reflect consensus. There are many areas that do not reflect practice, nor the views of the majority here. In order to develop a consensus, and in line with George Louis's suggestion, a proposal for the repeat link section has been developed, which I copy below.

Of course the repeat link section is not the only section that needs updating, but we have to start somewhere. Comments, please.

Repeated links

In general, link the first occurrence of an item in an article. In addition, consider linking again:

• Where an unlinked occurrence of an item is a long way from a linked occurence
• Where the first link is in transcluded content such as an infobox or sidebar template, in an image caption or a table, or otherwise outside the main article prose
• In multiple table cells, if links are needed in the table, as each row must stand on its own (tables are re-sortable).
• In a glossary (treat entries as if they were separate articles)
• In a group of other linked items.

Rationale

Two types of links must be distinguished. Explanatory links to terms, concepts, places, etc., that many readers may not be knowledgeable about or familiar with at all are often helpfully linked more than once in an article (e.g. in the lead, and again when they appear in context in later sections). Navigational links, to things that most readers are at least somewhat familiar with, are typically repeated less. But even nagivational links may be helpful to repeat in particularly long or complicated articles, as readers often use the table of contents feature to jump from section to section and do not necessarily read articles in linear order or in their entirety. Always bear in mind that content may be repurposed in ways not immediately envisioned by editors of linear articles (e.g. in a mobile computing version that displays only one entry at a time with no ability to present the full article at all). With restricted screen sizes (e.g. mobiles, again) some readers will find items to be a long way apart, that others would deem closer. Stand-alone items should be linked independently, since they are often read in isolation. For example infoboxes, navboxes and other transcluded content. Items should be linked appropriately in main prose as if the infobox did not exist, and in a navbox or sidebar as if the main prose were not there, since they are all separate, stand-alone entities. Readers can also re-sort tables in any order, so link items repeatedly in a table, since no occurrence can be assumed to always be the "first". A stand-alone glossary article or similar stand-alone list with entries that are mostly linked to directly from other articles, and which is not principally intended to be read from start to finish, but referred to piecemeal, should use repeat links: Each entry should appropriately link to terms both within and outside the page as if entries were independent articles, as few readers ever find the first occurrences of the terms. It is occasionally better to link a second rather than first occurrence of an item, when both are very close to each other, and linking the first would distract from readability or focus. Not linking "transclusion", "infobox" and "glossary" in this the above list but in the follow-on prose instead is an example; it is more important for readers of this guideline to absorb these points quickly than to provide explanatory links for these terms immediately. Linking a second, explantory appearance of an item is most common when the first occurrence is in a list, table, quotation, formula or complicated passage, or where it would result in two different links being side-by-side giving the misleading appearance of a single link.

--Michael C. Price ^talk 12:37, 30 September 2010 (UTC) & --Michael C. Price ^talk 03:00, 6 October 2010 (UTC)

• I *get* what you’re driving at. But what you propose here seems too complex and wordy. It also seems to be a prescription for too much re-linking; maybe that is just an unfortunate side-effect of trying to be so detailed and definitive. I think the tweaks to the basic principle can be addressed with much less verbosity. All the community needs on this are the basic principles. Less is more for these sort of things. Greg L (talk) 17:41, 30 September 2010 (UTC)
  • It is more complex, but current simplistic approach (which I mean without prejudice) is not working, as this talk page testifies. Sometimes more, even just a bit more, is more.

• Well, I didn’t mean “so simple that what we currently have suffices.” I meant “simpler than your proposed addition.” Sorry, I should have been clearer. I’m editing on the road with a laptop, don’t have my standard macros or even a mouse, and everything is awkward. See my below green‑div. Greg L (talk) 21:26, 30 September 2010 (UTC)

• Re verbosity, perhaps the explanation can be trimmed or hidden in a note? --Michael C. Price ^talk 19:00, 30 September 2010 (UTC)

When you say, "It is clear that the current version of the MoS does not reflect consensus.", it contradicts the fact that the current text is long-standing and is the result of carefully worked-through consensus. The fact that a few editors have been voicing objections to the current practices does not at all mean there is no consensus. Tony (talk) 07:32, 1 October 2010 (UTC)

And people have been telling you that it is wrong ever since you and a few other editors rammed it through. The number of unanswered points on this talk page should shame you. I predict that my apples and oranges argument has joined the list.... --Michael C. Price ^talk 10:41, 1 October 2010 (UTC)

Not only is the text long-standing, but the effects of the text have been long-standing. This debate must consider article space. It has been the experience of the (experienced) editors involved in this arena (over the past year or two) that delinking nebulous and dictionary-type terms has overwhelmingly not led to reverts and wailing-and-gnashing-of-teeth-type edits. Of course we're all happy to discuss change, but we must stay grounded and be careful how we introduce creep into the MOS wording in order to solve a "problem".  HWV258.  20:34, 1 October 2010 (UTC)

Prediction confirmed. --Michael C. Price ^talk 20:57, 1 October 2010 (UTC)

Links, the whole links, and nothing but repeating links, so help me RfC (and other wikidrama)

IMHO, the entire principle of what we need for linking on the entire page could be captured with the following green‑div:

What and what not to link

Link only those words and terms that help the reader to better understand the subject matter of that particular article. Even if a word or term is linked in a sidebar (or caption), its first occurrence in the body text should be linked.

Imagine an article titled “Notable mistakes in the history of science”, linking as follows is good practice:

12. In 1966, an American scientist presents a paper on what would later be known as polywater.

However, if the bullet point in the same article was as follows:

12. In 1966, a Ni-Vanuatu scientist presents a paper on what would later be known as polywater.

…it might be better to also link where the scientist came from because the demonym, which is for someone from Vanuatu, has an odd construction and Vanuatu is much less familiar to the majority of our readership than is a country like America.

Do not link an entry in an article titled “Notable mistakes in the history of science” like this…

12. In 1966, a American scientist presents a paper on what would later be known as polywater.

This latter example is rich with extraneous links that do not help the vast majority of the readership to understand the subject of “scientific goofs”. The vast majority of our readership are coming to the article to learn about scientific goofs throughout history and they already know what a “scientist,” “scientific paper,” and “American” are. Moreover, the “1966” article has precious little else to do with “scientific goofs.”

Repeating links
The majority of Wikipedia’s articles are sufficiently short and linear that editors should not relink the same word or term in the main body text.

However, if the article is of a long, nonlinear nature featuring strongly standalone sections to which readers tend to directly skip or are often directly linked to from elsewhere in the project (like our “Anwar al-Awlaki” article), editors should consider relinking important, critical, or unusual words or terms.

There, how simple is that? Greg L (talk) 21:15, 30 September 2010 (UTC)

I have a problem with the first sentence; specifically the word "only" :-) As we have already discussed, links are also to provide navigation to any pages the reader may be interested in. As the MoS says:

Ask yourself, "How likely is it that the reader will also want to read that other article?"

And at the end I would replace "important links to critical or unusual words or terms" with "terms" - even simpler! --Michael C. Price ^talk 21:46, 30 September 2010 (UTC)

• Hmmm… Well, if you have a problem with this sentence: “Link only those words and terms that help the reader to better understand the subject matter of that particular article.” …because it contains “only”, then I think your views are truly 180° out of synch with the modern view of linking, which is to avoid blue turds like “The sewer cover in front of Greg L’s house”

  BTW, I consider the above green-div to be a live document. I just tweaked the section regarding re-linking (that whole concept ought to make you happy).

  I’ll let others weigh in and will no longer deal with you for a while since you just made your views abundantly clear and engaging you on this subject seems pointless. Moreover, your above call for an RfC (a rhetorical question that is in fact a solicitation of support of the proposal) shows to me that you exhibit a fondness for wikidrama and needless confrontation. It seems all my effort here at finding a middle ground was utterly pointless and a waste of my time. Happy editing. Greg L (talk) 22:03, 30 September 2010 (UTC)

• Quite how you rationalise away the line I quoted from the MoS is beyond me.

  No matter, the principle is quite clear. If the article on "oranges" mentioned "apple", I would expect it to be linked, yet reading about apples doesn't "deepen my understanding" of oranges - but a reader of one might like to read about the other.

  --Michael C. Price ^talk 22:11, 30 September 2010 (UTC)

• How is a "nonlinear" article different from a "linear" article? What is "nonlinear" about Anwar al-Awlaki? Tony (talk) 03:21, 1 October 2010 (UTC)

While we're on the topic of the linking policy, I think there's another important exception, which is symmetry. If a sentence or a list contains links to a number of similar/related concepts or words, you shouldn't not link one of them just because it was linked earlier (say in the introduction). For example "The primary colors of light are red, green, and blue." looks really weird. Dcoetzee 05:19, 1 October 2010 (UTC)

Yes, a very ugly result. --Michael C. Price ^talk 05:24, 1 October 2010 (UTC)

I wouldn't call it "very ugly", but that one may be worth considering, along with table columns (but not where all items would normally be unlinked, such as years).

I am quite unconvinced of the calls for repeat linking based on "a long way down" or "nonlinearity"; these are nefarious concepts. Should we then call for acronyms and initialisms and all other abbreviations to be spelled out on a similar basis? This is not practical and defies standard practice. For a long time, WP articles have been treated as whole, integral entities WRT a number of major stylistic aspects. Suddenly changing tack just for wikilinking is a departure from this that would need the bigger picture examined. Tony (talk) 07:16, 1 October 2010 (UTC)

I'm ignoring your strawman argument.

WP articles are not treated as a whole - that's why we can link to sections and subsections, for example.

--Michael C. Price ^talk 07:34, 1 October 2010 (UTC)

I agree with Tony1. "A long way down" is very difficult to define, and we would be taking a step backwards with such a phrase.

I'm against repeated links in table columns. I've had a lot of experience with table editing at WP, and the repeated linking of column entries (irrespective of sorting) is not desirable. There are almost always better ways to handle the problem of repeated links in tables—it just takes more time and skill to solve the problem. For an example, have a look at how I handled the problem here.

 HWV258.  07:49, 1 October 2010 (UTC)

Of course "a long way down" is subjective and difficult to define - as are most terms in the MoS, which is why we leave them to editorial judgement. --Michael C. Price ^talk 10:37, 1 October 2010 (UTC)

A MOS must set a reasonable framework to facilitate good judgement. In my opinion, the amateurish "a long way down" does not set a reasonable foundation.  HWV258.  20:19, 1 October 2010 (UTC)

Greg L's version also mentioned "long", as does the current MoS version. --Michael C. Price ^talk 20:54, 1 October 2010 (UTC)

I wouldn't link to "red" in "red, green and blue" if "red" has already been linked; this is akin to the standard practice of giving the full name for people being mentioned the first time and the last name only for people already mentioned, even in coordinations. What I wouldn't do is writing something like "Coventry, Leicester, Manchester and Wolverhampton" on the ground that Manchester happens to be on some list of cities the readers are presumed to already know: even though a reader actually had already heard of Manchester but not of Coventry, Leicester or Wolverhampton, they'll still wonder why Manchester is not linked, and maybe even suspect that there was a previous linked mention of Manchester in the article that they've somehow missed. A. di M. (talk) 13:01, 1 October 2010 (UTC)

I prefer the currently existing "What should and what shouldn't be linked" section to the one in the box above, but I prefer the "Repeating links" section above to the currently existing one (though I'd rather use Glossary of cue sports terms than Anwar al-Awlaki as an example of an article where readers often skip sections). A. di M. (talk) 13:01, 1 October 2010 (UTC)

Bye

I thank you all for wasting my time. There have been more than two dozen edits to this page since my last one and hardly any of them did contain any useful idea to improve the guideline. I'm unwatching this page; if you need me (which I think you won't), you know where to find me. As for me, if I need more drama (which I think I won't), I know where to find it. A. di M. (talk) 11:35, 2 October 2010 (UTC)

• Ditto. Me too: Out of here. Adios. And thanks for giving your endorsement of sorts to my suggested guideline for re-linking. It’s actually a practice I and some other of the “de-link jihaders” do (or whatever the heck we’re referred to). I thought I’d throw out there as it seems to reflect the practical reality. The proposal went about as far as I could drool when I was asleep in 7th-grade history class. Greg L (talk) 13:55, 2 October 2010 (UTC)

Consensus and arbitrary removal of relevant common terms. Again

Since no one has seen fit to respond to these very simple and specific questions in all the months this debate has rumbled on - and since one editor has now on three separate occasions shunted the thread into the archives, which acts as effective deletion, on the second occasion bizarrely claiming that there was "nothing productive here but diatribe and personal insults" - I'm going to ask them again, and give people another opportunity to answer.

• Where is the consensus to delink common terms/countries in every instance, regardless of context; and
• Where is the consensus as to which terms/countries fall within the definition of "common"?

.. and also add a couple of newer examples that highlight the problem here (disclaimer - I'm not necessarily rejecting these edits in their entirety, nor am I necessarily condoning those that I have not picked out)

• Example 1 - Bolivia. This edit removed links to "Argentina" and "Brazil" from the lead. But left "Chile" and other neighbouring countries. It also removed a link in the infobox to "Spanish language", but left another - admittedly more obscure language - linked. Why? These are all relevant terms/topics to the subject of Bolivia. We can all argue theoretically about "utility", but the point is more simple than that - what benefit is there to removing them? Why leave an inconsistent and arbitrary hodge-podge of links?

• Example 2 - Mayors in England. This edit removed links to "England" plus "town" and "city". Again, navigability is reduced as links to rather obviously relevant terms - when discussing local administration for *urban areas* in *England* (the clue's in the page title) - are removed. And, again, what benefit is there to this loss of functionality?

Now I may be repeating myself here, but that is because the problem repeats itself. Edits like this are still being made all the time by a small group of editors, marked as "minor" or "fixes", and justified by reference to wp:overlink. But I still can't see how edits like this - or that highlighted portion of them - make sense in any way, whether by reference to basic editorial judgment or to the guidelines we have here. I can only assume that were they put to the broad mass of editors, they'd be scratching their heads as well. N-HH talk/edits 16:42, 4 October 2010 (UTC)

I agree. The choice of links to unlink in your two examples makes no sense, and if the examples are typical then there is a problem with the way these two editors go about it. I am guessing that they simply have a list of overlinked words, and that both are using a script that mechanically removes every single link to these words. That can make sense if you check what the script is doing and manually re-link everything that does make sense, or in some cases also remove other, related instances of overlinking. But these two examples show that it does not make sense if such a script is used sloppily or carelessly. Hans Adler 20:22, 4 October 2010 (UTC)

You [N-HH] probably don't like my answers to those two old questions, but I did notice the questions (after I joined the discussion) better than you noticed my answers. OVERLINK says to unlink major countries. We don't define "major", but Argentina and Brazil are the most obvious choices in South America, because they have more population. Removing only two countries from a list is debatable, but there were only five countries, and in this context neither Argentina nor Peru will get jealous, so it doesn't seem worth going to war over. Similarly for languages. Given that this is a plausible interpretation of OVERLINK, what benefit is there? There's no obvious way to describe the benefit of OVERLINK without upsetting you over an azure body of water you don't believe in. How relevant are "city", "town" and "England" to "Mayors in England"? Oh, more than in most articles I suppose, though of course not as relevant as some other nearby links in the article like mayor and lord mayor. What I'm sure of is that removing two borderline links should be marked as minor, although I don't mark my edits as minor when they include as many subchanges as OhConfucius's edit. Art LaPella (talk) 20:53, 4 October 2010 (UTC)

I think you are missing the point. “Link only what is required for the reader to understand that particular subject” means the context of the article is all-important in deciding whether countries should be linked at all.

First off, Americans are notoriously ignorant of other countries. At least for en.Wikipedia (where a large portion of our readership is American), many readers would not know any South American countries. But everyone who uses Wikipedia knows they can type something into the Search field if they want to know more about it. If the article is about countries, then I don’t see a problem linking the first occurrence of every country no matter how “well recognized” it supposedly is. But for most articles, if I’m reading up on Sucralose, I don’t need to see “Sucralose was first approved for use in Canada in 1991.” And in fact, the article currently does not have “Canada” linked. Good so far.

But I also note that the “Sucralose” article still has this:

Tate & Lyle manufactures sucralose at a plant in Jurong, Singapore. Formerly, production was completed at a plant in McIntosh, Alabama. It is manufactured by the selective chlorination of sucrose (table sugar), which converts three of the hydroxyl groups to chlorides.

Only two terms or words are valuable ones that are necessary to help the reader understand the subject “sucralose” (hydroxyl groups and chlorides). Cluttering up an article with more blue so a reader can stop what they’re doing and go learn about McIntosh, Alabama is just not using the noggin. One in a ten thousand readers will want to stop reading about sucralose so they can take a tangent about some town in Alabama. So de-link it. It doesn’t matter if it is an *obscure* location; that just makes it even less likely it will be clicked on considering the subject matter of the article. Greg L (talk) 21:44, 4 October 2010 (UTC)

"Where is the consensus..." is the wrong question. It is on the MOS (and has been for ages) so consensus is deemed to exist. I recognize that you don't like that, but it's a fact. Also a fact is that the associated delinking, which has been happening for a long time now with many, many thousands of edits and no significant backlash, reinforces the current position. No one has stated that every single edit has been correct (this is a wiki so we fix the problems as they happen), but the wording of "links should be created to...relevant connections to the subject of another article that will help readers to understand the current article more fully...as long as the link is relevant to the article in question" has broad and deep support (time and edits have proven that). It is that wording that has lead dedicated and experienced editors to act to improve WP for our readers—and that is their right (and I thank and encourage them for their hard work). If you are unhappy about any of the edits, take it up on the particular talk pages affected. Have you considered that the edits (although appearing like a deep pruning) are designed to set a new baseline in an article, and to instil a paradigm (hopefully causing the localized editors to take note and think about their future linking)? No one denies the right of localised editors to relink on their favorite pages, but it must be observed that in an overwhelming number of cases, they don't. If any local issues do arise, my experience has been that the delinking editors have been more than happy to take the time to politely discuss the edits. I'm asking the community here to debate with the context of the entire project in mind, and not just the handful of editors that can bring no significant real-world examples of problems to support their attempt to change the wording of the MOS.

Put simply: one or two eggs may break in this process, but we're making an awfully big omelette (with a perfectly cooked main part, and just the right amount of tasty bits to make you want to come back for more).

Personally, I'm surprised that Example 2 was raised above as it is a sensible delinking that addresses a number of problems, and four days after that edit no one has blinked a revert-type eyelid at that page. If that's a counter-example demonstrating some type of "problem" then I can can rest easy and log off (which I now intend to do).

 HWV258.  23:26, 4 October 2010 (UTC)

• No kidding with regard to “Where is the consensus?” It’s ludicrous to think that the current policy was just some “oopsy” or skullduggery pulled off at 2 AM by Evil De‑linkers©™® so that the *real* community just woke up one morning not realizing that a fast one had been pulled on them. Does N‑HH think the current guidelines were not the product of vigorous debate and compromise over an extended period and was the product of a general consensus? And now someone has to prove to N‑HH’s satisfaction that all that outcome was legitimate because he is experiencing a sense of wholesale dissatisfaction with it??? Puuuuhleeze.

  I don’t know about anyone else here, but I consider myself to be a volunteer on a hobby of mine, which is the collaborative writing project that is Wikipedia. I elect to contribute my time to the project in the manner of my choosing and certainly don’t intend to spend any of it taking N‑HH by the hand and guiding him through the months-long blow-by-blow that got us to where we are today.

  If he elects to add a bunch of extraneous links to articles just because they can be linked to even though they don’t have jack to do with helping the reader to understand the subject at hand, that is his prerogative. He shouldn’t act all surprised though when others de‑link all those extraneous links. Greg L (talk) 00:59, 5 October 2010 (UTC)

Why does diverting to town or city deepen a reader's understanding of that topic? Pillar: "WP is not a dictionary". Tony (talk) 02:12, 5 October 2010 (UTC)

• BTW, Tony, I liked the way you linked words “town” and “city” in your above post. Greg L (talk) 03:12, 5 October 2010 (UTC)

• My Bolivia edit was the result of a fair amount of deliberate manual work – a good half hour, IIRC. Yes, I do have a list I work to. Many articles suffer from overlinking of the same group of terms, so it is no surprise that my systematic work may look automated. It did cross my mind, and I'm sorry that I didn't also unlink Paraguay, Chile and Peru when I unlinked Argentina and Brazil. I say this only half in jest, because the fact is it is simpler for me to delink them than to replace them with higher value pipes. What is more, even those I didn't unlink are low value links that are increasing link density in the lede; their value would only be marginally improved by piping to Geography of Argentina Geography of Brazil, Geography of Paraguay, Geography of Chile and Geography of Peru – readers are already numb to always seeing terms such as country names in blue that the probability of them being clicked on is minimal. --Ohconfucius ^¡digame! 02:16, 5 October 2010 (UTC)

• I take ALP's point. I will cease to mark such edits 'minor'. --Ohconfucius ^¡digame! 02:26, 5 October 2010 (UTC)

Attempts to portray this as a "broad and deep consensus" are, sorry to say, rubbish. In effect, someone did pull a fast one on the community by unilaterally changing the guideline to support a personal viewpoint about delinking two years ago. Said change was then vigorously defended by a small group on this page, and it is being implemented through an aggressive (and often careless) delinking script and a list of so-called "common terms" that allows no opportunity for input from the community. If the small group of people - and yes, it really is a very small group - who are responsible for the most problematic behaviour would simply demonstrate a willingness to compromise, then we might actually be able to move forward. No-one is questioning their concerns about blatant overlinking. The problem lies in the attitude that only they can decide what warrants a link. They are using the current wording, language that invites due care and attention when delinking, in a manner that demonstrates neither. --Ckatz^chat_spy 09:21, 5 October 2010 (UTC)

Let me get this right: you've just been running around linking "Australia" in sentences such as this (I've underlined the item at issue):

"'''Perth''' is the [[List of Australian capital cities|capital]] and largest city of the Australian [[States and territories of Australia|state]] of [[Western Australia]] and the [[List of cities in Australia by population|fourth most populous city]] in Australia."

(1) This opening sentence is already fairly crowded with blue, so the reader is presented with an awful lot of choice within three seconds of starting the topic. (2) Why would we want to add yet another link to the most general article thinkable (unless Southern Hemisphere), when there is already a plethora of more specific links to the topic in terms of Australia? (3) Are you attempting to provide for readers who don't really want to read about Perth, but are more interested in wandering through a fairly unfocused, even random pathway, in en.WP? If so, it is at the expense of the wikilinking system as a whole—or at least in the specific place—where the link density is so high that readers are likely to ignore the whole facility and read on (which is why the article was written, actually). (4) What is it about the general article on Australia that the reader would so urgently seek to know, that isn't in the current article on Perth? I note that the reader hasn't gone to Australia to start with: clearly, they are interested in more specific information. It's not Tajikistan, either; even second-language speakers who consult Perth are likely to know where Australia is, and more. Let's not treat the readers as dummies, and debase the wikilinking system at the same time. Tony (talk) 09:46, 5 October 2010 (UTC)

Not "running around linking", Tony, but rather fixing yet more mistakes made by script-based delinking. As mentioned over and over again, it is not the real "common term" delinking that is the problem, it is the insistence by a very small group that they know best. There's never even been a response to repeated requests for an explanation as to why the list should be secreted away, with no opportunity for the average editor to have any meaningful input. There are countless examples of delinking that lack any apparent rational thought, such as the removal of neighbouring states in the lead of the geographical article Bolivia. There has never been any consensus for the heavy-handed approach that this small group is using, as you are well aware. Again, this nasty affair would be much easier to resolve if only there were some willingness to compromise, rather than the current climate of "damn the torpedoes, full speed ahead". --Ckatz^chat_spy 10:18, 5 October 2010 (UTC)

• Effective linking is not about doing it in a scattergun fashion. Does the example Tony cited above look like a mistake of script-based delinking? To me, it looks like a very carefully crafted, if overly so, attempt at giving four links which could be of interest to the reader should (s)he choose to click upon any of them, rather than the unfocussed 'Australia' article you appear to insist on linking to. --Ohconfucius ^¡digame! 10:32, 5 October 2010 (UTC)

Ckatz, you have not yet responded to my question: why did you insist on linking "Australia/n" in that sentence? It seems to degrade rather than improve the linking. I note also that "Australian" is jammed up with "state", yet they are two different links; the style guide has long before I ever posted here warned against such deceptive bunches. Also, could you be more careful when you add links, to avoid such glitches as [[ Western Australia]]? Tony (talk)

• Ckatz, quoting you: no opportunity for the average editor to have any meaningful input. Tony is actually rather remarkable for a Wikipedian. Considering how exceedingly active he is, and for as long, and how he hasn’t ever shied away from controversial issues, I would have expected him to have a half dozen blocks by now if he were as close-minded as you make him out to be. In reality, he has two blocks, and they were both in error, by admins who had been hoodwinked by Daffy Ducks who ran somewhere, jumping up and down, screaming “Shoot him – SHOOT him,” hoping some Elmer Fudd of an admin would blow Tony’s beak off. After a few hours, the admins were directed to proof that there had indeed been a proper consensus (or ruling, whatever), and the blocks were immediately reversed.

  Tony knows what he is doing and is reasonable enough. If you have a concern, bring it up on his talk page. If you two can’t come to an understanding there, don’t jump to the conclusion that Tony’s full of shit and needs his beak blown onto the top of his head again; just bring the concern here. And if you don’t see a crowd flocking to your way of thinking, then just shrug it off as one of those *let’s celebrate diversity*-things. Greg L (talk) 12:42, 5 October 2010 (UTC)

Well there's been quite some text since I last looked in. Trying to make sense of it and offering some rejoinders, in one place rather than scattering them throughout ..

@Art: I saw your answers at the time, but don't see how they answered the questions. Where is the consensus to link regardless of context? People are doing it, as my examples showed. You - despite saying people don't say it - say it yourself when you assert that "overlink says to unlink major countries". It doesn't, not without qualification at least. If Peru is relevant to Bolivia and worth linking, so is Argentina. Sorry, end of story by common sense editing principles or by wp:overlink as read properly.

@Greg 1: I agree with your points about the Sucralose article. Those places are mentioned in passing in an article unrelated to geography - I don't think it's wrong as such to link them, but I'd never argue with someone who took them out. My issues don't relate to those sorts of links, as I would have thought is clear.

@HWV/Greg 2: my point - as, again, should be clear - is as much that actions such as the ones I cited are in breach of overlink as written, especially its "relevance" qualifications. I'm not taking issue with any past changes to the guidelines, although I appreciate that is an issue - it's just not one that I've ever looked into or commented on. I don't know where you got that idea from Greg, or why you feel the need to then rave at me for it with words like "ludicrous" and "puuhleeze". What I am asking is where is the consensus to go even further than those - evidently disputed, at least by some - changes seem to allow. Nor am I going around adding links, as some kind of mirror to those going around removing them. People aren't delinking things I've added, they're delinking things 100s of other editors have added. And I'm sure that no one has re-linked in example 2, or often elsewhere, because no one ever looks at the page. And most of those who do wouldn't have an opinion one way or the other. Plus there's the broader point that many of these delinking edits are made to look semi-official.

@OhConfucius: I know you tend to do these things manually rather than by script, however, often the results exhibit the same issues in my view. One problem is that you spend a lot of time editing here, and, as you say, go through lots of articles one after the other, seeing the same terms linked over and over, which prompts you to think they are not worth linking. However, is it not worth bearing in mind that this is a very artificial way of looking at the site? Many or even most readers, especially those who do not edit, are as likely dip in and out of WP and have a very different perspective. The links that appear frequently are often those to the most viewed pages (eg United States). Per wp:link, we should be providing navigational aids to those pages where they are related and relevant. If I may, your comment about readers being "numb" to links is a) conjecture, as are a lot of claims about linking that are made; and b) probably says more about how regular busy editors see links rather than casual readers, if I may indulge in some conjecture myself here.

@Tony (and Art/HWV): "town" and "city" are related and relevant terms when looking at mayorships. For some people, they no doubt will aid understanding, in terms of providing details about the different status that different urban areas can acquire. And, as pointed out previously, that is not the sole criteria for linking anyway - see also the words "links bind the project together" and ".. help the reader find related information" and "navigability". Meanwhile, I've never understood the relevance of "WP is not a dictionary" to this issue, and I still don't.

No one has answered the two simple questions I asked clearly and concisely. Views seem 50-50 split on this page about issues with delinking (especially if you focus on the more occasional passers-by), and most times I check in on delinking editors talk pages, the comments that I see about delinking are 95-5 queries or complaints. That seems very far from consensus to me. N-HH talk/edits 13:18, 5 October 2010 (UTC)

Point of fact, Greg - I once went to Tony's talk page, for the first time ever, and he simply deleted my post without reading it, even though it was raising legitimate concerns of this sort and also directing him to outright delinking errors (ie those that buggered up formatting). That led, by a slightly tortuous route, which saw Tony shouting at others on their talk pages, to one of his blocks, which was not reverted once admins realised there was "consensus" or that he was right. I'm very over it, and discussing this is very off-topic and needless personalisation, but if we going to have hagiography, we can also have fact-checking. Cheers, and apologies to Tony. N-HH talk/edits 13:26, 5 October 2010 (UTC)

• I see. Tony deleted your post and didn’t give you the time of day. There is no chance your arguments didn’t reflect the consensus of the community and you were being tendentious beyond all comprehension, is there? Tony was simply being mean and bad and unreasonable and all that, right? Your behavior down in this thread isn’t so bad. So let’s cut to the chase without the flowery oratory of how “Links preserve Truth, Justice, and the Wikipedian Way.” Why not post a green‑div below with a freaky-concise nucleus of a guideline you’d like to see. Or is this about how you fully agree with the current guidelines, but think Tony’s activities aren’t sufficiently adhering to them? Greg L (talk) 13:48, 5 October 2010 (UTC)

Sorry, the consensus of the community is that editors should wreck category lists and other formatting - ie introduce uncontroversial and outright errors to pages - and ignore attempts to bring that to their attention? That's what happened on that occasion. What on earth it has to do with being "tendentious" is beyond me. I suggest you go and look back at that issue, as well as the posts Tony left on various editors talk pages after he felt they had crossed him at ANI before rushing to judgment. Also, I thought your point was that Tony very much would give people the light of day if they went to his talk page, which you now seem to be contradicting. As for now, on the more subjective issues, yes, I believe a lot of the delinking - by Tony and two or three others - goes further than the current version of the guidelines suggests it should. I have said that five times now in this thread alone. Thanks for getting close to working that out, although, amid the overblown rhetoric, you still seem to be phrasing your understanding as a question. N-HH talk/edits 13:57, 5 October 2010 (UTC)

Answering my own section: Now you say no one has answered you "clearly and concisely". Much better than just pretending nobody answered. I haven't seen anyone argue for delinking "regardless of context"; show me a diff if I'm wrong. Instead, we're arguing about which contexts are sufficient for delinking. As for my comment on Bolivia, you're right; as GregL put it, I missed the point. Art LaPella (talk) 22:23, 5 October 2010 (UTC)

Sorry, we're going round in circles here - I've shown examples of delinking what are arguably "common terms" where is it self-evidently being done regardless of context, eg the removal of a link to Argentina in the article of a neighbouring country. Here's another - the removal of a link to "Australia" from an article which even has Australia in the title. You yourself said, as I pointed out, that "major countries" should be unlinked [full stop]. That's the very definition of an assertion that disregards the possibility of taking decisions according to the context in which those links are found. Equally, I can't recall any response where anyone on your "side" has offered a discussion on specific contexts or principles, or when they've given an explanation of when they would link such terms (possibly because they never would, QED) - although I can't provide links to a negative of course. By contrast, I've given examples of when, in my view, they should and when they shouldn't. Not that I assume I'm right, but at least I'm engaging with the issue and offering one opinion for discussion. N-HH talk/edits 13:35, 6 October 2010 (UTC)

OK, I said "OVERLINK says to unlink major countries", and I didn't say "except when particularly relevant to the topic". That was for the same reason I didn't mention religions, languages, and quote the rest of the rule verbatim. I didn't mean that was the full text of OVERLINK. Of course I know about the "particularly relevant" clause, even though I somehow didn't connect it to Bolivia; this has been a long debate. More importantly, I have occasionally undone major country unlinks that my own AWB software proposed to me, as part of the manual review AWB recommends. That is, I don't unlink all major countries in practice, regardless of how you interpret my words. And the link I'm pretty sure you can't provide is anyone saying anything like "Let's delink all major countries, regardless of context; the 'particularly relevant' clause doesn't matter." Art LaPella (talk) 20:01, 6 October 2010 (UTC)

Consensus/common terms - arbitrary break #1

• This has gone way off track, it personalises in a way that is offensive to me, and most editors, fortunately will disregard it as WP:TLDR. One editor has already left in disgust (see section above). What is this emotive stuff doing on a style-guide talk page? I cannot participate here if it's going to be a hate page. Tony (talk) 14:19, 5 October 2010 (UTC)

As noted, I agree. I did not wish to see comments about "daffy ducks" and somewhat inaccurate - by my experience at least - claims about talk page welcomes go unnoticed. I should have done, as I should have let go further, bizarre, claims that on that occasion I was acting "against consensus" by raising the issue of outright and continuing script errors. Unfortunately that dragged your name into it. N-HH talk/edits 14:56, 5 October 2010 (UTC)

• Very well, N-HH. Unlike Michael C. Price, who doesn’t like the current guidelines, your issue is simply a content dispute over how the De-Link Jihadists were overzealous. Let’s start out really simple. Please provide just one edit‑dif that has gone full-circle with the other editor (that is, the edit has been clearly communicated to the opposing editor and his views were clear and he stuck to his guns) and which was not resolved to your satisfaction. I assume you are going to chose a humdinger where the link is important for the reader to better understand the subject matter and this offending editor won’t budge. I wanna see that one. Greg L (talk) 14:25, 5 October 2010 (UTC)

I see plenty of fairly pointless link additions. I also see plenty of pointless link removals, as well as quite a few link removals that are, in my view, outright detrimental and that go far beyond the requirement of wp:overlink. I don't comment on or revert each and every one of them, in fact I hardly do it at all. I did though provide two examples above (for one of them, the response from the editor in question was to go back to that page and remove even more links to neighbouring countries, before someone else, correctly, restored them). What are you asking for - evidence of an edit war over linking that I've lost? Or of interminable linking debates on individual article talk pages, which rightly irritates other editors? Why should I - or anyone - have to review each edit and each link removal, or even get involved at all in individual cases? The onus is of course on those trawling hundreds of pages removing links to think about and justify what they do on each occasion, not for others to follow them round page by page and delink by delink. All I am asking for, at a central discussion point, is the general application of a bit of common sense and for people to actually follow the guidelines they cite. N-HH talk/edits 14:56, 5 October 2010 (UTC)

• I gave you an opportunity to be specific so the community could see if we have a chronic edit-dispute problem because of incorrect interpretation of the guidelines by Tony, or by you. Rather than point to a single edit-diff that had been discussed with Tony and was unresolved, all you come back with is a metric butt-load of weapons-grade WP:IDON'TLIKEIT.

  Now, I happen to be in the middle of an medical trial surrounded by people with Ph.D.s wearing white lab coats as we get some pills fired up that transmit pH, temperature, and pressure from inside the gastrointestinal tract. Any time there is a break, I was coming here. And this is what you come back with?? Greg L (talk) 15:19, 5 October 2010 (UTC)

"What are you asking for - evidence of an edit war over linking that I've lost? Or of interminable linking debates on individual article talk pages, which rightly irritates other editors?" Well it seems that you are asking for that. I don't have any, fortunately. Why that should be a criticism, I don't know. As noted, I gave two examples above of pages that I have not touched, but where the delinking seemed crazy and in breach of guidelines. They were intended, as examples are, to illustrate the wider problem. I have also highlighted the specific points within wp:overlink - eg about related info, navigability - that are being ignored in a lot of the editing that goes on. Nothing to do with "I don't like it", which seems to apply more to others' wholly arbitrary attitude to links to certain terms. I'm not sure what your experiment has to do with anything either. N-HH talk/edits 15:39, 5 October 2010 (UTC)

One of the central problems here is that there is a real pattern of behaviour that fails to meet what one would expect in a civilized discussion. We're all adults (at least I'm presuming most of us who bother to participate here are) so it is not a playground-argument "oh, he's not playing nice today" situation. However, the overwhelming attitude presented here by certain people is one of "we're right, you're wrong" - which is completely out of sync with Wikipedia's standards. It is ridiculous that we have this situation where editors who dare to question the scope of the delinking are inevitably dismissed, slapped down, or quite simply ignored. You ask for evidence of drawn-out disputes over links, but the simple fact is that they do not exist because it would be counterproductive to engage in such actions. Speaking for myself, I've got too much to do in real life - and far more things I'd rather be doing when online here - than to patrol the thousands of edits Tony, OC etc are doing. At best, it is work enough to try to catch the mistakes and "overdelinking" that pop up on one's watchlist. None of the editors - N-HH, myself, the others who have added their comments in recent months - are pushing to "link everything" in a "sea of blue", despite the outlandish claims that are used repeatedly in an attempt to shut down debate. What people are asking is for a more measured, reasonable approach, minus the obvious personal agenda. This is a guideline discussion page, so we are trying to keep it from being personal. That aside, I think it is clear (through actions and attitudes demonstrated here and on other Wikimedia projects) that there is a definite disconnect between what the guideline calls for and what certain people interpreting the guideline are calling for. If we can address that problem, and achieve a more consensus-based approach, then perhaps this matter can finally be put behind us. --Ckatz^chat_spy 16:38, 5 October 2010 (UTC)

Having watched this debate, and while favoring the minimal approach taken by Tony et al for linking, I have to agree with Ckatz that this discussion is far from civilized - this is almost the date-delinking stuff all over again.

What we need is some type of pointers to past discuss that includes more than just the regular MOS members (as such would come from an RFC or the like) to show there is wide-scale consensus for minimal linking. There may be something already for this; I dunno. If the discussion of limiting links as being discussed above has only taken place here with no wider feedback, it is not necessarily a reflection of global consensus, which is what I believe N-HH is looking for. It might be a true reflection, it may be out in left field. There is fair argument that Tony, Greg and others provide that there is a lack of edit warring over link removals, but it's impossible to prove that the approach is accepted by the absence of dissent for it. The only really good way to assert that is through a larger call for input just like with date delinking, even if it is just a straw poll. --MASEM (t) 17:25, 5 October 2010 (UTC)

For info, here's a recent section from another discussion page, which I think highlights the usual reaction of most disinterested and uninvolved editors when they come across this sort of thing. I've seen many similar comments on other talk pages, although usually from individual editors rather than a group of them. Now maybe some of the editors there will be won round, but it's a further indication that consensus across the wider site - not to mention the guidelines themselves as currently drafted - is not with any over-zealous delinking of every single link to "common terms", regardless of context; equally that most people will broadly disagree with it, but will also more or less shrug their shoulders and not make as big a deal of it as perhaps I and a couple of others have. That is, do not take the relatively small number of editors commenting at length against the more radical delinking on this talk page - just as the numbers in favour are also small - as proving consent. It does not, by itself. N-HH talk/edits 17:31, 5 October 2010 (UTC)

Your spots a clear, N-HH, when you attempt to mischaracterize the nature of the dispute as one of the relatively small number of editors commenting at length against the more radical delinking. Radical delinking??? The link you provided shows only that some editors were unfamiliar with the linking guidelines with regard to how a common place name like “British Isles” should not be linked. With 47,454,107 registered Wikipedians, there is no way—even via e-mail if it were desired—to educate them to proper linking practices. Quite understandably, there are going to be dozens of editors dropping a note wondering “Where’d my precious United States link go???”

Once again, N-HH, it is becoming increasingly clear that you keep on insisting that your problem with Tony’s edits is that he is radically departing from the guidelines but the real truth is that you clearly and simply disagree with the guideline and therefore would like to put a stop to a prolific editor who edits precisely as the guidelines require. That much is clear from your rhetorical question, above, about “Where’s the consensus?” I can’t help it if you weren’t active when the consensus was arrived at. Your railing against Tony for properly implementing the guidelines is a pure and simple case of WP:IDON'TLIKEIT and your personal attacks against Tony for implementing are a violation of WP:NPA. Period.

Oh, and Ckatz, please don’t ride in all tall, proud, and handsome on your whiter-than-white steed, it’s nostrils flaring in the morning air, and direct a post to me about “civilized discussion”; I haven’t done a personal attack against anyone here other than to point out the shortcomings in their logic. N-HH has been making outrageous personal attacks (via mischaracterizations of fact) against Tony no end here on this page. So your selecting me out for a facefull of humble pie on the “civilized discussion” front is laughably partisan. We all know where you stand on delinking. If you don’t want an unsound and non-factual post exposed for what it is in public, try leaving a sound and factual one here. Greg L (talk) 21:08, 5 October 2010 (UTC)

Greg, please actually read my post. If you had, it would be clear to you that I'm not singling you out for the aforementioned behaviour (although it would certainly be better for all if you avoided the snarky insults). My post doesn't even suggest anything about personal attacks, let alone accusing you personally. You made a post, I responded and rebutted. Deal with it, and perhaps please actually address the points raised. If you disagree with what I wrote, then explain why. Otherwise, claims of "unsound" and "non-factual" are meaningless distractions from the real issue. --Ckatz^chat_spy 21:25, 5 October 2010 (UTC)

I read your post. It begins with this: Greg, one of the central problems is that there is a real pattern of behaviour that fails to meet what one would expect in a civilized discussion. It’s a classic technique to suggest the problem lies with the individual being addressed. I suggest you go back up and strike “Greg” if you don’t want to “single me out.” Greg L (talk) 21:29, 5 October 2010 (UTC)

P.S. Oh, and the bit about your stunt of riding in all tall, proud and handsome on your steed was spot on, as was the bit about your being quite partisan on this issue. So I’d appreciate that you dismount from said stallion, sheath your saber, and spell out for the record if you think Tony delinking “British Isles” and “America” (like the average reader just fell off the turnip truck) is contrary to the guideline, or if like N-HH, you just don’t like the guideline. Greg L (talk) 21:37, 5 October 2010 (UTC)

You can interpret the use of your name any way you like, I suppose. For the record, though, your current guess is completely incorrect, however. My post followed from several comments you'd made, and was directed to you because it felt like the natural person to speak directly to. That, and only that, is why your name was used. Other than that, the post does not even mention you, let alone make any direct accusations against you. However, if it eases your concerns, I'm certainly amenable to removing it. --Ckatz^chat_spy 21:41, 5 October 2010 (UTC)

As for your cutesy insults, again, it would help a lot if you could just avoid them. They add nothing at all to the debate, and serve only to muddy the waters. Enough about that, though. Yes, Tony (and OC and a few others) are taking far too much of a hard-line approach with the guideline; it does not endorse the militant manner in which they've chosen to operate. Many of their edits do serve some purpose; we don't necessarily need so many links to the actor article, after all. However, there is no good reason whatsoever to delink United States in a geographical article about a neighbouring country; there is no need to use a scattergun approach and selectively delink some countries while leaving others linked in the same sentence; and there is no certainly no reason to presume that their opinion as to what is a useful link is the only opinion that matters. It's a classic game that is being played; "our way is right, you're just wrong if you don't get it". Anyone who objects is misguided or simple; apparently, the entire French Wikipedia is a lost cause because they didn't accept the shiny new Gospel of Delinking, at least according to some here. That attitude has to go. --Ckatz^chat_spy 21:52, 5 October 2010 (UTC)

Hmmm. Your real message point gets lost through all your high-road posturing about “cutesy insults” and “muddy waters.” It is not an insult to expose fallacious arguments for what they are. And with regard to *muddying the waters*, you stole a page from my playbook because of N-HH constantly railing about Tony’s flagrant violations of the guidelines when all the evidence he provides shows Tony is editing precisely with the guidelines. In fact, N-HH is clearly opposed to the guidelines. Nothing more. Nothing less. That much is infinitely clear. So please don’t hop right atop your whiter-than-white stallion in a transparent attempt to seize the moral high ground here by talking about muddying the waters. I want to be exceedingly clear here with what your issue is with the current state of affairs. Is it because…

1. Some of the *delinkers* edits don’t adhere to the guideline.
2. The guideline sucks in your opinion.
3. The guideline clearly got here somehow but did not enjoy a true consensus in the process.
4. You hoped the guideline wouldn’t really matter but the delinkers prolific editing using a script to delink no-added-value words like “American“ exceeds your ability to add links to articles and you don’t like that.
5. All the above.
6. Other (succinctly specify please).

Once we have a clear understanding of precisely what each editor thinks the issues are, then we can logically address the issues. Greg L (talk) 22:08, 5 October 2010 (UTC)

Oh, with regard to That attitude has to go, I assume you are once again not directing that towards me or any one particular partisan side to this debate? Because your post seems to not be very clear that you disapprove of how N-HH has been in breach of WP:NPA by mischaracterizing the facts surrounding what Tony supposedly has been doing and he had to strike a couple paragraphs above to bring his behavior here in compliance with Wikipedia’s rules. Yet your vitriol about conduct here seems to somehow not be directed at him. Was that just an unfortunate accident on your part? Greg L (talk) 22:20, 5 October 2010 (UTC)

From a discussion standpoint, no one's attitude is above reproach right now from what I can see. And if it keeps up, history will repeat itself, likely with another Arbcom case over this. It can be avoided, let's assume both sides are working good faith to understand where this claimed "consensus" for limited linking came from, and not worry about who is right and wrong right now. And I'm going to point out again, I strongly favor the limited version (I still am working best how to insert my "tourist agency" analogy into this discussion), so this is not a matter of shutting out N-HH, Ckatz or others that question that approach, but as a means to make sure that WP-wide policy is not being set by only 5-6 people, the regulars on this page. There is a fair question of "where did this consensus come from" that either can be easily answered by pointing to a talk page discussion from the past, or otherwise we've just discovered that this just happened to gel without any real discussion - which certainly can happen but the calls of concern suggest we may want to revisit and clearly establish consensus on it. Linking, when and how much, are critical issues so it is better to caution on checking what the community wants than to play dictatorial games with one's personal opinion of how they should be handled. And just like date delinking - there is no rush to answer these questions; no one is about to sue WP for having a sea of blue or for methodological removal of links of common countries. --MASEM (t) 00:24, 6 October 2010 (UTC)

• There still seems to be too much impressive oratory going on and insufficient laying out the cards as to where different editors’ opinions really lie. I would very much liked to have seen a succinct and honest declaration as to whether it was #1–6, above. But such a straightforward question is met only with stretch of silence followed by still more grand oratory by another editor who seems to want to be perceived as a middle-of-the-roader in search of the Truth.

  Reading between the lines, it appears Masem is suggesting his position is #3 (the guideline is there but somehow managed to get there without a proper consensus). But others here seem chronically incapable of flat laying their cards on the table and prefer instead to berate other editors for editing against the guidelines. Upon inspection, the *evidence* they offer for other editors’ editing contrary to the guidelines proves hollow and it becomes exceedingly clear that the problem is they simply disagree with the guidelines. So…

  I am profoundly disillusioned with the lack of candor and the tactics being used. Some editors seem to prefer evading discussing the central issues and instead specialize in criticizing other editors’ conduct—and quite selectively so in a highly biased fashion. It is a cheap tactic, useful for when things go to ANI where volunteer admins with limited time have to sort things out. This place reeks of bad cess and ill will. I want absolutely nothing more to do with this. Don’t leave messages on my talk page about it. Wikipedia is supposed to be a fun hobby and I have free will and the right to say I want nothing more to do with this when things are like this. Goodbye. Greg L (talk) 01:49, 6 October 2010 (UTC)

I would consider this overreaction to the simple starting point of working on good faith. I don't challenge this MOS at all, but realize that the basis of the complaints about it, is that no one that is defending this is pointing out where the discussion occurred about this. I would if I knew where to look; I only started following this MOS relatively recently where the language of minimal links was already present. A casual search of the archives don't clearly bring up results. It should not be that hard, if this was discussed, to point to that. If it wasn't discussed but simply mutually agreed on without comment, it's probably worth discussing. Very simple, straightforward, and the path of less aggression. --MASEM (t) 02:03, 6 October 2010 (UTC)

The delinkers have been asked repeatedly where the consensus was established. There was no consensus, just a rewrite of the MoS and a refusal to allow any changes.

Oh, and BTW, #1-3, Greg.

--Michael C. Price ^talk 02:44, 6 October 2010 (UTC)

Since my AWB edits make me a "delinker", I wasn't here when the guideline was written, so this post is just to say "I heard you." Art LaPella (talk) 04:35, 6 October 2010 (UTC)

The "where is the consensus" question was asked and answered above.  HWV258.  07:49, 6 October 2010 (UTC)

Since this a claim of some importance you'd better be explicit. If you refer to the claim that the MoS is the consensus, that is of course tautologous and hence invalid. --Michael C. Price ^talk 08:45, 6 October 2010 (UTC)

Could you please link to where, in this thread or elsewhere, there was consensus to delink all "common terms", regardless of context. Despite the fact that wp:link explicitly mandates links to relevant common terms. And also says that links should help readers "find relevant information" and should "bind the project together". Or where, and on what basis, it is being decided that some terms are "common". These are the key questions that have been asked over and over, and just ignored I'm afraid. I would go straight to ArbCom with this, but that's an expensive waste of time and I know where it ends up. And Greg, to be quite frank, everything you've posted since I last looked in is fairly base invective or an utter falsehood - eg that I am "clearly opposed to the guidelines" and that I am making "outrageous personal attacks" - which shows you have not understood a word I've said, even though I've been very explicit, clear and detailed in what I've written. Repeating it all yet again obviously won't help. I'm used to seing my position completely mischaracterised and twisted, but it's still boring. Please stop lashing out with wild claims like that. By contrast, I have certainly not mischaracterised Tony or his actions. One error I will admit - and have already accepted - is that I was sidetracked into commenting on the ANI issue and the events that led to it (which simply corrected some misleading claims in respect of it), but I have struck my comments there. N-HH talk/edits 13:45, 6 October 2010 (UTC)

ps: by contrast, Greg, you have left for example your comment standing that I was being "tendentious beyond all comprehension" for, yes, deigning to attempt to point out a trail of formatting errors a couple of months ago on another editor's talk page. I struck out my bid to explain the facts of what happened in response to that egregious personal attack as a gesture and because it was an irrelevance to the point at hand, not in order to "comply with the rules"

pps: in your numbered list, I am at point 1. I have no opinion on the changes to the guidelines, because I was not involved in that issue and have not looked into it now. I'm just reading the guidelines as they are. If the changes didn't, on a true reading, go as far as some people had intended them to or assumed they had, it's too late now

Past discussions and clarity?

Having not been involved in the debates over changes to the wording of the guidelines a way back, as noted above, I did some research, which sheds some light on the current debate about common terms, especially geographic ones, and on the broader issues of "consensus". What I found was this change back in 2008, where a very specific list of deprecated or even proscribed links was added, along with the unqualified suggestion that they should pretty much never be linked at all. It was expanded by the same editor here. It now seems that some people are interpreting the guidelines as if that text, or some version of it, were still there. It is, of course, not there anymore. It was temporarily struck with a request for discussion - there appears to be have been no detailed discussion or consensus before it was added, other than the inconclusive debate here - with this edit and has not been seen since as far as I can tell. Nor, as far as I can tell, has any discussion about the contents of such a specific list, and certainly no consensus on it. The only discussion post-removal I could find was this one, that discussed the issue more generally and was split pretty evenly numbers-wise, and which seems to have led to the broad wording that we more or less have currently, subject to any later minor changes. The current wording of course has no proscribed list and also clearly excludes common terms that are "relevant to the topic of the article" from any non- or de-linking requirement.

All that might explain why both "sides" are arguing that the guidelines support their case: one is reading them as they found them as of a couple of months ago, the other is reading them as they wish them to be and as they tried - unsuccesfully - to make them two years ago. It also explains why it's been so hard to get straight and simple answers to the questions I asked above: the record indicates that despite vague assertions to the contrary there is no such consensus to delink all common terms, regardless of context, or that there is any consensus as to which terms are indeed "common" in the first place. If the links I found above had showed that consensus, wouldn't someone have rushed to point me and others to them? Now, of course anyone can continue to argue the toss over what the content of guidelines should be - anyone and everyone, whichever side they stand on and whether you want slightly more or slighty less delinking, is entitled to do that. It's not a criticism surely to accuse someone of disagreeing with them as they stand. As it happens, despite claims by others who presume to know others' minds better, I'm broadly happy with them; others are not so keen on every aspect. But, for now, we have to deal with them as they are and with the apparent consensus that got them there. If I've missed anything in the historical record or in my observations on it, I am happy to be corrected. N-HH talk/edits 18:01, 6 October 2010 (UTC)

Thanks for taking the time to track down those links, N-HH. Here's an interesting quote from the 2008 discussion:

"We don't want to encourage robo-delinking. Judgment is necessary to determine when almost any term should or should not be linked." (Spacepotato)

--Ckatz^chat_spy 19:08, 6 October 2010 (UTC)

Nobody is "interpreting the guidelines as if that [list of countries], or some version of it, were still there". I have a list of countries, religions etc. in my AWB software, but I have never claimed that list is a guideline; it's just my own interpretation of the words "major countries ...". OhConfucius uses a longer list. That is to be expected when the guideline leaves the list undefined. Of course there is no consensus on what should be in the list; if that oft-repeated explanation isn't a "straight and simple answer", I can't imagine what would be.

As for robo-delinking, AWB allows and strongly encourages manual review of each change. If my manual review isn't considered sufficient judgment, that won't make me meditate for hours on each link. It would make me remove the unlinking section from my software, and assuming others do the same, OVERLINK would become as obscure and pointless as much of the rest of the Manual of Style.

None of this is to argue for or against a rewording of OVERLINK, or even whether the existing "particularly relevant" clause should be interpreted more broadly. Art LaPella (talk) 20:35, 6 October 2010 (UTC)

Concerning the first two posts here: whoa, that's ancient history. Please check out the more recent encyclopedia of negotiations that made possible the merger of wp:context, wp:btw with MOSLINK. Tony (talk) 04:47, 7 October 2010 (UTC)

Hardly "ancient history", given that they are directly relevant to the principles involved in this dispute (and also central to the discussions you've highlighted). When you factor in that many of the same participants from those discussions are still here today, it emphasizes the importance of reviewing and understanding the entire history rather than just a selective portion of it. --Ckatz^chat_spy 05:10, 7 October 2010 (UTC)

So let me guess: you weren't selecting, but I was? This is not a US election full of spin. I've provided a diff of negotiations that went on over three months, early 09, involving many editors. You are pointing to discussions more than two years ago. We could go back further and ask where the consensus was for widely liking common terms. Could you point to that, please? Going now. Tony (talk) 05:28, 7 October 2010 (UTC)

Tony, what exactly are you trying to prove? You're complaining about references to a unilateral fundamental change in the direction of the guideline from mid-to-late 2008 while simultaneously claiming that discussions from late 2008 to early 2009 to merge two linking guidelines are perfectly fine? --Ckatz^chat_spy 09:41, 7 October 2010 (UTC)

• I'm not claiming that anything is "fine" or "not fine", but that the subsequent extensive negotiations among many editors over a period of months in early 09 settled the wording well and truly. You seem to be claiming that there is no consensus for the current wording. I have limited time to go around in circles here. Tony (talk) 13:34, 7 October 2010 (UTC)
  • Since we are told there is a consensus now, perhaps we don't trust your judgement on whether there was a consensus then. Put our minds at rest and point to the section in the archives, please? --Michael C. Price ^talk 14:01, 7 October 2010 (UTC)

The merger was a hard negotation by many editors, as you can see from the diff I provided. You will note that Kotniski merged the three pages, then there were objections, and they were unmerged. Negotiations continued, and I can't recall whether they were merged and unmerged again, or just merged once more. This process represented a complete overhaul and rewriting of the guide line. You can put your mind at rest by reading the diff. You are welcome to consult User:Kotniski if you wish. He ran the whole thing. To dig up diffs from a year before this claiming there is no consensus for the current page seems to be a little odd.

Please do not personalise the discussion, including the insulting of other editors ("we don't trust your judgement"). Tony (talk) 14:10, 7 October 2010 (UTC)

Thanks for that link to the 2009 discussion Tony. My request was a genuine one, as to whether I'd missed anything. As you know, I am more recent entrant into this dispute. However, I don't see what it actually shows - it's a very long discussion, and as Michael says you haven't highlighted any particular thread within it, or tried to explain what you are arguing it shows. I scanned it quickly, and can't see much specific discussion about the near-blanket exclusion of links to specific terms, let alone any consensus on that. OK, so it's a more recent example of wider discussions that got us, along with the 2008 threads that I pointed to, to where we are now with the overall wording of wp:link. But, as I keep saying, I accept broadly the current wording, which says - as well as that "links bind the project together" ... "include [links that] help the reader find related information" ... "think before removing a link - it may be useful to other readers" - that common and geographic locations can be linked when they are "relevant to the topic of the article". My point is that the current wording does not mandate the removal of pretty much all links to a set list of "major" countries, professions etc. You tried to insert that point in 2008, with support from one other editor as far as I can tell, and you did not get consensus for it. The 2009 discussions do not seem to contradict or supersede that discussion, and the guidelines still do not include that wording.
Hence, it is quite clear that consensus was then - and remains now - against such rules and actions. Yet a small number of editors - in my reading of what is happening, I accept you dispute that Art - argue and delink in thousands of articles across the site regardless as if it did and as if there was such consensus (note: Tony when you comment here, you seem to be conflating the issues of consensus for the guideline wording and consensus in respect of some aspects of delinking practice. The point being made is that there is no consensus for the latter, and that indeed those actions go against the actual consensus as formulated over the years in the guidelines - that's the *whole point* of my argument that opened this subsection. Nor do I read anyone else's comments as them saying this all proves there is "no consensus for the current page", as you suggest they are. Maybe they do mean that, but it's certainly not my point). I also have to say it is noticeable from the 2009 discussion that most of the names who are not familiar to me are asking for less rigidity in delinking and make uncannily similar points to the ones I and others have been making more recently, which gives me some confidence as to the views of the more disinterested editor and of other great minds. N-HH talk/edits 15:22, 7 October 2010 (UTC)

.. and while it is clear that there was - and still is - no consensus for certain aspects of delinking ..

.. people simply sign off from the discussion here with a comment that fails to address the actual issue (as noted above, the issue is primarily about whether there is consensus for certain actions and whether they follow the guidelines, not consensus about the content of the guidelines themselves), and then continue to for example take out links to "Catholic" and "Italy" from a page about an, er, Italian trade union with Catholic roots, marking those edits as "minor" and with an edit summary claiming that they are simply "fixing dashes, unifying date formats and removing date links." N-HH talk/edits 17:22, 7 October 2010 (UTC)

RfC on the above

Per the section and sub-sections above and previous discussions, there is a long-running dispute about certain aspects of delinking, which is often done using scripts.
Question/issue: Does it make sense and is it in accord with wp:link as currently written for every individual link to "major" or "common" terms, countries, places, professions, nationalities etc - however those are to be defined exactly - to be removed from most articles, both main text and infoboxes? And, as a secondary question, is the use of scripts to do that across hundreds of articles appropriate? N-HH talk/edits 17:54, 7 October 2010 (UTC)
Note: The purpose is not to ask for comment on the removal of more obviously trivial, irrelevant or needlessly repetitive links, or links to major/common terms - eg "Italy" - where they are mentioned in passing in an article and have no immediate relation to the page/topic. N-HH talk/edits 17:54, 7 October 2010 (UTC)

Comments/response (no threaded discussion please)

• No. This the push to reduce overlinking is a good idea that has gone too far. Per some examples in the discussions above, wholly relevant links that aid navigation for both casual readers and regular editors, and which bind related and often very popular pages together, are often being removed, seemingly on an arbitrary basis and regardless of context and of the relevance of the terms to the main topic. This reduces the core functionality of an online encyclopedia, and introduces random inconsistencies, especially to infoboxes and tables. The guidelines are to me quite clear that where a term is "relevant to the topic of the article", there is no requirement to remove a link to it. Scripts are fine as a tool, but they are a blunt tool, and their effects need to be reviewed on each occasion to take account of the context of each page. N-HH talk/edits 17:54, 7 October 2010 (UTC)

• Yes and Yes This RfC was poorly crafted with a two-part question that the poster only answered with one “No.” Moreover, RfCs are ideally crafted by an uninvolved admin; either that, or a better job is done by the poster to achieve unbiased wording. The wording of this RfC was crafted in a slanted manner to advocate an outcome in the form of biased questions that have a rhetorical ring to them. Now…

  The guidelines on linking are well drafted and are a reasonable compromise that best serves our readership. There is always room for interpretation but the editors N-HH has a problem with are doing a good job of adhering to the spirit and word of the guideline. This is just a content dispute over the implementation of guidelines, where it is the guidelines themselves that N-HH has been outspokenly in opposition of. Unfortunately, his views are too extreme and are not shared by the majority of wikipedians. Greg L (talk) 18:23, 7 October 2010 (UTC)

• No and no. Good to see you back, Greg. I never really believed you'd stay away long :-) --Michael C. Price ^talk 18:51, 7 October 2010 (UTC)

• Yes and Yes. Note the question specifies "most articles", not all. A randomly chosen example from Intelligence quotient#Genetic influences: "Various studies find the heritability of IQ between 0.4 and 0.8 in the United States ..." But I didn't find anything in the United States article about IQ. So that link is no more helpful than linking "helpful" would be in this sentence. Sure the IQ sentence is about the United States, but my sentence is about helpfulness; so what? Most any word could be linked by that logic. Another random example: History of Cuba links United States once down in the middle at History of Cuba#Bay of Pigs invasion, apparently because all the other references to the U.S. were unlinked, but they missed that one. A link to U.S. history in the lede would make sense, but not the way it is. Since most such links should be unlinked, scripts are a more efficient way to unlink them, even though some United States links should remain. A misconception is that the alternative to scripts, is careful consideration of each link that will come to the same conclusion you would have come to. No. The alternative to scripts is represented by my examples above. There are plenty of articles I haven't run my scripts on at all, so that would be a more efficient use of time than trying to guess what the consensus here would be for each link, beyond briefly glancing at each, especially in the lede and infoboxes. Art LaPella (talk) 22:58, 7 October 2010 (UTC)

• Yes and Yes. The current wp:Link wording has been in place for a considerable time, and has led to thousands and thousands of edits—the overwhelming majority of which have been well received in article space. Common and dictionary links "to be removed from most articles" is precisely the intention of the hard-working and skilled editors who have undertaken this task—solely with the aim of improving WP for all readers. The use of a script (note: not a bot) is the only practical way to identify and remove links that don't deepen the understand of the article in which they have been placed—as long as the result of running the script is checked manually (which the editors involved have given an undertaking that they do). Linking is a difficult skill to master, and the delinking in question serves two additional purposes: to prune an article to a new baseline (from which local editors can continue to evolve the article), and to remind local editors to carefully consider the value of links they add in the future.  HWV258.  23:06, 7 October 2010 (UTC)
• Yes to both. Indeed, it makes perfect sense, and is in absolute accord with wp:link as currently written for every individual link to "major" or "common" terms, countries, places, professions, nationalities to be removed from most articles. There are many instances where linking to 'relevant' terms fails the 'germane' test. Links are often used by some, ostensibly for navigation, for purposes that can only be described as providing definition. This may be fair enough in specialised articles about technical subjects, but we should indeed draw the line at [not] linking 'Australian actress' in the Nicole Kidman article, 'billionaire American businessman' in the Warren Buffet article, or 'multi-instrumentalist British musician' in the one on Paul McCartney, relevant though the terms may be. The extent of linking in en.wp, although comparing favourably with our sister wikis, can still be improved. I don't quite understand this nebulous concept of 'navigation', and Nick's example of 'Italy' does not shed any light on how this dichotomy can be easily resolved.

  In my experience, there are a great number of non-overlinked articles; there are also a great number of extremely overlinked articles, most of which few editors care about. I have seen articles, where many proper nouns, including most dictionary terms, are linked and more than once. I have come across articles where The New York Times has been linked more than 30 times in the reference section alone! – go figure. At present, there is no way to identify overlinked articles; I wish there was. Removing the overlinking manually, or even by using the standard module of AWB, is painstakingly slow and inefficient. Use of scripts (but not bots) for this purpose is desirable, in line with wp:link, and with WP:BRD. Scripts must be supplemented by further manual effort, as purely mechanical link-removal is often inadequate, and links to some words can only be removed (or reinserted) through interpretation of relevance (read 'germane'); they reduce workload. Also, some articles have links which are so tortuously piped in order to squeeze in an extra link or three that manual unlinking is unavoidable (I believe Tony cited one about an Australian location the other day). Editors objecting to removal of certain of those links are always free to relink those instances they believe jeopardises the understanding of the article. In fact, that should be the acid test, rather than the insistence that where a link is "'relevant to the topic of the article', there is no requirement to remove a link to it". Even this way, we slowly undo the ratchet of overlinking because some editors fail to appreciate the need to balance quality with quantity when linking. --Ohconfucius ^¡digame! 02:20, 8 October 2010 (UTC)

• Yes and yes. Per Ohconfucius above, who puts it well; Masem, below, makes good points, too. Wikis (at least this one) have been evolving more sophisticated approaches to intralinking over the past five years, after the first four years in which no one really thought much about the issue. This has involved embracing more than just "navigation" in the decision-making process: now, the focus and likely usefulness of each link to an English-speaker, and the link-density (with its corollaries of diluting high-value links, adjacent "bunching", and unnecessary reader distraction) are widely accepted as balancing factors. The notion that article targets should be linked just because they are somehow similar to the current topic (e.g, Tent from House, and Dog from Cat, as suggested by Hans A.), works against efforts to optimise wikilinking by blurring its role with that of our Category system (not to mention the "See also" section).

  Nowhere is poor linking practice more obvious than when whole country articles are target-linked without thought. Show me one, and I'll usually suggest a more focused target to a daughter articles or a section—or that it should not be linked in the first place. It is damaging the wikilinking system as a whole to throw such generalised, vague links before the readers in the hope they will generate a click. Readers soon switch off our valuable links when we dish up distractions like that. Past linking practices have also led to carelessness in piping. These are reasons I developed a tutorial on linking skills (av. 80 visits a month, feedback welcome).

  All of this is born out in the wording of the style guide, while at the same time it extols the virtues of the system as a whole. Gnoming activity is essential if linking standards are to be improved, since we have a legacy of past poor practice to deal with. Matching the high standards of linking in featured content, standards that have become well-established (take a look), is a challenge for the whole project. Human-supervised automation is an important part of this, given the scale of the problem. Sometimes mistakes are made (not the examples above, I must say), but overall, this is the right way to go. It would be productive if editors who who have qualms about it engaged those who work to reduce overlinking not WRT obvious examples of overlinking, but WRT the occasional grey area. Skill, insight and hard work are the keys for all parties in collaborating to improve wikilinking system. Tony (talk) 03:32, 8 October 2010 (UTC)

Further comments/threaded discussion

I think RFC is too simplified for an encompassing discussion of linking. It definitely cannot be answered by a simple yes/no answer. We need to establish to exactly what degree interwiki links in prose, and separately in infoboxes, are to be doing for our readers. Only once that question is fairly answered by consensus can one make a determination of when it is appropriate delink the use of common english words. --MASEM (t) 17:59, 7 October 2010 (UTC)

But how would we do that? Surely the point is that different readers use links in many different ways at different times - for basic navigation, to move on to related topics of interest, to read into or "drill down" to a more specialised or technical page etc. I'm not sure we'd ever pin that down, and I think there's a risk of over-intellectualising and over-analysing the whole issue. Also I think there's a risk that people enter that debate saying how they think others ought to be using links, which surely isn't the point of WP unless we're all at school here. And even without any such agreement, I think people can still offer a general judgment or opinion as to whether, broadly, it is OK to have links to "England" or "China" on related pages. Anyway, let's see what happens here. N-HH talk/edits 18:12, 7 October 2010 (UTC)

But it is a very deep issues, not one that can be waived away by a yes/no question. And it affects the entire WP population; if you're calling for an RFC and wide discussion, we need to come to a wide consensus on the nature of linking. Effectively, there is a simple question: how "valuable" is an intrawiki link, being that they stand out from text and require readers to click-thru to learn more, possibly breaking reading flow or the like? If they are highly valuable - we want readers to be able to read w/o distraction, a subdued approach to linking where only essential terms make sense. If they are "cheap" in that it doesn't break flow, they we can allow for linking a lot more common terms. But I don't know the answer on a global scale (I know, however, that I prefer considering them valuable to avoid high link density, but that's just me), and if you're going to RFC anything, getting at the core of the issue makes more sense than a challenge to what some are currently practice; that is, addressing the cause and not the effect of it. --MASEM (t) 18:21, 7 October 2010 (UTC)

My sentiments exactly. The problem with the question, and the whole concept behind it, seems to be the nom's apparent preference for lowest common denominator linking, rather than one which is adapted to the man on the Clapham omnibus. --Ohconfucius ^¡digame! 02:28, 8 October 2010 (UTC)

@Greg: per this comment in the section above. If you have issues with the wording of the RfC, feel free to say exactly where the problem is. I did answer my own supplementary question - which seems a fair one - just without a simple yes or no. And could you provide links, or evidence for the statements in your last sentence? Or, better still, avoid outright falsehoods ("outspokenly opposed to the guidelines"), cheap labels ("extremist") and prejudging an RfC ("not shared by the majority of Wikipedians")? Half your comment engages with the issues, the rest does not. N-HH talk/edits 18:48, 7 October 2010 (UTC)

I believe it would be useful to broaden the scope of the RfC to include issues about repeat links.--Michael C. Price ^talk 18:53, 7 October 2010 (UTC)