Wikipedia talk:Manual of Style/Linking/Archive 14

Archive 10

←

Archive 12

Archive 13

Archive 14

Archive 15

Archive 16

→

Archive 20

Link Lank Lunk

If one thing is more obvious than another, it is that this discussion is going nowhere. The obstructive side is too pig-headed, and the logical, constructive side is too determined and too obviously has right on its side to have any option of caving in. Personally I could have contributed more pearls to the talk page indefinitely, but venerable injunctions forbad. Instead let me point out a few considerations that I hope will be persuasive enough to settle the matter so decisively that everyone will approve. (Fat chance, but one must try, right?)

Look, the fundamental problem is that the negative side doesn't like the appearance of the links. They claim that they (and therefore the users) find links unnecessary, excessive, unaesthetic, unhelpful and distracting. (Tick as many as apply.) They unlink connections to relevant pages on no better basis than their own taste, in opposition to any readers stupid enough to prefer seeing links, and editors idiotic enough to have a more flexible and constructive contrary opinion. They can always counter opposing preferences with flat statement, flat contradiction and personal abuse. The pages look tidier and therefore are more useful. Right? Riiight…

Nothing new so far, right?

So isn't it time we ineffably brilliant, constructive, creative, cooperative Wikipedians explored a new approach? We already have an option (which I for one have embraced with gratification) for inspecting a link by hovering over it. There is a straw in the wind if ever I saw one. All we need do is add a new option with a very similar facility for not showing links in any special way at all. I reckon that a toggle button "show/hide links" at the top of the page would suffice. An invisible link would still behave like a visible link in every other way, so if any user found links distracting, he could simply shut them up.

The user then would only have to move his mouse cursor over a suspect spot to check whether there was a useful link for him to follow. The cursor would immediately change to give a hint, and if he had the floating option active, he would soon afterwards know whether to bother or not. In case the cursor were insufficient, we could (I do not insist on this point) have the colour spring into visibility while the cursor remained over it. Our programming staff should be able to implement something of the type almost on the trot. Thereafter it would hardly matter whether obsessive linkers bracketed every word in an article, French or not. All the nitpicking about linking to place names, languages, repetitive relevances and so on simply would stop mattering. And so they bloody should. Some of us may have time to spare for such nonsense, but I certainly do not.

The rest of us could then stop squabbling and proceed with sweet mutual tolerance and enhanced efficiency. I can well understand some of us worrying that that the resulting atmosphere would strike some of us as cloyingly peaceful, but let me reassure them: plenty of bases for gratifying slanging matches would remain.

Furthermore, on a slightly more demanding basis, we could ask for a new, semi-automatic linking facility to supplement the linking, hidden or not. If implemented, then when anyone selected text, whether linked or unlinked, and then, depending on the details of the implementation, hovers over the selected area or holds down the right click button, or whatever activation suited users and designers best, it would have much the same effect as entering that text into the search box. This would not be a universal substitute for linking, because, though some delink-zealots have not cottoned onto the fact, or not cared, there often is a need for the visibly linked text to differ in name from article or redir names. However, it still would greatly increase the flexibility of the existing linking facility, and without harming any delicate sensibilities.

Now, all that was a quick, halfbaked thumbsuck; any better ideas? Any reactions? JonRichfield (talk) 19:20, 28 April 2012 (UTC)

Anything that is going to be based on a user-preference that impacts the direct display of an article is not going to fly. We used to have date linking that would format dates on a user's preference, but the problem is that this ignored what would happen with non-registered users, and leads to problems with editing articles as they would view them, instead as how the editor sees them. This same logic applies to the linking approach you're talking about. --MASEM (t) 02:26, 29 April 2012 (UTC)

Actually, JonRichfield proposed "a toggle button ... at the top of the page", so registration would be unnecessary. My main problem is that Jon seems to assume delinkers oppose all links, but nobody does. The best argument for delinkers is that there was formerly an expectation that countries would automatically be linked, even if the country is no more likely to be clicked than linking words at random – those links should be unlinked. If delinkers can predict which links are useful to the reader better than linkers give them credit for, then Jon's suggestion makes it impossible to show readers the best links on the page, remembering that readers can always type other words into the search box if they really want to. My solution is that careless over-delinking (if any) should be treated like any other kind of editing against consensus, with more emphasis on specific examples and specific editors, and less on more generalized debate on whether removing a link is more sacrilegious that removing any other kind of content. Art LaPella (talk) 03:54, 29 April 2012 (UTC)

I, for one, am trying to make the linking system stronger, not trying to get rid of it. Jon, please don't characterise editors as wanting to abolish wikilinking. A toggle switch is not the way to go; careful linking is the way to go, and it should appear the same for all viewers—readers and editors. Just one tip: please consider writing about 20% of what you do now, and paragraphing a bit more; most people, like me, won't make it further than the first few sentences. Tony (talk) 06:22, 29 April 2012 (UTC)

Tony, what on Earth gave you the idea that I wanted to abolish linking? Last time you attributed views to me, it was along the lines of: "a link maximiser—someone, like N-HH and CKatz—who'd be happy for wikilinking to be boosted back to levels resembling the old days, when it was virtually "link what you like", without discipline". In the light of such remarks, you will pardon my reserve in acceding to your judgment in such matters. As it happens, I really don't care whether you want to make the system stronger or weaker, more self-effacing, or an Apollyon on the road of the editor. Nor did I characterise editors as wanting to abolish linking, only as opposing the linking or non-linking of every other editor whose linking diverges from their own preference, and justifying their opposition as being "thoughtful" as opposed to the "link what you like without discipline" of the idiots who disagree with their own lucid insights.

You assert that a toggle switch is not the way to go, even if its use or its very discovery by the user is optional; how many users have you polled who have tried it so far? Are you undertaking to dictate to them why or how they should use the system? I hope not! How do you see linking being weakened (whatever that means) by the user not wanting to see blue and red words among black text, when, even if he elects to use the hiding setting, he can easily move his cursor to any word that he is curious about, link to it directly if it is a link, and search for it if it is not? If you really want the best of both worlds, you could have a second generation ternary toggle that hides all but the first link of a given word. (Heaven, heaven! No more arguments about over-linking!) That would strengthen linking dramatically, I reckon.

I find it hard to take your statement seriously that "careful linking is the way to go". On what conceivable basis is anyone to justify his imposition of his personal opinions as "careful", as opposed to the linking of other editors? They must be careless because they use more links than you do? Or because they suggest that optionally a link pops up only when consulted? Banning a word from being linked more than once in an article is "careful"??? Assumption that every reader reads every article top-down, or doesn't deserve consideration in one's writing, is careful? And calling someone who has opposing views a link-maximiser is careful? Do me a favour!

With the possible exceptions of some of my earliest efforts, when material I had written was "corrected" by editors who certainly would not have earned your approval, leading me to assume that I was what we here are calling "under-linking", I defy you to find a single link or multiple link that I had inserted carelessly. You would find plenty that you would have considered redundant, and perhaps some of them you would be right about; for example, I have occasionally inserted a link where a previous editor had linked the same word elsewhere, and sometimes I would not have done so had I noticed it, and sometimes I would have linked anyway, because in my careful opinion it would be helpful. Accidents do happen, and not every apparent accident is accidental. Conversely, not every reasoned action compels every dissenting agent to accept it as the only justifiable way to act. As long as you cannot define a particular version of care as exclusively correct, or even as nearly enough so, as to to be clearly superior to alternatives, in serving every need, you are in no position to urge "careful" linking or editing as being adequate, let alone "the way to go". It isn't working, no matter how you gird at the efforts of other editors less "careful" than you.

The toggle, whether binary or ternary, is not a perfect solution, but it is a cheap and easy way to deal with certain classes of problem. I could propose others, such as a new class of link that never shows except in certain contexts, or for all words that have been linked earlier in the article being non-highlighted, but behaving as links; there are many possibilities. But the toggle as proposed so far should be effective and inoffensive. Why this "weakening of linking" bit?

I am sorry if you cannot make it past the first few sentences, but if you cannot do that, I am left wondering whether you could make it past the first few points. If you did, I wonder how you could have made some of the statements that I have referred to. Better luck in future. Cheers. JonRichfield (talk) 11:23, 29 April 2012 (UTC)

The idea of links only being visible by hovering on them doesn't work on devices without a mouse such as smart phones or tablets (which is also a reason for avoiding WP:EGG links, BTW). (And by the way you can already achieve that by adding a:link, a:visited {color: inherit} or something to your style sheet – either the one in Wikipedia or, if you're not logged in, the one in your browser if there's one.) ― A. di M.​  09:56, 29 April 2012 (UTC)

Errr... Really? Aren't you mixing up link visibility with sample link content? Doesn't the device of that type have a cursor at all? (I don't own one.) If it does, then does its cursor shape not change on hitting a hot button? That would suffice if necessary; in fact, it is what a PC user would get if he had not activated the floating option. If the portable supports neither facility, then is there no matching facility at all that would suffice instead? Usually one would expect something of the type to be available.

Adding stuff to my style sheet to achieve the effect is irrelevant to the problem. I can just see a heading on WP's front page: "Go and copy jonrichfield's style sheet if..." Sorta lacks a certain something, wouldn't you say? I appreciate the importance of avoiding facilities that cannot benefit a significant section of the user population, but at the same time one must be cautious about how stringently one avoids every facility just because someone cannot or will not use it. Until you can assure me that the link hiding is not the only part of the suggestion that would not work, I cannot see any reason to modify what I proposed, all the more so if there are alternatives, as I suspect there might well be.

Your other suggestions do not strike me as relevant. I may well have misunderstood. Perhaps you could elucidate for the benefit of people who do not wish to hand-craft individual links? Thanks. JonRichfield (talk) 11:45, 29 April 2012 (UTC)

Mobile devices like phones and tablets have no cursor, whatsoever. When you touch them, they then react, and often this itself is a point of confusion (you can "tap", "tap and hold", "tap and drag", "tap, hold and drag", etc. and the function varies program to program. By far the simpliest interaction is a simple tap to initiate an action). So no, we can't even consider the popup functionality of use here. It's nice on systems that have it, but we need to recognize we can't build on that feature for all users. --MASEM (t) 20:06, 30 April 2012 (UTC)

Folks, I am unsure what I said to elicit some of those responses. As Art says, the whole point is that this permits the editor (and re-editor) to decide what s/he wants within reason, and independently the user, registered or not, can decide what s/he wants and change minds at the click of a button as often as desired. That is as it should be, isn't it? It also has no correspondence to the problems with the date example. If the user doesn't mind the appearance of links then ethically and practically it makes no sense for the de-linkers to insist that he bloody-well is going to hate them and make him do without, like it or not. Conversely, if he (the reader) does in fact hate them, it is not for the writer/editor to insist that he put up with the distraction and aesthetics. About the only tricky aspect to the vanilla version of the scheme (the one without the automatic searching/linking) is the question of whether the toggle button should be bistable or monostable, and if monostable, in which default state it should start. If we cannot come to a general agreement on that point (guess my opinion on that! ;-( ) then we can add a user option similar to the floating cursor link option. How hard can that be? Whose preferences and opinions would suffer?

Masem speaks in dire tones of "user-preference that impacts the direct display of an article", but firstly, user preferences deserve more serious respect than that implies, and secondly, what I said implied nothing of the kind. As Art pointed out, the proposal is nothing of the type. Ignoring the links is an option that the reader is entitled to, just as much as he is entitled to change the size of the window and font, examples of actions that affect the display a lot more drastically. Hiding a link does nothing to the layout. It is reversible on the fly at any time without impact on any other facility or creation of confusion. Some things we may claim authority to insist on, but not that. And as for non-registered users, the button would be as much at their disposal as anyone else's. To make sure they know the links are there, we either could make the visible link the monostable option, or make the toggle button blink when the display is off, or something.

But one of the key points is that with this approach implemented in some effective form, the investment would be minimal, and the linker/delinker war should be over except for those who are fanatics in the Churchillian sense of "one who can't change his mind and won't change the subject". Why should anyone care about links that they consider unnecessary as long as they don't need to put up with the sight of them? Nor is there any need for link-averse newbies to do without the advantages of linking when by (double?) clicking on any word they can find an available link, whether it is visible or not. It would be either a tailored, explicit link, or the best the search engine could come up with. (And btw, IMO, the floating cursor link option should be the default. Anyone who dislikes it, can go and disable it without any problem at any time after registering.)

And Art, my point is that I don't want to give a damn, nor want anyone else to give a damn, what either linkers or delinkers oppose or approve. This is not a matter that can be or should be settled one way or the other by reason; there are opposing material reasons and for a long time there have been opposing (variably reasonable) interested parties willing, if not to slit each others' throats, at least to waste everybody's time enlarging the flood that rolls hoarser with anguish as the pages scroll. Didn't I get the impression that my contribution struck no one as particularly novel? That this mutual cage-rattling had been going on since gammer Gurton slipped into the pig midden? Was it just my illusion that no one took longer to get bored with the whole mess than I did, but felt trapped in trying to come to some sort of practical, acceptable arrangement? Apart from those who proverbially enjoy wrestling in it? So why not just navigate round the whole sorry mess, letting all participants choose their own styles in tranquil superiority instead of sweating blood trying to force them onto everyone else? It is not as if it would be expensive! Perfect? Of course not! Close enough to justify a compromise... Errr... sorry about the 4-letter c-word. I'll try not to let it happen again.

But Art, as far as I can tell, your position is

• that the de-linkers can do a better job than the linkers can. Let's get real, shall we? Some de-linkers certainly can -- for some readers in some articles, not for all readers of any article. Some of them make a dogs' breakfast of it in most articles. You don't become a competent linker/delinker by feeling strongly about it. And you don't become a lousy editor by preferring more links than some Dunning–Kruger de-linkers with delusions of competence, or fewer links than their opponents.
• that the suggestion makes it impossible to show readers the best links. Not only do I deny that; the reader who wants the links doesn't have to hide them at all. If he does want to hide them, the author no right to insist on jamming his personal opinions up his nose if he finds them objectionable or distracting. Anyway, if he floats he can see any links that he wonders about; even in the most vanilla version the cursor shows when there is a link, and besides it should not be a major schlepp to make a link show up whenever the cursor hits it.
• that readers can get the same effect by typing other words into the search box if they really want to. Come on! Sure they can, but if that would suffice to give the same smooth non-obtrusive assistance that floating links give, why have links at all? Just let them type in everything they want to query. Anyway, isn't this confusing the two kinds of linking -- the (in)visible link and the assistance in linking unlinked text? The latter is just a seamless improvement on searching; the former is everything we have at present -- plus function and minus talk-wars. Think about it.
• that I assume delinkers oppose all links. Well, that is taking my rhetoric a bit too literally. My reaction is to hypersensitive individuals whose skin crawls every time they see a link that they personally do not feel a need to use, or that they imagine that they have seen before in the same article a few pages back (where the same linked word might not even refer to the same material). They might even be right. But that does not justify their foisting their preferences on every reader. Let's face it, most readers pass a lot more links unclicked than they follow, whether they should follow it or not. Let them do the choosing, not the linkers or unlinkers. Any lesser degree of reader freedom makes about as much sense as assuming or trying to insist that they read each article desuper deorsum, which is in every way contrary to the encyclopaedic idiom or function. The reader might be a youngster who does not know enough to absorb much beyond the paragraph he is wading through, or he might be an expert who knows ten times as much as the authors ever will, but happens to be checking on a particular point of view, and doesn't need any links at all. Leser, Leser ueber alles. If we are not writing for the reader, what are we writing for? Vanity? I can primp my own ego more cheaply than doing this kind of work.
• that "careless over-delinking (if any) should be treated like any other kind of editing against consensus". This is flying in the face of the gods of systemantics. One of the first laws of systemantics is that a system works best when it runs downhill. Forcing choices is trying to force the system uphill. Consensus? You are joking, right? What on earth is the point of consensus when it is unnecessary and vocal and reasonable dissenters object to blanket enforcement of rules based on arbitrary opinions? You might prefer X, and you might even be able to justify your choice, but that makes not a scrap of difference when any user is entitled to a contrary opinion and has to put up with your own, personal, ideas of what he should prefer. The costs of the dissent have long since exceeded in labour and goodwill, all the costs of if implementing a facility to let the system run downhill by letting readers vote with their fingertips. The personal assessment of careless over-delinking or over-linking hasn't worked yet; what should give us any hope that (de)linking will suddenly start working now that we have carefully explained it?

I am putting my views where my mouth is. I say that if we implement a display facility as cheap and handy to use as possible, with at least a transient moratorium forbidding fiddling with other folks' links except where there is a gross error, gross omission, gross misdirection, or gross functionless repetition, permitting readers their own preferences; then we shall have less wasted time, labour and spittle, and the articles will look more attractive to both sides of the battle. Perfection? Be serious! But better is better than worse, which is what we have now. JonRichfield (talk) 09:28, 29 April 2012 (UTC)

You're downplaying the impact of user preferences. We've been there before with the date delinking thing, this is not some random concern. Take, for instance, the case where in the sentence "John Smith is an American lexiographer". "Lexiographer" would likely be linked regardless of prefs, so that's ok. But the schemes proposed, "American" would likely be linked with visiblity set by user prefs. This makes a "chain link" for those users which is a layout problem. Trust me, we don't want to be writing two different sets of pages here due to links. --MASEM (t) 15:04, 29 April 2012 (UTC)

Masem, I just don't know where you get the connection between your stated date concern and this proposal. Let us consider two hard examples, and concentrate on them proportionately hard. Let us start from the simpler one first: Yevgvey Tsintnist is a Tungusic lexicographer. This being the first mention of those words in the short article.

That one is easy because you know and I know Yev very well, but we also know some persons who shall here be nameless never have heard of him and certainly will want to look him up as soon as you and I collaborate on an article on such a famous philologist. You and I are also of course fluent in Tungusic, so we need no such link either, but in our generosity we indulge the ignorant with this link too. Against all probability that anyone needs such a link, we also link lexicographer, just as we link any other pentasyllable, such as orangemarmalade or pentasyllable. So far so simple, right? Then here we have seen what the text would look like on the page of anyone who is not pathologically link-averse. Anyone who had (binary) toggled off link display would see just: Yevgvey Tsintnist is a Tungusic lexicographer.

Unlike what we have here however (since we have not yet implemented the facility), if he moved his cursor over any of the linked words, he would see the cursor change in appearance, indicating the presence of a link. This is exactly what we would see here if we moved our cursors over the coloured links. If he had activated the floating interpreter as well, he would see some of the content of the link if he held the cursor there long enough.

OK so far?

Right. Then lets try a harder example: "John Smith is an American lexicographer". No one has heard of JS, so we don't have a link for him, not even a redlink. Nothing has changed with our favourite pentasyllable, so it still is there. But what about "American"? Everyone speaks American nowadays, OK? So why link it? Or alternatively, how dare you not link it; have a little respect! So we have a long-drawn fight about that one, and you don't need me to tell you what the outcome looks like when the reader has not toggled his button. No matter what the outcome of the fight however, everyone who has toggled will see this: "John Smith is an American lexicographer". The only difference is that the toggled version will show which words are linked whenever the cursor passes over.

First let us observe that this has nothing whatsoever in common with your date niggle. It has no effect on the facilities available to the user, except that he has the option of not seeing links (spelt "clutter" in some lexicons) that he doesn't want to see, but can find and use whenever he wants to, even if he does not trouble to toggle ad hoc. It is hardly or not at all sensitive to the resources he has available. If we add the fancier searching option he still can search on "American" without typing it in, whether it has been linked or not.

Now, are we a little closer to a common understanding? If not, please supply an illustrative example of a difficulty. Cheers. JonRichfield (talk) 17:51, 29 April 2012 (UTC)

I hope you're joking about the pentasyllable thing. university and individual are pentasyllables, and squark and brougham are monosyllables, but the fraction of occurrences of individual in Wikipedia which ought to be linked will be a helluva lot smaller than for squark. :-) ― A. di M.​  21:40, 29 April 2012 (UTC)

Joking? Me JOKING? Indulging in persiflage like a lesser mortal? AdM, you do violence to my finer feelings... That which lies there bleeding is my heart. Tread softly on it; it is slippery! <snnnnffff...> JonRichfield (talk) 20:00, 30 April 2012 (UTC)

The larger problem here is that your proposing a system where perhaps any difficult word is linked to allow a reader to look that up. The problem there is that that means there will be a lot of random interconnections between pages just because page A used the word page B is about, even though A and B have no other relationship. The concept of germane links that Tony et al. are proposing make every link provide clearly relevant information to the topic at hand by limiting it to links that extend on the information in the current article. Hence why its rare to need to link to a country from an article about a person, as the reader will gain no new information about the person clicking through the country link. The "What Links Here" facility on the left side of the page is extremely useful in research too, and if it is filled with random un-connected links short of sharing a common term, its usefulness goes away.

WP is not a dictionary: our linking shouldn't be like that too. Yes, other parts of the web use more frequent linking with abandon, but that should not be the case for here when we're a reference site. --MASEM (t) 15:13, 29 April 2012 (UTC)

Ooohhh dear! back to the looking glass, Dumpty, Humpty and all!

What a horrible thought: "a system where perhaps any difficult word is linked to allow a reader to look that up"! Much more of this, and people will begin to think that WP is an encyclopaedia! Mind you, you had better watch out -- subversive elements have been known to link to lesser wikis without the law (pause for hushed horror: WIKTIONARY!!! <gaaasppp...>) What shall we dooo?

No doubt it is such thoughts that drive you to inappropriate terms such as "random" as a four-letter term of abuse.

Masem, the links are anything but random; you know better than that surely? Consider your very own means of expression: "interconnections between pages just because page A used the word page B is about, even though A and B have no other relationship". FTLOM, they are not supposed to have any other relationship! The point of a link is to direct anyone, and only such a one, who finds a need for explanation of a term, to get the relevant information with a click or so. And the relevant information is to be found where? At the linked article, surely?

That is not only sufficient relationship, Masem, it is the precise relationship for which we link. Right? Why do you link, if not for that; surely not to find the articles that do not have such a relationship? You really will have to put this more plainly.

And as for the ideaof WP being a dictionary; why, we do often link to Wiktionary, thereby gratefully availing ourselves of the dictionary function, and WK in turn often returns incestuously to WP to return the favour. A finer and more gratifying, not to say constructive, example of mutualism would take some finding.

There is no question of random links; they are spontaneously highly organised; any link that does not direct the reader to the right address would soon appear in a revision list. Let's put it to the test, shall we? You go to some article with lots of links, and change some of them at random (coin tossing etc) and see what the effect is. Then come back, tell us, and tell us why and in what way the proposed display toggle will cause similar problems. Cheers, JonRichfield (talk) 18:17, 29 April 2012 (UTC)

If this proposal is a compromise between linkers and delinkers, then each position should be addressed according to its own assumptions. Delinkers say "John Smith was an American lexicographer" tells readers in 1/10 of a second that they can read about lexicographers if they want to, which has a probability of around 1%. "John Smith was an American lexicographer" tells readers in 2/10 of a second that they can read either article if they want to. But the probability that someone reading about a lexicographer will suddenly want to read about the United States is more like 0.001%. So let's save the 1/10 of a second and show them a more realistic choice. For the 0.001% of readers, they can take 3 seconds to type "United States" into the search box.

You may not agree with the above, but if you want to compromise with it, please address that opinion for what it is, not what you want it to be. The distracting effect of the link to "American" isn't solved with an option that shows "John Smith was an American lexicographer". "American" is distracting from the lexicography link, so removing both doesn't solve the problem. Even if the lexicography link still exists if you hover over it, the reader won't know that without reaching for his mouse first. Art LaPella (talk) 19:13, 29 April 2012 (UTC)

Art, I refer you to the entry under "Syllogism" in Bierce's "Devil's Dictionary", urging particular attention to the Syllogism Arithmetical. He seems to have had some of your forerunners in mind, who by combining mathematics with logic, achieved a double certainty and were twice blessed. Your .001% is truly creative, but not truly relevant, let alone credible even on an order-of-magnitude basis. Never mind the magnitude though; it is the sheer unreality of your vision that leaves me blinking. Whether the words are linked or not, and whether the links are visible or not no one should wonder about them unless they did not understand their import (except for the super-sensitive souls who find links so cripplingly disruptive that they cannot bear to see them). So far no problem; the super-sensitives would be thanking me brokenly for the link hiding if only they knew about it. But now, suppose that an insensitive reader who either doesn't know about pentasyllables, or about America, or wonders what the heck they were mentioned for, reads the passage and says "HMMmmm???" Maybe he wonders for a few seconds what the heck... Then if the links are showing, he links one or more of them and either blesses the writer for the illumination or curses the pointless interruption. If the links are invisible, he will not pause unless he does not understand what he is reading, in which case, after a few re-takes he wonders... "Hm?", and moves his mouse over. Of course, he may have omitted to activate floating cursors, in which case his cursor says "Click me!" and all is sweetness. Any wasted time is not the reading or clicking speed, and if it were, it would be too trivial to measure. After reading a ten-page article with dozens of links, we would be speaking of a smaller delay than stopping for a single yawn or sneeze. If the links were absent, it might have taken hours extra. If they were present, it might have taken minutes for him to speculate about the disruptions. If they were present, but hidden, so that he would only check those that he wondered about, there would be a net saving of time. And as for your ignorant user who could simply have typed in the words in <snigger> three seconds, not all readers are touch typists. Not all of them would get the spelling right first time.

And please remember that I am etherially indifferent to the squabbles about whether you link to America, visibly or not; please refrain from raising that rather trivial case in this connection. It passeth me by like the idle duck's water which I respect not. The time saved by hiding such squabble bubbles would justify the facility alone. The distraction would simply vanish in either case. Must rush. Bcak later to soothe the wrinkled brows. JonRichfield (talk) 20:00, 30 April 2012 (UTC)

Once again, nobody "find[s] links so cripplingly disruptive that they cannot bear to see them", because you mean all links not some links. The argument is that some links make it harder to find more useful links, not that anyone wants to see no links (which is hard to discuss without the basic example you don't like). After reading past dozens of unnecessary links, while trying to notice more relevant links, the delay would indeed be comparable to a yawn – but the same can be said for dozens of readable but technically incorrect misspellings. And we can't assume that "he will not pause unless he does not understand what he is reading"; someone might want to read more about lexicography because he generally understands what it is, and will appreciate seeing the blue line that says we have an easily accessible article about it. Art LaPella (talk) 20:56, 30 April 2012 (UTC)

Art, the problem would not have arisen if everyone agreed with everyone. The problem is that certain persons know exactly what everyone else ought to like and refuse to countenance giving people choices. They also accuse people who disagree with their (dis)approval of link choice of falling short of the careful choice of links that would make the WP look neat and generally inform everyone who needs information. There would be no one arguing about whether to link to America or Albizzia in any context, nor would anyone be saying the likes of "For goodness sake I wish those guys would stop shoving links up my nostrils; I know what I want to query" or else, "Dammitall, why don't those guys link things; don't the know the value of context and perspective?" There would be no time wasted with people shouting at each other because someone wants a rigid formula for linking and wants everyone else to adhere to his view of the formula. All that would be achieved as long as everyone has the simple insight that you are right; right? So when someone says: "Hey folks, let's face it; we cannot please everyone all the time, sure, and we cannot agree all the time surer, but lets take a simple step of giving most of the people the choice they prefer most of the time, without depriving them of smooth access to any of the available options all the time, instead of assuring them that they will just love our preferences if they will only hold their noses and swallow, why, then the prescriptives go all shaky and explain how horrible it will be. Right? I am willing to put my views to the test. Your views have been open to approval for yonks, and look where we are: fighting for American Lexicography! Real soon now I will have time for a longer discussion, but now, once again I must do you the discourtesy of a hasty half discussion. Sorreee! JonRichfield (talk) 19:52, 1 May 2012 (UTC)

If you want a toggle that gives either side what they want, then one toggle position should be more links, and the other toggle position should be fewer links (although I presume that would be considered impractical for other reasons). Showing no visible links wouldn't please anyone. Art LaPella (talk) 20:11, 1 May 2012 (UTC)

Art has it right. Wikipedia's linking approach should be developing a linking network that is useful for research purposes that, in conjunction with our categorization system and allowance for list articles, makes cross-referencing between similar topics easy to trace and follow. It is holistic approach. The style where we link any "difficult" or "interesting" word is another possible approach but is not research oriented, more "immediate satisfaction"-derived. I don't think this really works well for Wikipedia's end purpose. --MASEM (t) 00:11, 30 April 2012 (UTC)

Masem, I have learned to treat any medicine with "holistic" in the title as a priori probable fruitloopery. For one thing, it generally assumes or implies that the vendors are the only true holists on the block. The very idea of a prescriptive approach to contextual linking is contrary to any constructive approach to holism, and poison to anything resembling (you should excuse the term!) research. JonRichfield (talk) 19:52, 1 May 2012 (UTC)

I strongly disagree. One of WP's lofty goals is to be holistic, and ergo there's no reason to aim for holistic approaches in our policies. "What Links Here" can be a very useful research tool that is diluted by linking done just because a word might seem interesting to a reader or to provide a dictionary definition. There are intelligent computer tools that can use this valuable information to help aid research quickly than humans can but it does require that we are careful and not spam it. --MASEM (t) 22:03, 1 May 2012 (UTC)

You make my point for me. I am the one in favour of holism; you oppose it. Links mean far more, and in far more dimensions, than you propose. In consequences they may be included for far more reasons than "a network for research purposes". You need not declaim to me on the importance of research or of research aids, thank you kindly; in fact in the light of what you have written so far, it takes me all my charity to avoid wondering how you can justify doing so. However, the primary function of links is to increase the value, attraction, interest, versatility and power of the WP as a tool for the user. To cripple such functions for fancied research benefits is self-indulgent sabotage; no doubt you think that one can achieve all the objectives by careful application of "art", but I am too distracted by flying pigs to see your justification clearly. Art in such cases Masem, is a matter of balancing of the relevance of conflicting functions, prioritising those of the most fundamental value, and accommodating as many as one can, consonant with the primary functions. It also requires that one can foresee the nature of the research that one is likely to value in the future, before one undertakes to be maximally conservative with one's investments. Don't speak of "dilution" until you can point out non-functional diluents in the mix. Links to points of interest, links to particular significance, links judiciously repeated within long texts, all are functional, helpful, and rewarding to people who follow them as far as makes sense, no matter who is howling "over-linking" every time he can tick off a point on a counter-holistic, perennially out-of-date list of perceived transgressions. (There's a holist in my bucketist, dear Lisist, dear Lisist...") Pardon my whistling, you need not link to it. ;-) JonRichfield (talk) 09:59, 2 May 2012 (UTC)

"But that does not justify their foisting their preferences on every reader. Let's face it, most readers pass a lot more links unclicked than they follow, whether they should follow it or not. Let them do the choosing, not the linkers or unlinkers. Any lesser degree of reader freedom makes about as much sense as assuming or trying to insist that they read each article desuper deorsum" That, if anything, sounds like an argument to link every word with more than 4 characters. ;-) Linking/unlinking is done by editors, clicking is done by readers. No editor has the divine right to foist their links upon the reader; (s)he must exercise editorial judgement as to what is potentially of strongest interest of the reader. Readers are diverse; editors have different viewpoints. JonRichfield accepts that most readers bypass more links than they follow. That is reason enough to be economical with the links. --Ohconfucius ^¡digame! 07:33, 2 May 2012 (UTC)

OC, for shame! That, if anything, sounds like an argument to link every word with more than 4 characters. ;-)? Why leave 4-letter words in the cold? What about head, head, head, carr, ooid sten, Sten , aeon mole mole mole, germ, germ, germ, bore, or bore? I bet there are hundreds of 4-6-letter words that justify a link for reasons of reference or of connectivity, or even, if some of us have our way, for holistic purposes of research. OC, please remember that proper linking is a skill of art. It is not easily done. (Hm. Is there an echo in here?) Because it is a difficult art, rather than a deterministic technique, we can be sure that it cannot be perfected by any particular authority. Not unless that authority is in a position to tell each reader just what to want to link to. Some of us might lay claim to such an ability, but pardon my polite reserve. Not unless the authority also is familiar with not only the meaning of the word in its place, but with its relevance in its place and its value to each reader. These are points in the art, and not every pruner and weeder is privy to such insights, as is demonstrated daily by the delinkers snug in their righteous citation of WP rules sans sense. Simply because there is no simple criterion, we need to employ such tools as might reduce the worst cases to bearability or even to usefulness. Rules need sense in their application, as I am sure you heartily agree (but do correct me if I am wrong.

"JonRichfield accepts that most readers bypass more links than they follow"? Not so; he asserts. He accepts that if 100% of readers bypass 100% of links we could and should omit 100% of links in 100% of articles. But OC, it is not a quantitative relationship; If 50% of readers bypass 50% of links it does not follow that we should omit 50% of links in 50% of articles. It does not mean that we should omit any particular proportion of links, but only those that are definitely acceptably dispensable, which is not always practical to determine. A link that is of positive use to 1% of readers of that article has earned its place handsomely. The reasons for economy with links we have all in turn mentioned, albeit with varying, even inverted, emphasis. The fact that most readers skip most, is no valid reason. It is neither good faith nor good sense to misrepresent claims that there could be multiple links, commensurately with the configuration of the article, the nature of the topic, and the nature of related articles, as demands that every word be linked. That sort of thing is all very well as point scoring, but it is neither constructive, sensible, nor even honest, and it distorts the rhetoric duels that should have been reasoned discussion. Refer back to the list I posted at the start of the previous section; how much of it do you recognise in the scrolled tirades that follow? Check on the proposal at the beginning of this section; same story again. Not much of an advertisement for the clarity of our link artists, is it? Not much of a prospect for ending the tedium either, you reckon?

Once more. What is intended? How about for example: Permit those allergic to links to hide them entirely. Permit those who are distracted, but wish for linking capability, to find links at the shove of a mouse. Permit anyone to change his mode at a click. Slightly more ambitiously, permit anyone who wants to see every distinct link once only, to do so, and assist anyone who wishes there were a link to search for subject matter without going to the top of the page and typing or copy-and-pasting. If we really want to get fancy with our "research" then we would collect stats for individual links (or do we already? I never asked.) and any link that no one has visited, we think about clearing automatically after a year or some such period, whether it is a work of art or not. Now OC, will you please explain which aspects of such or similar criteria strike you as sounding the knell of civilisation? JonRichfield (talk) 12:24, 2 May 2012 (UTC)

Well, thank you for that most sarcastic and provocative rant. I'm really sorry I said 'words of four letters or fewer', it would appear that some would go so far as to link words of what, three, two, or maybe even one character? The mind boggles. As to making links invisible, that can be done by including a { color: #000000} at the top of one's style sheet. But why would I want to miss out on all the fun?

Death knell of civilisation? I wouldn't ever dare to characterise your rants in similar hyperbole. Simply said, it's WP:TLDR. --Ohconfucius ^¡digame! 13:00, 2 May 2012 (UTC)

And thanks in return for a most reassuringly diagnostic response. Dismissal of material as sarcastic and provocative ranting, material that you didn't read, and points that you could not address, concerning matters on which you have been very voluble, is comfortable grounds for dismissal in anyone's book, I'd say, O Confucius, wisest and most virtuous of mortals. "color: #000000" indeed! You really didn't get it did you? That is what happens when you refuse either to read or think about a problem; system design is more challenging than that, dear! It is a hard world isn't it? At least when A. di M. suggested something like that, he was more creatively practical. And... links of less than 4 characters? You mean like i, e, Om, Ohm, and the like? Your mind boggled at that? Boggling sure isn't what it was in the old days! All of those (and several more, if you bother to think for a moment) not only have their links, but also thoroughly justify them (unless you reckon that anyone too ignorant to know their significance by heart, is too ignorant to benefit from a link and needs to be protected by authorities such as yourself). In case you read as far as this, do have a nice day. JonRichfield (talk) 15:00, 2 May 2012 (UTC)

For lengthy text not to be TLDR presupposes that it's worth reading and non-aggravating. I find your contribution to this discussion has been dismissive from the beginning, and more recently condescending to boot. You obviously have a more elevated opinion of yourself that seems justified. Note that playing with my username in that way would count as being uncivil if not a personal attack. No, I don't merely unlink, thank you; I retrain plenty of them to make them more useful to the reader. Now if you don't mind, I'm out of here. --Ohconfucius ^¡digame! 16:49, 2 May 2012 (UTC)

Jon, you're approaching this from "what our readers want" instead of "what the Foundation wants". The Foundation, ultimately, is paying the bills to build a research-grade encyclopedia, not a random collection of information. That includes making sure the work is useful as a research tool, not just another page on the internet. --MASEM (t) 14:04, 2 May 2012 (UTC)

Masem, you nonplus me. Could you please explain how you infer that the foundation wants anything better than "what our readers want"? I don't believe it for a moment, but if you are correct, kiss WP goodbye; it is doomed, free or not. (Incidentally, I assume that you mean "our readers and volunteer workers such as editors", right?) But even if your inference were justified, you still fail to explain how and why you you conclude that the proposed measure would do anything but improve the research accommodation facilities along with the user accommodation facilities? and in what way do you imagine that your difficult art is contributing to the solution of the problem? I am running out of ways to explain it more simply, or examples to illustrate it more clearly, while you just keep repeating the dilution/research bit without presenting any justification. Why don't you try presenting a few concrete examples for a change? (Just, please not OC's style sheet; MEGO!) JonRichfield (talk) 15:00, 2 May 2012 (UTC)

Oh, and my apologies; you still have not explained your "random collection of information" idea. Functional and usable links reduce the "random collection" aspect; they don't increase it. Lets imagine a WP with no links at all; wouldn't you say that that would decrease the structure and value most of all? Remember that I (and I have been assuming you too) am all in favour of links carefully chosen to be as helpful as possible either to a schoolchild or a post-doctoral reader or editor. IFAIK, where we differ is the presentation and selection; or have I missed something? Please explain, if you can. JonRichfield (talk) 15:08, 2 May 2012 (UTC)

The Foundation wants a "free content encyclopedia". Readers would prefer one that would allow a lot more fair use images, popular culture facets, and the like - aka what Wikia is (this is evidenced when you consider readership counts on WP). We could satisfy what the readership looks for, but that would fail the Foundation, the people paying to run the servers that the information is on. It also would fail in the larger social experiment of how to functionally create a research-grade encyclopedia via open wiki. Furthermore, readers that are interesting in improving the work are encouraged to become editors with little to no barrier to entry, and there they can influence how the work should grow. But we still have the core mission of the project to worry about.

You're pulling the "no link" argument that no one here is proposing, so that's a fallacy. Second, we have to dismiss any type of changes that are based on how links are presented if there would be different appearances for different users, as described before from the date delinking issue - it makes the editing a complete mess and creates an undesirable situation. Third, we are not writing for the age group you are describing; our choice of language and approach, by necessity, assumes strong familiarity with the English language and how to use a computer; otherwise we'd be here forever to explain how to navigate WP, much less use it; similarly with linking we assume they understand broad categories of human knowledge - otherwise, we might as well link every noun in every article. The issue that others have pointed out is understand that a link from within the prose of an article should be highly relevant - not just to learn what that term means or is, but to understand the connection between it and the article it is linked on. In other words, we should be aiming to making our linking structure somewhat bidirectional: if I link topic A to topic B, there should be a natural route back from topic B to topic A through links. This is why, for instance, the country linking is typically useless - maybe the relevance from A to B (the country) makes sense, but going from B to A makes no sense. --MASEM (t) 15:25, 2 May 2012 (UTC)

Masem, may I recommend a very careful separation of concerns in this discussion. I don't know when I have seen more cross purposes in debate between educated people. Let's consider some of your text and see whether we can pin down points instead of arguing against the unsaid and unintended. If that does not strike you as a comfortable target, I am not quite sure where to go next.

The Foundation wants a "free content encyclopedia". Readers would prefer one that would allow a lot more fair use images, popular culture facets, and the like

That is a sample. Which readers? No one that you and I are considering at the moment I am sure. I am a bit frustrated about the rigid picture rights policy etc, but big deal, more important things to worry about and at least they are erring on the safe side. Nothing I have been discussing on this page has anything to do with anything of the type. If you want to discuss it, fine, but please not here, with me. What I work on is varied, but it is notable and mainly technical in various fields. The readership I was talking about would be the ones who want links and flexibility, whether reading or writing about quantum dynamics or evolutionary dynamics or for that matter poetic dynamics. So please be a little more selective about your use of terms such as "the readership". The acne brigade may be very important for FB or vandalism, but they will always come in large numbers when they want homework, and hardly at all otherwise.

It also would fail in the larger social experiment of how to functionally create a research-grade encyclopedia via open wiki.

Impressively put, but you still have not even dropped a hint as to how this should happen, in particular how a modest difference in links would affect the quality of a research-grade encyclopedia via open wiki. You haven't even said what a research-grade encyclopedia might be. You just keep on repeating the phrase. Any encyclopaedia of decent quality, making allowance for the academic level of the readership, should offer breadth, depth, interest, attraction, stratification, quality, accessibility, efficiency, versatility and accessibility. And a few other –ities that do not as glibly spring to mind. Not one of these is intimately related to the precise adherence of linking to your own unformulated standards, nor have you shown how it should or could be.

You're pulling the "no link" argument that no one here is proposing, so that's a fallacy.

I am doing nothing of the kind, never have been and your repeated claim is accordingly ill-taken. The fallacy is all your own, so you can forget about saying anything meaningful before you drop it.

Second, we have to dismiss any type of changes that are based on how links are presented if there would be different appearances for different users, as described before from the date delinking issue - it makes the editing a complete mess and creates an undesirable situation.

Unlike you, I have been specific about my claims and have presented several examples of what I mean. You have not paid any attention to my repeated challenges to you to demonstrate a solitary scrap of harm that might befall if someone were to choose an option that someone else chose not to use. So let me help you. For example, I use the floating cursor option; I use it heavily. Lots of people do not. Obviously, according to you something horrible will happen if I do not relinquish this vile habit, or alternatively if others fail to adopt it. Please explain how and why. Or explain what I misunderstood about using a facility that others don't like.

Third, we are not writing for the age group you are describing;

Masem, where did I mention age groups? What the blip does age have to do with it, once the kid can walk, talk and read, and has not yet gone gaga? I have taught eight-year-olds who shocked me by learning the basics of programming faster and in greater depth than most practising graduate professionals. Conversely, one of the first men I assisted with computer problems turned out to have sat down with a Fortran manual one weekend and produced his first working program unassisted on the Monday. This was in batch job mainframe days, please note! He used to come to me when his reps and support couldn't help him. He was in his late fifties nearing retirement at the time, but razor sharp; I never had to explain a point twice or at length. I have dozens of similar anecdotes whose plural is close enough to data for me to treat students with respect reflecting their performance, so don't come telling me about age groups I never wold have referred to.

our choice of language and approach, by necessity, assumes strong familiarity with the English language and how to use a computer; otherwise we'd be here forever to explain how to navigate WP, much less use it;

That is not much better. I also have worked with people whose English I could hardly understand and who suffered in trying to understand me, who none the less performed excellently in original and productive work, so it would not become you to sneer at the likes of them. Sure there were duds too, but there always were duds; even among the Anglophones.

The issue that others have pointed out is understand that a link from within the prose of an article should be highly relevant - not just to learn what that term means or is, but to understand the connection between it and the article it is linked on

If you did not understand that from the word go, what are you doing here? And you lectured us on linking being an art?

In other words, we should be aiming to making our linking structure somewhat bidirectional: if I link topic A to topic B, there should be a natural route back from topic B to topic A through links.

Come off it. Represented as a graph, the arcs between the nodes in a relevance relationship can be unidirectional in either direction, or bidirectional. They even can be asymmetrically indirect. That is a day 1 insight. In particular, it is common for a link to be monodirectional. If I follow a link to cataphyll from Searsia, that is quite reasonable, because most Searsia species have cataphylls around certain classes of buds. However, there is not a solitary thing I need to want to know about Searsia when I want to know about the nature of cataphylls in general, or about cataphylls in reference to say a Crocosmia. At such a time I might never want to know whether such a genus as Searsia exists at all. And neither that nor your country-linking example is relevant to the format of links either. And the date option you mentioned has even less to do with it; it is not even qualitatively related. The date thing affected data format. The link only affects presentation without affecting data or demanding special hardware. We have had a couple of folks in this very discussion independently and correctly, though not relevantly, pointing out that one could achieve the same appearance by fiddling one's own style sheet. Remember? It was just a pity that the option did not meet our needs.

Your move. JonRichfield (talk) 19:19, 2 May 2012 (UTC)

Indirect links etc Does anyone have definitive advice, please?

For a change I have a question rather than a proposal, so maybe this will go more peacefully, I hope. I have had several of my links that had referred to redirs changed (I was vaguely under the impression that sometimes a bot was involved, but as far as I can confirm, the bot was only referring to disambigs. If I am right about that, the relinks must have been done manually by do-gooders.) So in my sweet conscientiousness, I have been avoiding redirs in links, and short-cutting them when I do notice them. Now I see this in the manual:

Piping and redirects. Per Link specificity above, do not use a piped link where it is possible to use a redirected term that fits well within the scope of the text. For example, let's assume the page A Dirge for Sabis is a redirect to the page The Sword of Knowledge, and while you're editing some other article, you want to add a link to A Dirge for Sabis. You may be tempted to avoid the redirect by directly linking to it with a pipe like this: A Dirge for Sabis. Instead, write simply A Dirge for Sabis and let the system handle the rest. This has the added advantage that if an article is written later about the more specific subject (in this case, A Dirge for Sabis), fewer links will need to be changed to accommodate for the new article.

OK, so that makes sense and I have no gripe. But do the guys who have been shortcutting and incidentally misleading me know something I don't know, but should know? Subject to anything anyone has to say here, I now intend to link to redirs when their titles fit my text, whether the redir is a synonym for the main title or not. (If the main title suits just as well however, I'll change the text accordingly. No point having a redir without a particular function.)

Any correction?

Cheers, JonRichfield (talk) 15:47, 2 May 2012 (UTC)

See WP:NOTBROKEN. In general, whether it's better to write [[foo|bar]] rather than just [[bar]] (assuming that "Bar" redirects to "Foo") depends on whether "Bar" is a {{redirect with possibilities}}. (Personally, I'd go further and say that it only makes sense to write [[foo|bar]] if "Foo" is an {{unprintworthy}} redirect – I can't see the point of writing [[gasoline|petrol]] when [[petrol]] suffices, for example, except possibly in individual I-know-when-I-see-it special cases of which I can't think of any right now.) ― A. di M.​  15:56, 2 May 2012 (UTC)

Thanks AdiM. That sounds good. Go well, JonRichfield (talk) 18:51, 2 May 2012 (UTC)

Also consider that there are bots that go through to correct piped redirects directly to the pages in question. Redirects are "cheap" and we should encourage users to not worry about being exact (though if you do know what you're doing, go right ahead). --MASEM (t) 19:06, 2 May 2012 (UTC)

Thanks Masem, that is helpful. I'll be applying it. JonRichfield (talk) 19:20, 2 May 2012 (UTC)

Do said bots at least check whether said redirects are in Category:Redirects with possibilities (or in any of its subcategories)? If they don't, someone should their operators about that. ― A. di M.​  20:04, 2 May 2012 (UTC)

• It's a Wonderful Life‎ (1946 James Stewart film) This evening, I reverted an edit by Lobo512 (talk):(removed Category:Films set in the 1940s using HotCat) – My edit summary reads: Wrong! World War II finishes (1945) during the story of this film.

Surely, these quasi-automated revisions may cause many incorrect deletions and should be curbed. -- Gareth Griffith-Jones (talk) 21:24, 2 May 2012 (UTC)

Hotcat is not an automated tool. (See WP:Hotcat). This isn't a linking issue. --MASEM (t) 21:40, 2 May 2012 (UTC)

Understood. -- Gareth Griffith-Jones (talk) 22:25, 2 May 2012 (UTC)

Major geographic regions

RE: "Avoid linking the names of major geographic features and locations, languages, religions, and common professions." The term "major" is very vague and subject to interpretation. So in an article about a movie shot in Spain, do we not link Spain? In a movie shot in Albuquerque, New Mexico, do we link Albuquerque but not New Mexico? Is there any way we could make this guideline a little clearer? It would help save a lot of time and debate on article talk pages over a trivial question that suddenly becomes the subject of edit-warring, which no one wants. But because the guideline is so vague, this happens. --Tenebrae (talk) 15:05, 8 May 2012 (UTC)

Let's start with the easy case: if you have a movie that's documented to have been filmed in Albuquerque NM, the only geographic link you may need is to Albuquerque, New Mexico, since the link to New Mexico will be found in the city article and is not germane to the film article. As to whether the city needs to be linked, Albuquerque probably falls at the cusp of a major metropolitan area that most English-speaking people would know about, so I would argue it's probably okay to leave linked.

On the other hand, if the only thing about filming a movie in Spain is the country itself and no specific localities, then there's no need to link Spain since that's a nation that we'd expect most English speakers to know about.

As to generalities, my take is that most links to the continents, major oceans, nations, and large metropolises (New York, Paris, London, Tokyo, etc.) are rarely needed in prose on articles that otherwise don't deal with geography or politics (on these articles, I'd expect such links). When you can get more specific than those, linking such places can be helpful. --MASEM (t) 15:15, 8 May 2012 (UTC)

But the point of linking isn't just to enable people to find out what something is, it's to enable navigation to related and relevant entries in the encyclopedia - which are detailed pages, not simple dictionary definitions - whether some putative average user might be expected to already know what those things are or not. The guidance against linking to "major geographic features and locations" expressly and explicitly has an exception for where such things are "relevant to the topic of the article". So, in such cases, even "major" locations are OK to link, along with the more obscure ones, and hence the issue doesn't arise (although we then have a separate issue about what constitutes "relevance"; as noted, it would probably come into play for entries about geography or politics. I'm personally not sure about film locations). However, when the terms come up more in passing or with limited direct relevance, such that the distinction might apply, I agree there's a huge problem here with defining "major", as well as "common" or "well known". I've never known how you are supposed to do that for the millions of people from different backgrounds who alight on WP pages. N-HH talk/edits 15:41, 8 May 2012 (UTC)

My take (and only my take, nothing more) is that while we should be highly selective about what English words are linked in prose, whenever we can link a proper name for the first time with a blue link, we should. We expect people to be fluent in English when reading on en.wiki, but because of localities, they may not be familiar with specific names of people, places, and things that are outside their realm. Thus, linking them in prose the first time they are used is helpful. But to that end, things that we expect every English speak to recognize easily shouldn't be linked, and hence the list of geographical features I offer barring their germaneness to the article. There is another consideration that I use that's a bit harder to define but its based on the bidirectionality of links, and this may limit linking of some proper names. For example, I don't think we need to link Roger Ebert on every film article even if we use his review and cite his name, as conversely we're not going to be linking every film off his page. This, however, is much more the art of linking and not always an absolute. The case of linking Albuquerque above is a case where there's unlikely the return link from the city, but the link is still helpful as to define the real setting the film was created in, and Albuquerque, as noted, is not well known unlike Spain. --MASEM (t) 16:34, 8 May 2012 (UTC)

At the very least, major should be replaced by well-known – Uttar Pradesh is more likely to need linking than Belgium. ― A. di M.​  21:16, 8 May 2012 (UTC)

Is "well known" any easier to define or assess than "major"? Both are pretty subjective and depend entirely on the perspective of the reader. On a related point, your example is interesting as well, since it shows how we should be wary of relying too much on our own assumptions - it's quite possible in fact that more English speakers are going to be familiar with India's most populous state than a small European country. Also as noted, in straight geographic articles where the main topic is related to either of the two terms respectively, each is as link-worthy as the other, regardless of how "well known" or "major" either is. N-HH talk/edits 22:00, 8 May 2012 (UTC)

I think we can be very specific as to what either "major" or "well known", allowing for users to extend to uncited cases, using a footnote to iterate these with IAR/common sense allowances. Were I to iterate them, the continents, major international bodies of water and other major geographic features spanning multiple nations (Atlantic Ocean, Mediterrian Sea, the Amazon river, the Alps, the Gobi desert, etc.), multi-national regions (eg Middle East, Central America, etc.), all countries, and nearly all major metropolitan areas (based on 5M+ metro area population counts, as a rough cut-off line). --MASEM (t) 22:17, 8 May 2012 (UTC)

Wuhan. Sydney. (In other words, a criterion based on international relevance, e.g. “cities listed in the article Global city, would be better than one based on sheer size.) ― A. di M.​  22:37, 8 May 2012 (UTC)

“it's quite possible in fact that more English speakers are going to be familiar with India's most populous state than a small European country” – seems unlikely to me. (En.wiki has more readers from Germany, the Netherlands and France combined than from the whole India.[1]) ― A. di M.​  22:37, 8 May 2012 (UTC)

Although the second stats show India as fourth in the list of countries where most WP visits come from. Anyway, even with stats like that and others, it's hard to make a definitive judgment about what specific things people are familiar with. I'm just wary of making assumptions about such things one way or the other, which is easy to do - my point was quite general in that respect. And even if we were able to assert with some confidence that some things are better known than others, where do we draw the line? It's all very well saying "well, obviously the Atlantic Ocean" for example - it's probably safe to say that most people might well know what it is; but what if that figure is only 95%? Are the 5% discounted? And even the 95% who know what it is might not know much about it (me for one). If we're really saying everyone knows what it is and hence let's not provide links to it, why have an article at all? Plus, again, when directly relevant or related to the topic at hand - say in the entry on the Pacific Ocean or the Bay of Biscay - we should be linking to the Atlantic page regardless of any debate about familiarity or not, for basic navigational purposes. People interested in one are likely to be interested in the other, and the entry on the Atlantic also provides rather obvious comparison and/or context for the other two. N-HH talk/edits 10:24, 9 May 2012 (UTC)

Nick, you say "when directly relevant or related to the topic at hand – say in the entry on the Pacific Ocean or the Bay of Biscay – we should be linking to the Atlantic page regardless of any debate about familiarity or not, for basic navigational purposes." Certainly from "Pacific Ocean"; from the "Bay of Biscay" perhaps – depends on the context. I find Masem's stress on research-related navigation compelling, I must say. In that respect there are so many articles in which you'd probably want to link to specific oceans. Like those listed at Category:Oceans, Category:Coastal and oceanic landforms, Category:Oceanography, Category:Physical oceanography, plus hundreds more. I'm not saying all of those listed articles deserve a direct link to Atlantic Ocean, but many probably do. This is the kind of navigational network that really makes wikilinking hum, in my view. Tony (talk) 10:51, 9 May 2012 (UTC)

Once again, while the rest of you debate unusual cases like the Bay of Biscay, I would rather go to the Atlantic Ocean article, click "What links here", and see some random links. How much do you need to know about the Atlantic to understand that the Queen Elizabeth 2 crossed it, or that a Live Aid broadcast crossed it, or that a plane crashed in it off Nantucket, Massachusetts? Admittedly, in the first two cases it would be somewhat relevant to click that ocean link to see how wide it is – but that isn't the sort of thing you guys are debating. That's the kind of link that has the most effect on Wikipedia, because there are so many of them. Art LaPella (talk) 01:31, 10 May 2012 (UTC).

Oh, and only 95% know what the Atlantic is (and can still read the rest of the article)? In the Atlantic sentence in the QE2 article, the word "refitted" is much less familiar than "Atlantic", according to Google hits, and thus more worthy of a wikilink. Art LaPella (talk) 01:50, 10 May 2012 (UTC)

I don't see way Bay of Biscay stands as unusual vis-a-vis Atlantic or unusual as a theoretical example to raise - it's obviously a related topic (as, per Tony, would be other entries directly related to oceans and oceanography) yet such links often get wiped out, with "overlink" being cited. It's an example that illustrates the debate, as yours do; and, btw, I'd agree with you on them - Atlantic is neither obscure nor, in that context, directly relevant to the main topic and therefore there's a good case for losing the link. I for one would happily see links go when they fail on both counts - but it's the "both" that's important. The key issue for linking locations in the context of films - ie the original question - is how relevant we think locations are to the film per se. N-HH talk/edits 07:38, 10 May 2012 (UTC)

Then I conclude there is now a broad consensus for enforcing the WP:OVERLINK guideline with AWB, if the operator is conservative enough. Slogans like "let the reader decide" apparently have the invisible disclaimer "in a minority of disputed links like the Bay of Biscay's link to the Atlantic, not the majority of cases like the QE2". Art LaPella (talk) 13:46, 11 May 2012 (UTC)

I don't see anything of the sort. Nor do I see in what's been said above (or previously in the wider debate) anything that could be described as "slogans" or "invisible disclaimers". Nor, indeed, do I see any conclusions whatsoever in respect of the main question asked by the OP as to how we would define "major", as used in wp:overlink as written. N-HH talk/edits 14:21, 11 May 2012 (UTC)

ps:in addition of course, much delinking of "well known" things is not done by AWB - the issue transcends the tools. Witness this example, which strips any link to Catholicism from the entry on a noted and explicitly, er Catholic, writer. Someone actually put thought into taking that one out, which makes one doubt what the effect would be of asking people merely to be a bit more "conservative" when using AWB for similar tasks. N-HH talk/edits 14:40, 11 May 2012 (UTC)

That article has an entire paragraph about religion. So if that editor were more conservative, he would have left that link to Catholicism, but removed it from the article about Ricky Martin for instance. Assuming the Martin article resembles the QE2 example, you would presumably be less willing to "let the reader decide". If you agree so far, then the Catholicism link in Ricky Martin exists because nobody has gotten to it yet, an example of why we have AWB. As for how we define "major", I don't think anyone is advocating an explicit list because each delinking requires a subjective decision, and so does almost any guideline. Art LaPella (talk) 16:44, 11 May 2012 (UTC)

How about this concrete example: Canada. In the lede they have North America, Atlantic Ocean, Pacific Ocean and Arctic Ocean linked. Is this overlinking or is it justified? --Walter Görlitz (talk) 01:38, 16 May 2012 (UTC)

Those entities are necessarily linked, and were for a long time earlier, since the lead of the article describes the country's location, size, and such in the context of those entities. More can be found at the article's talk page. Ubiquinoid (talk) 06:26, 16 May 2012 (UTC)

Restating two rather simple points - a) as a general point, links are not simply to explain what something is at a basic level, as if WP entries were dictionary definitions, they are to enable navigation to the detailed pages on related topics; b) the suggestion in the guidelines to not link to "well known" or "major" terms has the explicit exemption where those terms are relevant to the topic at hand. Taking either of those points into account - let alone both of them - it's hard to see how there's any valid arguments in favour of removing the cited links from the lead of the Canada page. N-HH talk/edits 14:45, 16 May 2012 (UTC)

You do realize you contradicted yourself or at least conveniently ignored the text you're quoting. Are you somehow suggesting that the United States are somehow not well known or major terms? The same can be said for North America and both the Atlantic and Pacific oceans. They are no more relevant to the topic at hand than linking Europe or Sweden are to an article on Norway. So why do we need to explain what those are, and why would we want to redirect readers in the lede to subjects that are being used simply to place the subject into a context when there's a map to the right of the image? --Walter Görlitz (talk) 16:04, 17 May 2012 (UTC)

Where on earth have I contradicted myself or ignored the text I'm quoting? Could you be more specific? As it happens I do dispute how exactly we would define well known or major, although do accept that if any terms can be so described, US and Atlantic etc probably fall within that category. But the point is that the guidelines accept links to such terms where relevant to the topic - that's the bit of the guidelines that you and others are "ignoring" when you've been arguing that we should not be linking to such terms at all. Now you're shifting slightly and saying, "Oh, OK, but they're not relevant". The idea that US is not related to Canada or relevant to a page on Canada, or Norway to Sweden or Europe, seems a little odd by any conceptual or category standard. And you've also ignored the rather obvious and endlessly made point that linking is not simply about explaining in two words what or where something is for those who don't know. What has the map got to do with anything at all? And finally, links only "redirect" if readers choose to click on them or open a separate tab. I don't quite understand who appointed one or two editors to remove that option, when those links are clearly relevant and possibly of interest to many - yes, not all - of the people looking at the Canada article. N-HH talk/edits 16:16, 17 May 2012 (UTC)

Well, let's link every noun in that lead. Why stop at 31% of the words? Tony (talk) 13:24, 18 May 2012 (UTC)

There's an idea in this, in perhaps there are separate degrees of germane-ness in the lead and in the body. I'm not sure yet how to say it, but because the lead certainly isn't going to have a lot of repeat links, it can be prone to become the sea of blue that we're trying to avoid. Also consider that infoboxes should have all said terms linked as well. This is a separate issue from the body itself. Again, I'm tossing this out there, it's not well-formulated yet, but I think this might help... --MASEM (t) 13:47, 18 May 2012 (UTC)

Tony. Please, don't you ever get bored throwing that childish and pointless response back at me and others, which has nothing to do with anything given that I've never argued for that or anything close to it? I think that's about the 67th time you've done it now in the last couple of years. N-HH talk/edits 14:37, 18 May 2012 (UTC)

This issue is well illustrated at Canada. We have some links to major geographical features that some editors want linked; others just don't want it crowding out other links in the lead. The solution to move the linking to a more relevant section, where the links show up to better advantage was proposed – a compromise that in no way detrimental to the prized "navigabilty" concept (seems to be code for linking as many words as possible, but no more than once, please...) espoused by NHH. Then some dude rudely shot that down, saying it was a "peanut gallery edit". Go figure. --Ohconfucius ^¡digame! 14:48, 18 May 2012 (UTC)

Actually, the assessment was more due to the editor not having discussed anything regarding this amidst the dispute. But, really, what justification is there to link the major terms half the way down the article and not when they are mentioned initially in the lead, when the country's location is being described (like most other country articles)? Ubiquinoid (talk) 03:56, 19 May 2012 (UTC)

No, the peanut gallery" assessment" was a cheap shot, and I think you know it. The justification, was greater relevance to the particular aspect of the topic, and thus the link is more focussed and potentially useful to the reader. --Ohconfucius ^¡digame! 04:50, 19 May 2012 (UTC)

The comment would not have been made had the editor actually participated in the discussion instead of reverting regardless and possibly inflaming matters - that is what the 'peanut gallery' does. Anyhow, no guideline supports what was an inferior (albeit possibly good-intentioned) effort to deprecate the links by moving them down instead of upfront where they rightly belong. Ubiquinoid (talk) 06:42, 19 May 2012 (UTC)

Can we please stop with the pointless, inaccurate and (ultimately) disruptive suggestion that anyone who questions link removal wants to "link as many words as possible"? That spurious claim has been used over and over again in multiple debates, despite clear explanations that it is completely wrong, and has become nothing more than a transparent attempt to derail the discussions. --Ckatz^chat_spy 16:56, 18 May 2012 (UTC)

The rhetoric is strong, but I have yet to see concrete examples of how you would apply the sensible linking practices you are referring to. The currently available 'proof' is when you follow up one of my edits, replace one or maybe two links, usually of countries with the gracious edit summary 'fix script error', and leave 98 percent of my unlinkings intact. Is that sufficient to prove that I'm so in the wrong as you seem to make out? From where I am standing, it actually seems we aren't too far apart after all. --Ohconfucius ^¡digame! 18:29, 18 May 2012 (UTC)

Who has the time to sort through every single edit you race through with your script, in order to catch the mistakes in delinking? Generally speaking, the articles you pass through that draw my attention are geographical in nature simply because I have a lot of those on my watchlist. As for the edit summary, I would suggest that "script error" is much more gracious than "who in their right mind would delink those terms in this context"... --Ckatz^chat_spy 23:46, 18 May 2012 (UTC)

And OhConfucius, no, "navigability" is not "code" for anything, despite your bid to continue with this rather tedious tactic. The word means what it says and I've always quoted it with reference to the actual wording of WP:LINK and WP:LEAD, which talk about using links to bind the project together and provide context respectively, and encourage links to relevant and related topics within articles. As to the issue of linking in the lead as opposed to the body and "crowding out" etc, see the section below. N-HH talk/edits 17:08, 18 May 2012 (UTC)

See my comment @Katz above. --Ohconfucius ^¡digame! 18:29, 18 May 2012 (UTC)

PS: Where do you draw the line at "navigational linking" . IIRC, it was once called "build the web" on WP but has now been sadly superseded by this guideline. Do we stop linking only when we reach pareto optimality? Any word/term/concept that isn't well know by at least 20% of the readership? --Ohconfucius ^¡digame! 18:35, 18 May 2012 (UTC)

Expanding on the idea of reducing lead linking (without affecting body linking)

My take on leads is that it is there for two primary purposes: first, in a well-developed article, it is summarizing the concepts and organization that would be present in the article - an executive summary if you would - so that you know what's coming as you read through it in one block. Secondary, if you have only a few minutes to understand what a topic is about, that lead is there to quickly familiarize yourself with it. These are not the only reasons for the lead, but to me seem to be 90% of the uses for it.

In either case, the user reading it is likely not going to be following links at that point; if they are going to follow links, they are probably researching for a specific topic and coming at it from a related tangent. (Eg: "Oh, I know there was a movie made by that guy that made 'Scarface' but I have no idea what it is...") That's not to say no links are allowable in the lead, but I would argue that this gives us a rationale to be much more selective than links in the prose. Because the lead is necessarily shorter, trying to reuse links that would appear in the body prose normally in the lead would likely also create a denser sea of blue that can be more difficult to read than the body, also pointing to a need to reduce the link use in the lead. And as noted above, many articles have infoboxes, and often there are repeated links between the infobox and the lead. There's also something about blue links in the lead that personally draws my eye, and I'd hope that following it would lead me to something interesting.

Take for instance an article I was primary author on that just was at TFA, Limbo (video game). Thinking about this approach, the only links I'd leave in the lead would include "puzzle-platform" (the category of game) "Playdead" (the company that made it), Trails HD and Splosion Man (but perhaps even then cutting this line a bit), "film noir" and "German Expressionalism", and "video games as art". Those are all highly valuable and/or highly interesting links that would not distract from reading it as given. All the other links, particularly the list of platforms are duplicated in the infobox. The 5-7 remaining blue links all provide high-value germane links that either help immediately on the research side or are important enough to draw your attention and tease you to follow the link for good reason. The links that are removed are otherwise repeated in the prose.

It's not an exact science but I bet we can make lead linking more selective than the body without too much work. --MASEM (t) 15:22, 18 May 2012 (UTC)

Masem, this sounds very sensible to me. The Canada article has made me think through the role of the lead and its relationship with wikilinking more clearly than before. The guideline, of course, is perfectly consistent with what Walter G has just done at that article, which is to unlink the oceans in the lead, already heaving with links, and to link the same items in the thematic section (Geography). It's much more reader-friendly, and in textual terms, it's logical:

"Generally, a link should appear only once in an article, but if helpful for readers, links may be repeated in infoboxes, tables, image captions, footnotes, and at the first occurrence after the lead."

The only thing one might quibble with is "at the first occurrence after the lead" (alternative: "at a suitable occurrence after the lead", but since the first occurrence is usually likely to be a more focused and logical context for departure via a link than the lead, it sort of works. Tony (talk) 15:32, 18 May 2012 (UTC)

If the argument for not linking to Berkshire if Reading is linked to in the same sentence is that Reading will link to Berkshire anyway, then the link from Reading to Berkshire had better be as near the top as possible, because if I'm using the article about Reading just as a stepping stone to get to the one about Berkshire, I wouldn't want to have to scroll down to do that. ― A. di M.​  16:33, 18 May 2012 (UTC)

That would follow my "research by tangent" example, and thus would make sense for the allowance of links on logical connections (a film linking to its director where'd we expect to find a list of his films, a state/providence linking to its country where'd we expect to find a list of states/providences it contains, and arguably in the case of the Canada lead, a country linking to the nearby major geographic features it borders, assuming those all list out their neighboring features). As a counterexample, I would not link Canada to Russia in the logic that that is the only country that Canada falls behind in land area, as that's not a common sense logical connection. --MASEM (t) 16:38, 18 May 2012 (UTC)

Nice example. Indeed, what makes links valuable is that they favor navigation to the topics that have a close logical connection with the present topic. It has nothing to do with whether the linked-to topic is a major geographical area or not. —Ben Kovitz (talk) 09:04, 19 May 2012 (UTC)

Practically speaking, delinking focuses on major geographical areas because it was once the practice to link every geographical area automatically. Links are ordinarily helpful in case of an obscure geographical place (it's almost meaningless to know that the Battle of Waterloo was fought at Waterloo, unless you can link to a map). But links are ordinarily unhelpful for major geographical areas; for all the debate about special cases like whether to link the U.S. in the Canada article, most links to the U.S. are no more helpful than linking any other random word. Yes, "Wikipedia is not a dictionary", but I never understood why not; aren't dictionaries and encyclopedias separate because paper volumes will break if they are too big? Our interwiki links solve that problem to a degree, and make Wikipedia a kind of dictionary anyway. Art LaPella (talk) 17:52, 19 May 2012 (UTC)

Masem, no offence intended, but I think that you're incorrect in your assessment regarding the likelihood of people following links from the lead. The arguments I've seen with regard to this concept in the past appear to be based on the idea that readers will not - upon starting an article - immediately want to click away to a different one. However, that does not accurately reflect the modern, multi-tabbed browsing experience we have today. Readers can easily open interesting links in separate tabs without interrupting their reading of the initial article. --Ckatz^chat_spy 16:50, 18 May 2012 (UTC)

No current policy or guideline suggests going too far in reducing links in the lead, in fact WP:CONTEXTLINK in wp:lead and WP:LINK itself all suggest more extensive linking than it would seem is being proposed above. I'm also wary of saying "oh we can just link in the relevant section instead" - sometimes that may be the better option, but we need to bear in mind: a) that people often don't go much beyond the lead in articles (and nor should they have to given the state of most WP entries); and b) partly in consequence of that, A. di M.'s point about Reading-Berkshire etc, where without realising it we'll end up more or less disabling any linking options and sending people through endless loops and article scroll-downs to get nowhere. Often the people on this page seem to discuss this topic from an overly theoretical standpoint and also assume or even prescribe what and in what way general - and often infrequent readers - are reading or should be reading. Also, similar to CKatz (edit conflict), I don't think it's up to us to talk too much or too definitively about "distraction", "crowding out" or "diversion". Not everyone's diverted by blue text, and even if they are, who are we to declare that a sin in each and every case? Also, as a final point/alternative, as I've suggested before, I've always been amenable to the concept of the infobox more as a navbox, where links of marginal relevancy can be placed instead of the lead (eg nationality, birthplace, profession etc for biogs for example). N-HH talk/edits 17:00, 18 May 2012 (UTC)

Right now, I'm only postulating on a possible approach with the lead that would necessitate changes to appropriate guidelines. I'm not making any of these suggested changes (in guidelines or in articles), only putting forth some ideas to see if there's any legs to it. By necessity I have to ignore that current guidance favors more links in the lead. --MASEM (t) 12:51, 19 May 2012 (UTC)

Right now, I don't like the idea of deferring links from the lead to the body. This is because of how I usually browse Wikipedia: when I start, I often zip through a bunch of articles very quickly, skimming only the leads. Often I look only the first sentence or paragraph. I click on links in the leads until I find an article that seems worth a closer look, and then I slow down. For example, I start at Eugene Onegin, and I notice that the first sentence links to novel in verse. What, there's a whole article on novels in verse? The reason I was looking into Eugene Onegin was because it's a novel in verse. Or, first I get "the lay of the land" and then I go back and look more closely at the articles that seem most interesting. It's a messy, chaotic process, and it depends heavily on those links in the lead. However, I don't really know that other people browse this way. Does anyone know of any empirical research about this? If not, I'm willing to do some. —Ben Kovitz (talk) 09:36, 19 May 2012 (UTC)

• Interesting choice. The logic here is invalid because although concrete, that article in actual fact seems to be a textbook case of optimal use of the wikilink. 'novel in verse' is hardly what anyone would call a common link. It's "unusualness" makes it a potential magnet. It's highly relevant to the topic at hand: the poet who writes in the style. I would expect few to click on 'Russian', 'poetry', or 'eponymous'; so while 'Russian' may be deemed "necessary" by some, 'poetry' isn't linked, and 'eponymous' seems to be the only potential dicdef link with a very low degree of relevance to the subject. --Ohconfucius ^¡digame! 10:36, 19 May 2012 (UTC)
• Going by my vague rules at this point, in Eugene Onegin, these links would remain in the lead: novel in verse (what it is, per categorization), Alexander Pushkin (who wrote it, supporting research by tangent), iambic tetrameter, Onegin stanza, superfluous men, and possibly both feminine and masculine rhymes, but in both those cases, there's a weird structure to the lead that I don't know if those need to be introduced at that stage (yes, the Onegin stanza is important, but its not important, in the lead, to establish what its format is given we have a separate article on it). The last links there are all "interesting" that I would want to have my attention drawn to in the lead and wouldn't affect the lead density. On the other hand, something like "stanza", while I'd link in the body, is basically sitting there as a dictionary definition that doesn't do much, same with the Russian Poetry (which links just to a list of Russian poets, an easter egg link here) --MASEM (t) 12:50, 19 May 2012 (UTC)

Maybe I misunderstood the idea being proposed in this section. I thought it was to be more selective about linking from the lead than from the body. No? The Eugene Onegin example was to illustrate rapid browsing by clicking links in the lead. The topic that was really of interest was in an article that I didn't even know existed when I started browsing. This example is extremely short: just one link. More commonly, I jump around more. If this kind of rapid browsing is common, this would be a factor weighing against being more selective about links in the lead than in the body. It's OK, even beneficial, to lead a reader away from the article almost immediately; sometimes you don't know what you're really looking for until you see a link to it. But I don't really know how common this style of browsing really is. Does anyone here know? —Ben Kovitz (talk) 13:25, 19 May 2012 (UTC)

The type of navigation you're talking about - akin to the nature of why TV Tropes is a time draw - is that you're browsing without a direct purpose beyond personal curiosity. That itself is fine, but it is counter to the fact that this is a reference work. We could probably talk alot about how various people use WP and how linking affects that use, but they all run counter-purposes to each other, and thus the goal is to try to figure out what mode of use is the one we need to focus on (most likely , research -oriented) and thus how linking strategies should be employed to help that. This is not to belittle the use of interesting links in leads. Eugene Onegin provides several examples of interesting links that a curious user can click on to follow within its lead, and it is not my intention to get rid of those links that can still draw in the curious. --MASEM (t) 13:32, 19 May 2012 (UTC)

I certainly agree that WP should be kept a serious reference work and not cow to certain uses of it just because they're popular. Please see what I've proposed in the following section. I don't follow the connection between this and rapidly browsing along links in the lead, though, and I don't think browsing for personal curiosity isn't research. Do you really mean that? As it happens, though, I was taking a class on translations of Eugene Onegin and one assignment was to write a poem of ten Onegin stanzas. Wikipedia could potentially provide insight into how one sustains a narrative over so much verse. But, why am I getting the Ph.D.? To make a vocation of my personal curiosity. —Ben Kovitz (talk) 14:10, 19 May 2012 (UTC)

As a fairly new editor, I found it odd that the terms of note in the lead -- United States, Atlantic/Pacific/Atlantic oceans ... all proper nouns, no less -- were unlinked. How does it enhance a reader's understanding to have to scoot someway down the page where (illogically, as a few have suggested) a reader may find the 'appropriately placed link' (in the geography section) to click, or make the experience less smooth by prompting them to type the terms out in that instance? I also find there little logic to proposals to link 'United States' yet not the three oceans with which Canada has some 20x as much border with. The country's motto, after all, is 'from sea to sea', interpretations aside. Why place human geography over physical geography? A minority of experienced editors may complain about a sea (or an ocean, or even a lake) of blue, and that is to rightly say that every word need not be linked (like 'country' or 'continent') but this purports to be an encyclopedic wiki -- if you don't want that, stick to paper. And, for those that invoke overlinking, that may apply throughout the rest of the article, but not with regards to the initial instances of these major topics. The guideline is so obtuse that it should be struck. Ubiquinoid (talk) 02:24, 20 May 2012 (UTC)

Striking WP:OVERLINK is a much more extreme position than linking the Atlantic from the Canada article. Click Atlantic, then click What links here, and choose random examples from that list, to find less controversial examples of why we need WP:OVERLINK. Art LaPella (talk) 04:13, 20 May 2012 (UTC)

I am not saying to strike WP:OVERLINK in totality, but the point relevant to major geographic entities, or at least heavily refactor it. While I may question what links to it, why shouldn't the Atlantic have a plethora of links to it? It fronts territories on 5+ continents. Ubiquinoid (talk) 04:27, 20 May 2012 (UTC)

My previous discussion of that question] Art LaPella (talk) 04:37, 20 May 2012 (UTC)

And I do not think that demonstrates anything, particularly in this instance. IMO, all of those examples are relevant in varying degrees - the Atlantic had an important element within the context of each topic - though that may not be with others, in which case editors must be judicious. The guideline unnecessarily causes more problems than it solves. Ubiquinoid (talk) 04:57, 20 May 2012 (UTC)

Those links were randomly chosen, so you want to keep most links to the Atlantic, and thus for you there is indeed no reason for a general guideline. Do you agree that "refitted" is a better candidate for a link than "Atlantic" in the QE2 article lead? How about "diesel" or "oil-fired" in the same sentence? To me, linking all 3 words would be more helpful to someone who is known to want to read about the QE2 than the word "Atlantic". We may assume the reader knows the Atlantic is between Europe and North America; if he didn't, he would stumble over words like "Elizabeth" and be unable to wade through the rest of the article. We may also assume that if the reader wants to know details like areas of higher salinity in the Atlantic, he would be unlikely to be reading about the QE2. And if you want to link "refitted", "diesel", and "oil-fired" as well as "Atlantic", then the sea of blue is nigh. Art LaPella (talk) 05:17, 20 May 2012 (UTC)

I prefer to link than not to link, but not excessively. In this instance, I actually do not agree that 'refit' is more fitting of a link (!) - there is no such article, at least directly, and such a link may need to be an internal one. Even a dissection of the word reveals its basic meaning ('to fit again') - so, no need. On their own, 'diesel' and 'gas-fired' are deserving of links, as they are different means of propulsion which are relevant. The sentence could be constructed so to minimise links, such that 'Before its diesel refit in 1986-87, QE2...' is singly linked to the section of the article where that is described. I would not assume what a reader knows or wants to know regarding the Atlantic - a reader may want to learn of notable crossings, ports and harbours, why it's so named, and the like. And, yes, perhaps its salinity. I would rather not deprive the reader of the opportunity to explore and utilise the article, as opposed to cherry-picking what we think they should be interested in ... particularly if they are wholly relevant to the topic at hand, as in the lead of the Canada article. Ubiquinoid (talk) 05:41, 20 May 2012 (UTC)

Actually, I proposed piping "refitted" to refitting of boats, which isn't anything like the ordinary meaning of the word "fit". Yes, someone reading about the QE2 might want to read about Atlantic ports and harbors, but only if you don't mind the page looking awfully blue. Areas of salinity? I don't think that's any more relevant than the chance the reader wants to see any unrelated article in Wikipedia, such as Babe Ruth, sin, or neutron. If we want to link a reader to anything he might want, the index is a linker's dream because it links to everything. But it isn't used much compared to the Main Page and its search field. Why not? Because the index is an extreme example of how seldom-used links distract from the links you're really looking for. Art LaPella (talk) 17:02, 20 May 2012 (UTC)

You were unclear to what article to link to. After pointing it out, it doesn't seem that linking to 'refitting of boats' is advantageous - the same effect could be fulfilled by linking through the reconstructed sentence I suggested, where the QE2's refit is described. (I'm unsure which dictionary you are using, but an ordinary meaning of 'fit'is to put into a condition of readiness, and 'refit' is 'to fit out or supply again'.) The other articles suggested above have no relevance with regards to this topic and seems a segue, and the index just as well. Do you really contend that the other links you initially suggested are distractions from the topic of this ocean liner? In sum, unlinked items of relevance in the lead, and the editors who promulgate them despite common sense, are the real distractions and oddities here ... so much so that I am surprised how much discussion has been devoted to it. Ubiquinoid (talk) 11:33, 21 May 2012 (UTC)

The main point is that we delink most occurrences of major countries, along with places like the Atlantic, because they were once all linked automatically. Without getting too distracted by "the ordinary meaning of the word 'fit'" (Mommy, my shoe doesn't fit), linking "Atlantic" in that example wouldn't be completely out of line if one is ready, as you are, to at least double the number of links in our articles. Most editors aren't, so I too am surprised at how much discussion this issue raises, and at how distracted we get by issues like the Canada article without noticing how statistically atypical that is – well, not really surprised; Manual of Style issues often go on and on without making much sense. Art LaPella (talk) 21:25, 21 May 2012 (UTC)

Who are 'we'? There seems little consensus about the propriety of delinking major countries and locales. That is, actually, seems a contradiction. I also see little to justify overgeneralisations about the prevalence of editors who frankly deprecate rather than embellish items of relevance. I would suspect that more editors choose to link than not, since this is a wiki where that activity is a basis of the exercise. And, to return to the main issue, no one has yet argued convincingly that the various terms in the lead of the Canada article are not of relevance to the topic at hand (particularly given the context) and, thus, should be devoid of links. Ubiquinoid (talk) 03:19, 22 May 2012 (UTC)

"We" is enough editors to keep the guideline in place, and to keep most of the debate on articles like Canada, not on the general principle that we shouldn't have linked countries like the U.S. wherever they appear. I presume that there is no consensus for doubling our link density (the basis of your rationale for linking "Atlantic"), or it would have already happened. And I have often repeated that Canada isn't my main issue, nor have I expressed an opinion on it; rather I have distinguished those special cases from the need to undo the blind, automatic linking of all countries. I'm having eye surgery tomorrow, so I could be gone for a while ... Art LaPella (talk) 05:58, 22 May 2012 (UTC)

Well, it is currently in place at Canada, and was for a long time before removed by a script. I am confident that the lead will remain moderately dense but not overpopulated, instead of a perpetuation of the dubious link sparsity advocated (deleteriously) by few others. "We" indeed. That is all. Ubiquinoid (talk) 06:29, 22 May 2012 (UTC)

• "I would rather not deprive the reader of the opportunity to explore and utilise the article"—here's the clever line that is spun by editors who want to maximise linking: the underlying assumption that readers are being deprived of something because it's not linked. I think that assumption needs to be examined—not even carefully, but just for a few seconds. Tony (talk) 14:59, 20 May 2012 (UTC)
  • Yes - make all the assumptions you choose to for comfort. Have you ever thought that those 'readers' who may have, say, learning or accessibility issues may be challenged to visit a topic that is unlinked appropriately? Get off your high horse. Ubiquinoid (talk) 15:12, 20 May 2012 (UTC)
    • Since you raised the matter, I asked User:Graham87, who's been blind since birth. One editor, I said, "is claiming that lots of links are helpful to those who have accessibility issues." His response? "Uh, no ... unless you're using Lynx ... in which case, you get what you deserve." Tony (talk) 02:22, 21 May 2012 (UTC)
  • You know, sometimes making it harder for someone to get something can be nearly as bad as making it impossible.[2] ― A. di M.​  19:37, 20 May 2012 (UTC)

Tony, I've examined the assumption that removing links is depriving readers of something, and found it to be a rather obviously sound one (certainly more sound than the various assumptions about how readers should be reading or using WP and how they might want to use links). Could you explain why it is not, possibly with something better than vague claims about distraction or crowding out? N-HH talk/edits 12:02, 21 May 2012 (UTC)

Could you enlighten us as to exactly how you examined the above assumption, and the process by which you were able to come to the conclusion that it was "obviously sound". --Ohconfucius ^¡digame! 12:46, 21 May 2012 (UTC)

Do you really want me to write up a thesis on the similarities in meaning between "removing" and "depriving" and the fact that if you remove something you are, by definition, depriving people of that thing? Is this why there's such an impasse on this issue, because we're not even speaking the same language? N-HH talk/edits 12:56, 21 May 2012 (UTC)

Given that my idea here has been only about the lead and not the body of the article, it is certainly not aimed of depriving such links to the reader, only to remove lower-value links from the lead based on the understanding they likely appear elsewhere in the infobox or the body. So there is a difference here that is getting lost in this discussion. --MASEM (t) 13:03, 21 May 2012 (UTC)

So, these major items are of low value in the lead, but not elsewhere? When a country's location is described succinctly upfront (as with Canada, and many other articles), there is NO better place for these links. I see no guideline, or even sound rationale, which justifies linking the latter and not earlier links. Until this farce, 'United States' wasn't linked at all there.[3] Really, how sensible is that? Ubiquinoid (talk) 13:10, 21 May 2012 (UTC)

I've put out that linking "United States" from Canada within its lead does make sense as a closely-related geography/political term, and ergo should be included. There are several other links, however, within the current lead of Canada that I see of low value if I'm using the lead for research purposes (that it, either I'm only reading the lead to learn what Canada is, or I'm reading the entire article). (However, I still point out that without US linked in the lead, a link to it wasn't deprived elsewhere, as claimed.) --MASEM (t) 13:17, 21 May 2012 (UTC)

Thank you. There was no other direct link to the US article in the article, AFAICT, though many others (e.g., Canada-US border, Canada-US relations, etc.). Ubiquinoid (talk) 13:25, 21 May 2012 (UTC)

There generally always has been a link to United States in the lead of this article; unfortunately, it was stripped away without discussion in March by application of a delinking script. --Ckatz^chat_spy 21:11, 21 May 2012 (UTC)

• So why would you put before readers a very generally scoped article such as "United States" when within seconds' reading there were or will be much more focused links within that topic? The guideline fitly raises this as the principle of link specificity: "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." Flooding readers with links of widely varied specificity within the same topic, in the same vicinity, is not designed to make the wikilinking system work well, I put it to you. Tony (talk) 05:32, 22 May 2012 (UTC)

Links in quotes

This page states that links should "generally" not appear in quotes. This is somewhat vague and confusing, and has at times been interpreted as a ban on links in quotes. I can see many cases where doing so is appropriate. In US politics, for example, Newt Gingrich has often been quoted as saying Obama uses Saul Alinsky tactics, while Obama once was quoted calling Republican policies social darwinism, and both quotes need links to make clear what the person is talking about. To add to the confusion, there is a template saying Wikilinks added that is currently up for deletion. If we aren't supposed to link, why do we have this template? Also, I'd think it's pretty obvious that Wikipedia added the links, so I can't think of any reason we need to say this. Anyways, could we get this policy clarified, both as to when (if ever) to link in quotes and whether we have to note that we added links? D O N D E groovily Talk to me 03:52, 24 June 2012 (UTC)

I believe that that idea is that for whatever scheme you may be using to determine what to link in regular, non-quoted prose, you should be even more selective when you do so within quotes. Sometimes you can avoid the issue altogether by rewriting around the quote to introduce the linked term before or after it, but when that's not always possible, use tighter judgement. I think both cases you list are good examples as understanding the quote's meaning relies on those terms. --MASEM (t) 13:17, 24 June 2012 (UTC)

It seems to me that allowing links on more than a specific basis allows for the possible introduction of spin to the quote rather than allowing the quote to speak for itself. --IznoRepeat (talk) 03:03, 27 June 2012 (UTC)

How about "Link should only be used in quotes if most users wouldn't understand the quote without one. Do not link routine and well-known topics such as place names or biographies." D O N D E groovily Talk to me 04:10, 27 June 2012 (UTC) - I take the biographies thing back, since one of my own examples was a biography :( D O N D E groovily Talk to me 04:16, 27 June 2012 (UTC)

If there's a link without which most users wouldn't understand the quote, then someone reading the article off-line wouldn't understand it either. In such a situation, there should be at least a minimal explanation within square brackets after the item (though having a link as well if the reader wants a fuller explanation is not necessarily a bad idea). ― A. di M.​  10:58, 27 June 2012 (UTC)

Another alternative, of course, is to couch the quotation in brief language (before or after) that does the explaining. Tony (talk) 11:45, 27 June 2012 (UTC)

So in your view, the guideline should be revised to remove the word "generally" and simply say that links shouldn't be used in quotes? Personally, I think linking an obscure term works better than an explanatory sentence before or after, especially when that sentence would be off-topic (which would likely be the case in any context using the Gingrich and Obama quotes above). So, if we simply say not to link in quotes, then the Wikilinks Added template should be deleted as contradicting policy. D O N D E groovily Talk to me 01:37, 28 June 2012 (UTC)

Do we need to note that we added links?

The second part is do we need to note that we added links to the quote if we do? D O N D E groovily Talk to me 04:10, 27 June 2012 (UTC)

Template:Wikilinks added

See the following TfD concerning a sup-note template for indicating when wikilinks have been added to quotations. Plastikspork _―Œ^(talk) 19:20, 1 July 2012 (UTC)

Repeated links in references

I have been editing Nailsea and Backwell railway station, and it was recently subjected to a blitz of delinking on the basis of "WP:OVERLINK". Now, first off, the idea that something should be linked only one in the whole article is just ridiculous, it should be per section as not everyone's going to read the whole article, they may just want to know about a certain aspect. Moreover, the references section should be exempted on the basis of allowing people easy access to the relevant links. There are ~60 refs, and I want to know about say #48, but it's been delinked, so I now need to search through all the rest to find the single one which hasn't been. -mattbuck (Talk) 11:49, 2 July 2012 (UTC)

Well, in the prose, it's long been determined that you only link the first time a term's used within the article, and not per section, and that's been long-standing advice. But for references, its up to the editors of the articles (and generally falling to the style selected by the first editor if nothing clear) that either you are 100% consistent with linking all applicable publications within the references, or linking none of them. Switching that style (delinking them) without consensus is never appropriate. --MASEM (t) 12:35, 2 July 2012 (UTC)

WP:REPEATLINK says: "Generally, a link should appear only once in an article, but if helpful for readers, links may be repeated in infoboxes, tables, image captions, footnotes, and at the first occurrence after the lead." I have always taken that to mean that you can put a link once in the infobox in addition to the link of the same item in the body text, and once in each table (possible exception: large sortable tables, see discussion currently at the top of this page), and so on. For footnotes it surely does not mean one should link the same item over and over again in line after line of the footnotes. I can see an argument for putting in a second link if it is many lines away in the reference block from the previous link, in an article where there are lots of references, but to link the same term repeatedly over a short distance just creates an irritating sea of blue to no useful purpose. -- Alarics (talk) 18:14, 2 July 2012 (UTC)

I look at it this way: people generally read an entire section of prose, but they are highly unlikely to read the entire section of references. They will see a [28], click it, and be taken to the references section to look at that particular reference. If the link isn't there, they have to search "is there a link in reference 27? No. Is there a link in reference 26? No." etc etc.

Regarding links in different sections, I think that guideline forgets that unlike a story where you start at the beginning and work forwards, an encyclopaedia is a reference work, and so people will skip the bits which don't interest them. -mattbuck (Talk) 19:23, 2 July 2012 (UTC)

Embedded citations

Earlier, I was looking for the guideline that says external links shouldn't be used in the body of an article, but I'm a bit confused by what I found. The "External links section" section of this page, specifically the sub-sections "Link titles" and "URLs as embedded (numbered) links" appear to give advice on how to place an external link into the body of an article, which I thought wasn't done. Indeed, WP:CITE#Avoid embedded links (also WP:ELPOINTS) categorically says not to do it. That page and this both link to the essay WP:Embedded citations, which says that use of embedded links is deprecated, but then goes on to explain how to add embedded links. I must be missing something, so can anyone enlighten me? Are embedded citations acceptable or not? And if not, why does this page appear to indicate otherwise? DoctorKubla (talk) 15:55, 4 July 2012 (UTC)

Briefly, it's much easier to get opinions on how other people should write, than to get consensus on fixing guidelines. It's a systemic problem. Art LaPella (talk) 19:35, 4 July 2012 (UTC)

At the end of the day we should avoid the embedded links for citations -- but we shouldn't be discouraging newer editors that can't figure out the citation templates from adding a link to a source that someone else can replace later. It's better to provide the advice of how embedded links work - and then note that they should ultimately be replaced with better inline citations, than to ignore the simple aspect that will leave new editors confused. --MASEM (t) 21:52, 4 July 2012 (UTC)

In that case, a brief "(However, see WP:ELPOINTS)" would relieve the contradiction. Art LaPella (talk) 22:13, 4 July 2012 (UTC)

Policy question - red links

To what extent is the creation of red-links either a) vandalism, or b) counter to "over linking" as currently defined?

This is part of an on-going dispute I am attempting to resolve. With thanks, doktorb _words^deeds 21:17, 7 July 2012 (UTC)

Did you and/or the other guys study Wikipedia:Manual of Style/Linking#Red links and/or WP:REDLINK? If not: Some red links are good. Some red links are bad, but that doesn't necessarily make them vandalism. Only bad faith edits are vandalism, like Art is a poopyhead. Most links to be removed according to WP:OVERLINK are links like United States, so it's hard to imagine them being red. If your question is more complicated, please explain. Art LaPella (talk) 23:33, 7 July 2012 (UTC)

"Avoid linking the names of major geographic features and locations"

I gather this is something of a recurrent topic. Would I be a million miles away if I were to suggest that the (italics) "major" should be primarily read as meaning linking larger, less specific entities when somewhat redundant with more particular ones? (Context: I'd never heard of this guideline, but its exact meaning and application is being discussed over at Talk:Madagascar#Country name links.) 84.203.39.242 (talk) 07:35, 20 August 2012 (UTC)

Yup, it recurs quite a lot ... and there is quite a split over exactly how to interpret the guidelines we have, such that you'll get different opinions here (my guess is that there's a 50-50 split). Some - such as myself - take a more generous view of "major" and "relevant" and, without linking everything over and over, are happy to have a few more links there on the basis that you'll never know exactly what every single passing reader might want or need to look at next or later; and are wary of cutting back reader choice and forcing them only to have the option of those links a small group of editors think are "better" for them, even if those readers don't know it themselves. "People know what China is, therefore we shall virtually never link it" seems to miss the point twice over - not least since the page does not simply explain what China is, dictionary style, but provides a detailed and lengthy encyclopedia entry that someone looking at, say, the article on Taiwan, is likely to be interested in as well. Anyway, in this case, the original removals there seem to based on someone misreading the guidelines, let alone overinterpreting them, such that even the more keener delinkers have come down against them. N-HH talk/edits 10:02, 20 August 2012 (UTC)

So a popular (if not universal) interpretation is that this does mean "don't link to countries, or at least not to ones that are on CNN regularly"? And not just where it's subsumed by a more specific link? (Like a subdivision or location in that country, or a more particular article on some abstract aspect of that country.) Isn't this getting us back to the pre-Johnsonian state of affairs of "horse: a beast well known"? I think I feel myself becoming a more keener relinker (as it were), even as we speak. (Though since the article concerned is permasemiprotected, I'll have to let that go for the time being.) 84.203.39.242 (talk) 10:20, 20 August 2012 (UTC)

Links to countries are best suited when you are talking about articles on geography or politics when they will likely be most germaine, eg the example of a link to China when discussing Taiwan's trade economy, and particularly an appropriate section. But if the country name is just being dropped as a location but otherwise not discussed in any depth, it likely shouldn't be linked. --MASEM (t) 13:27, 20 August 2012 (UTC)

And most often when I see a link to an almost universally known entity such as China, it would be far more useful and relevant if substituted with a link to a section (e.g. "Government", "Culture") or a daughter article. Or not linked at all if it's just a formulaic link. Tony (talk) 13:33, 20 August 2012 (UTC)

As specific a link as in sensibly possible is fine. It's more likely to be the "right" one, and if it's not, it'll at least get you there sooner or later, Baconishly. It's the "not linking at all" I have issue with, and the seeming whimsy with which "formulaic" is being determined. They speak two languages (officially) on Madagascar: linking one of them is OK, but to the other would be "formulaic"? In a list with both "fishing" and "forestry", how might a third-party observer determine which of these is "formulaic"? 84.203.39.242 (talk) 00:43, 21 August 2012 (UTC)

Like writing good prose, developing a sense for what to exclude from the linking system sometimes involves grey areas (but "fishing" and "forestry" are not in that grey area: if you speak English, you're meant to know what they mean). The English Wikipedia, unlike most other Wikipedias, has moved gradually but definitely towards a more selective use of wikilinks. For example, you'd allowed "bird" to be linked in that article, before I came along and unlinked it. This kind of trivial linking debases the currency. We owe it to the readers to use our knowledge of a topic to select only the most useful and relevant items for linking. There's a fiction about that not linking an item somehow denies a reader the ability to type that item into the search box. We assume readers are not that lazy if they really want to divert to "French language" instead of reading the rest of the article on Madagascar. Tony (talk) 00:48, 21 August 2012 (UTC)

If both those terms are quite so black and white, why did you delink one, and not the other? When they were right next to each other, indeed. (I hesitate to ask, lest doing so occasion a fresh round of delinking (of an article with which my original issue was manifest and now-seemingly-uncontested underlinking, mind), or retooling your script to delink more aggressively still, but curiosity compels.) Again, you're assuming that a link is only going to be beneficial if the reader is going to be so blindsided by the opacity of an term or concept that they're simply unable to proceed with further reading of the current article. One can only counsel against projecting your own apparent preferences onto what's likely to improve or worsen the experience of a range of readers. There seems to be some confusion as to my role here. I've never edited that article, and can't at present (or for the foreseeable future) do so, as it's indefinitely semiprotected. I take -- and indeed, can take -- no responsibility for any preexisting overlinking, present underlinking, or new-minted wonky capitalisation. Candidly, the "fiction" being peddled here is that "type it into the search box, you bum" is a cost-free or universally good solution. It's extra work, as is implicitly conceded by your "lazy" characterisation. Making people do extra work for no return isn't necessarily as morally improving as you evidently feel. No link means that existence of the article fails to be flagged. The reader doing so is left to do all the work of filtering and contextually disambiguating the results of said search, which they could have been entirely spared by a link. Or indeed, were spared that until someone decided to "redress" the existence of such. (I don't know why I bother to enumerate these -- it's a dead cert you're familiar with them, as your avowed crusade for link minimalism guarantees someone will have put them to you before, but I suppose we must go through the ritualised motions of "debate" on these. Never mind the whole "implicit in the premise of a wiki" concept; "list of disconnected pages on the web"-paedia, anyone?) 84.203.39.242 (talk) 02:43, 21 August 2012 (UTC)

None of these are insurmountable, but that cost is quite blatantly vastly higher than the nebulous "Defence of Link Sanctity" you're presenting as a balancing risk. 84.203.39.242 (talk) 02:43, 21 August 2012 (UTC)

I do agree with Tony about the linking of common, standard English-language words such as "fishing", since we can hopefully assume that most people will know what these things are. As the guidelines suggest, only when they have substantive and direct relevance to the main topic of the article should they probably be linked. However, the "go and use the search box" argument for manifestly relevant potential links seems pointlessly reader-unfriendly. No one has ever suggested that route is somehow not available - just that we have the ability to provide much easier and direct navigation to relevant and related other articles here (and not simply articles, or specialised sections of articles, that help explain the current one). Why deny people browsing here the opportunity to do that by suggesting that they are too "lazy" to go another way or even, apparently, too stupid to work out which related links they might want to click on - often in a new tab for later reading - or not as they go? And even if we want to do that, who selected the elite that supposedly can "select only the most useful and relevant links"? People are only "diverted" if they wish to be - both by choosing to click on the link and doing that by clicking on and away from the current article. Ultimately, a small number of editors have taken it upon themselves to hobble one of the most useful features of an online encyclopedia, on the basis of their own views and speculation about aesthetics, the average knowledge levels and interests of the average WP visitor, and what is supposedly "better" for everyone else. N-HH talk/edits 07:51, 21 August 2012 (UTC)

"Why deny people browsing here the opportunity" ... um ... this presupposes that we deny entry to other articles unless they're internally linked. We've been through this many many times. To anyone not used to this talk page, the matter comes up reliably every three months. Next will be November; then February. Tony (talk) 08:58, 21 August 2012 (UTC)

N-HH, it's not a question of "who selected the elite", it's a matter of forming a judgement about what best to link and what not. We have to draw the line somewhere, otherwise every other word will end up being linked and we shall have a user-hostile sea of blue. Ideally we want the reader to have confidence that any word or phrase that is blue-linked is actually worth clicking on for more information. As Tony says, having too many links devalues the currency of the whole linking system.

But having said "we have to draw the line somewhere", I do think there is room for debate about whether the present guidelines draw that line sufficiently clearly. At present they spell out that it is pointless to link "United States" in most articles that are not specifically about that country, and that is about all. Maybe the guidelines need at least to give some more examples of both good links and bad links. -- Alarics (talk) 06:14, 22 August 2012 (UTC)

“To anyone not used to this talk page, the matter comes up reliably every three months.” Yeah; so does the idea that people don't have to “divert” to anything to follow a link because they can just open it in another tab and keep on reading the original article. (I feel bad I'm repeating something that's already been told umpteen times, but you never seemed to acknowledge that, and kept on using that ‘diverting’ argument. This time I'm adding a link ;-) just in case you're not familiar with the idea.) — A. di M.  09:37, 22 August 2012 (UTC)

I agree there has to be a judgment somewhere - I for example (despite being told often that I do) don't want to see every word linked that can be linked. However, some people might take the view that that would be better and they are entitled to argue that case. The problem comes when a small group declare that they have found the objectively "correct" way to link and that this is something that has been "agreed" in recent years on WP (despite never being able to show where this consensus was established). They then insist that anyone who disagrees, even mildly about where the line might be best drawn, is wrong and that any articles containing offending links need to be purged, even when regular editors on those pages seem quite happy with the extent of linking and even when the guidelines allow for such links, when read properly.

Also, it's worth noting, for example, that you will find people running scripts or doing manual edits that automatically remove any link to US, France, English language etc, including from articles about things from - and even geography relating to - those countries, reducing navigability to important articles and leaving inconsistent linking in their wake. And finally, on the diversion point, the other point of course is - so what? Who are we to try to divert people from being diverted, as it were, if that's the choice they make (clicking on links is not compulsory of course)? I quite often enjoy flitting from page to page to related topics, something I could never do easily with the paper encyclopedias I used to have. And finally, finally, doesn't the fact that it comes up every three months - usually in a thread started by a rather baffled newcomer to these pages - suggest that the consensus on radical delinking among the wider group of readers and editors is not quite as strong as regularly asserted by some (very definitely not all) of the regulars here? N-HH talk/edits 10:47, 22 August 2012 (UTC)

Linking everyday English words

I notice there has been some to and fro with the sentence about linking to everyday English words, but I couldn't see any recent discussion. So I am starting a discussion now. Adding {{citation needed}}, a template designed for use in articles is not appropriate here and I have removed it. Perhaps the term "everyday word" is better than "non-ambiguous word" in this case. There are lots of non-ambiguous words which are highly technical and will certainly benefit from being linked. — Martin (MSGJ · talk) 09:31, 19 September 2012 (UTC)

I couldn't make head or tail of the language that has been added there recently. Consensus for change will be required. I'm unsure what is wrong with the current guideline. Tony (talk) 10:30, 19 September 2012 (UTC)

If you are unsure, then be sure to ask first. Incnis Mrsi (talk) 10:31, 19 September 2012 (UTC)

Yes, continue this way, please. Making a revert daily without providing an argumentation is an effective way to convince people. Incnis Mrsi (talk) 06:51, 20 September 2012 (UTC)

Thanks MSGJ, but you could provide more clear edit summary in this circumstances of an edit war. Not "partial revert", but "reverting a bare revert button pushing", or, even better, "restored wording of Incnis Mrsi". Incnis Mrsi (talk) 10:31, 19 September 2012 (UTC)

Now that the citation needed has disappeared I am much happier with the current version. However I still think that everyday is a better adjective than plain. — Martin (MSGJ · talk) 11:48, 19 September 2012 (UTC)

I have removed the rfc tag temporarily because I think it is somewhat premature. I'm not even sure what we are discussing now. If we are having a rfc, per WP:RFC we need a neutral and succinct statement about what the discussion is about. There is no such statement present yet, hence the removal of the tag. — Martin (MSGJ · talk) 11:31, 20 September 2012 (UTC)

My two cents:

1. When you revert an edit, you had better say what you think was wrong with it (unless it is really obvious).
2. The template that was used is {{clarification needed}} not {{citation needed}}; the former makes perfect sense to me in that context (what exactly is meant by “plain”?).

— A. di M.  15:45, 20 September 2012 (UTC)

(Personally, I would say: Unless they are particularly relevant to the topic of the article, avoid linking words and phrases whose meaning in context can be expected to be understood by almost all readers, such as: • everyday English words; • the names of etc. — A. di M.  15:53, 20 September 2012 (UTC))

It is also not bad. The only thing I added beyond a pure grammar was in a given context, which is, actually, an important clarification. Such word as "noise" is everyday English, but unlikely we should expect that almost all readers will understand "noise" correctly in the context of signal processing, or, more esoterically, genetics. Incnis Mrsi (talk) 16:28, 20 September 2012 (UTC)

Link dispute on Template:LGBT

Per RfC I've started a discussion at the project who is chiefly responsible for the integrity of the template, discussion at Wikipedia talk:WikiProject LGBT studies#Link dispute (another anti-gay hate group one) on Template:LGBT. Per WP:EGG and WP:LINKCLARITY I've moved the link to List of anti-gay hate groups since it's a list and not an article solely on anti-gay hate groups another editor disagrees. More input at the discussion linked above would be appreciated. Insomesia (talk) 06:52, 21 September 2012 (UTC)

Giving the horse a rest; best practice opinions anyone?

Some technical terms are what one might call nodal or key terms; they keep occurring in certain classes of articles; every person familiar with the respective fields knows them so well that it does not usually occur to them that the terms are any more technical than breadboard or barn, and yet outsiders may not even know that the words are used technically, or other than in everyday English, let alone what they mean in their technical context. However, the specimens that I have in mind are ubiquitous. Common examples include taxonomic terms, such as family and order, lab terms such as artefact, some morphological words such as bulb and fruit, and a lot of physical and chemical terms such as relativistic, work, heat, and compound. There even are some logical/maths/computer terms such as fuzzy logic, trellis, and structure. I am not even thinking of jargon such as heat map, jaggy, or backdoor. Some of these I have definite views on linking to, others I am frequently left in doubt. What I am sure of is that any single, simple, definitive rule is wrong. Any opinions or approaches? JonRichfield (talk) 13:01, 3 May 2012 (UTC)

Similarly, the terms "common law" and "precedent" in the law -- every lawyer knows them. For 90% of the non-lawyer public, what the person knows is just wrong. (I remember my educated layman's view from before law school -- I was wrong). Since so much of WP is directed to making specialist subject matter available to non-specialists, it seems to me that the best practice -- the only rational practice -- is to link specialist terms of art relatively heavily. Boundlessly (talk) 22:16, 23 October 2012 (UTC)

Proposed changes to Principles section

To reduce some of the common confusions about links, I'd like to make the following changes to the "Principles" section. Would you let me know, for each one, if you favor it, or, if you don't favor it, would you be so kind as to briefly tell me your objection? Reasoning for these is given above, under "Salient relations". In that discussion, it seems that most people on both sides of the on-going disputes about over/underlinking agreed, though the concept took some effort to grasp. Hopefully these changes would communicate the main principles of wikilinks to a typical editor much more clearly than our current explanation. Note that a lot of specifics and rules of thumb can be (and are) fleshed out in later sections. —Ben Kovitz (talk) 20:04, 18 May 2012 (UTC)

1. Add:

The structure of all the links between articles is itself part of Wikipedia's summary of all recorded knowledge. The presence or absence of a link from one article to another is information about the relationship between the topics covered by those articles. The links lead readers to clusters of related information while browsing articles, making browsing more fruitful and useful. Because the links are stored in a computer database, they also enable automated querying of Wikipedia through What links here and other methods.

2. Add:

A link from one topic to another means that the second topic plays some special, significant role in relation to the first topic, which gives it much greater importance for the first topic than almost all other topics have. For example, William Robinson is notable in large part because he led a movement against the use of greenhouses to grow plants for gardens. So, the article about him links to greenhouse. Only a tiny percentage of all of Wikipedia's topics have an importance for William Robinson comparable to that of greenhouse. These are the topics that the William Robinson page should link to.

3. Add:

Most links are naturally asymmetrical. For example, there is no link back from greenhouse to William Robinson, because William Robinson is drowned out by all the other topics that have even higher importance for greenhouse. Robinson is not even mentioned on the greenhouse page, not even in the history section.

4. Add:

In practice, many links appear on specialized words and phrases that most native English speakers don't know. However, defining unfamiliar terms is not the primary purpose of links. Wikipedia is not a dictionary. A link leads to further information about a topic that plays some special, important role for the present topic.

For example, algebra should link to mathematics even though anyone who knows what "algebra" means surely already knows what "mathematics" means. A high-school student looking up algebra might have little idea of the full scope of mathematics: what other branches it contains, its applications, its history, its controversies, etc. The mathematics article, or an article reached via a couple more links, might prove particularly enlightening for that reader, even though the reader didn't have that information in mind before starting to browse.

Similarly, the article common law should have frequently link back and forth to specialized terms of art such as precedent and case law. You can't understand one without the other, and the ordinary (non-lawyer) person on the street may or may not have an accurate idea of what each term means.

5. Replace: "Ask yourself, 'How likely is it that the reader will also want to read that other article?' " with the following (in a separate paragraph):

Links should not try to anticipate what a reader is most likely to find interesting or most likely to click. Links should faithfully reflect the relationships of relative importance between topics, so readers can use that information as they choose. Links naturally tend to provide many navigation paths to the "center" of important clusters of topics. Consequently, the path followed by a person browsing Wikipedia reflects both the reader's interests at that moment as well as considerable expertise in all the topics touched upon. By reflecting the real relationships between topics, the link structure tends to guide readers to the information they find most valuable, even when they don't know in advance what that information is. Wikipedia editors need not decide or predict what a reader is looking for--predicting possibility is entirely possible, predicting likelihood is not. Link the former.

comments

Longer is rarely better. 1. Wikipedia should not have the stuff about being the repositiory of all human knowledge in any page, and certainly it does not improve this either. 2. The example is unclear, and not going to help anyone AFAICT. 3. This page has precious little to do with "link symmetry" - it does not help the person reading this. 4. Why not just say that major topics in any article ought to be linked (wikilinked) to appropriate articles? Really short and simple. 5. Keep It Simple. In short, we should keep it short. Verbiage qua verbiage does not assist the user. Cheers. Collect (talk) 12:56, 19 May 2012 (UTC)

The problematic statement on the main page is this:

What generally should not be linked

Generally, a link should appear only once in an article, but if helpful for readers, links may be repeated in infoboxes, tables, image captions, footnotes, and at the first occurrence after the lead.

There are a lot of editors who see the words "a link should appear only once in an article" and stop reading/thinking there. To curb this, exceptions/counterexamples/contraindications/etc. need to be stated explicitly. I agree with BenKovitz' list substantively. And I think it shouldn't be necessary -- but it is. The religious "one link" zealots will only stand down in the face of specific guidelines. An alternative is to state some general principles, as I attempt below. I agree with Collect that BenKovitz' list is a bit too mcuh, a bit too heavy a hammer -- but better that than nothing. I'm fine with either.

Collect, the proposal does not use "being the repositiory of all human knowledge." It says "summary of all recorded knowledge." With the accurate statement, how do you feel?

23:05, 23 October 2012 (UTC)

• • "Ask yourself, 'How likely is it that the reader will also want to read that other article?' " — Heavens, if you asked that, not much would be linked at all. Tony (talk) 23:34, 23 October 2012 (UTC)

Tony, I see from your edit history why you might have that view. Those of us that work on articles to communicate something substantive and complicated see the function of links rather differently than someone who spends most of his time applying the Manual of Style without making substantive contributions. I agree with the point you raise on your User page, that common words ought not be overlinked. We can leave that general principle in place, while limiting the scope of the rule so that complex articles can link as appropriate in their context. Boundlessly (talk) 00:55, 24 October 2012 (UTC)

Complicated articles can already provide links to complicated things ... like technical terms. I wonder why the text needs to be changed for this purpose? What did you have in mind that is currently not included in the ambit? Tony (talk) 02:01, 24 October 2012 (UTC)

Repeating links in the lead and in the body

Because the lead is supposes to function as a summary of the rest of the article, able to be read separately, and because that often is the case, and vice versa, i.e., people not looking for a summary may skip the lead, and many people read the lead only and skip the body, should we generally treat linking in the lead separately from linking in the body, advising to repeat any links the first time they appear in the body even if already linked in the lead – especially if the distance between the lead and body mention is of significant length? I have some proposed text below, with changes highlighted in gray:

Current text:

Generally, a link should appear only once in an article, but if helpful for readers, links may be repeated in infoboxes, tables, image captions, footnotes, and at the first occurrence after the lead.

Lead section

Too many links can make the lead hard to read. In technical articles that use uncommon terms, a higher-than-usual link density in the lead section may be necessary. In such cases, try to provide an informal explanation in the lead, avoiding using too many technical terms until later in the article—see WP:Make technical articles accessible and point 7 of WP:NOT#Wikipedia is not a manual, guidebook, textbook, or scientific journal.

Proposed change:

Generally, a link should appear only once in the body of an article, but if helpful for readers, links may be repeated in infoboxes, tables, image captions, footnotes, etc. However, links that have been used in the lead section should be repeated in the body of an article of any length. This is because the lead is supposed to function as an overview of the body's content, with the body and lead able to be read entirely separately, each containing discrete elements of the other. This separation breaks down, however, in fairly short articles, and especially in ones where the lead is not written as a summary but provides separate content. In such cases, treat the lead and body as one and thus link only once.

Lead section

Nevetheless, too many links can make the lead hard to read. In technical articles that use uncommon terms, a higher-than-usual link density in the lead section may be necessary. In such cases, try to provide an informal explanation in the lead, avoiding using too many technical terms until later in the article—see WP:Make technical articles accessible and point 7 of WP:NOT#Wikipedia is not a manual, guidebook, textbook, or scientific journal.--Fuhghettaboutit (talk) 16:05, 29 July 2012 (UTC)

• People increasingly question why readers would want to divert from the article they've come to in the first few lines of the lead—in fact, anywhere in the lead: it seems like a dysfunctional way to read an article. It is now becoming the more important question: whether the first (and only) link to an item should be after the lead, in a deeper context than the big picture represented by the lead. Perhaps highly technical terms are an exception (as are infoboxes). Tony (talk) 16:21, 29 July 2012 (UTC)

Can you point to where people have indicated this (it may be difficult I know – just memory of loose conversations happening here and there). On my side, I've multiple times seen people say words to the effect that they read mostly leads only (what a shame), as well as the opposite condition; the lead is skipped and only body read. Anyway. I'm not sure I understand what you're saying. Is it that people should not be diverted to a new article while they're reading the lead, which is what a link does? I personally, open up links I come across while reading in new tabs for later perusal, which I do almost without conscious thought because the ability is so ingrained, and continue reading. BTW, I will not be able to respond to anything after this post for many hours.--Fuhghettaboutit (talk) 16:34, 29 July 2012 (UTC)

People can approach and read pages any way they wish to and if they only want to read the lead, or one specific section, or want to divert off quickly, wherever they started, why shouldn't they be able to? Wikipedia is an online reference work, not a paper novel (plus people are even allowed to put their novels down to go and watch some TV - or is that dysfunctional too?). Plus of course, as noted, people can use links to open tabs for later review, such that no diversion occurs anyway. The proposed changes seem broadly sensible to me. Giving readers more options and greater accessibility seems more reader-friendly to me than making unsubstantiated assumptions about - or, worse, effectively giving orders on - how they are or should be reading pages. N-HH talk/edits 16:59, 29 July 2012 (UTC)

The guideline already says "links may be repeated ... at the first occurrence after the lead". Turning the "may" into a "should" seems rather pointless if you then have to explain the exceptions to the rule. Sometimes links should be repeated, sometimes they shouldn't. That's what "may" implies. Aren't we better leaving this to common sense and editorial judgement? DoctorKubla (talk) 17:46, 29 July 2012 (UTC)

It does say that, but as an aside after announcing a general rule: "Generally, a link should appear only once in an article." This does not leave us a neutral stage from which to simply choose, as a matter of editorial judgement, whether to link in the lead and in the body. That decision now becomes a variance from the general rule, only to be done under extraordinary circumstances. Also, to link or not to link is not the type of matter that is innate to people with common sense. You can see this from watching the multitudes of new users who must learn what to link and what not after getting here, even if they appear very clueful. This is far more true of the lead verses body issue, which is something people only begin to question after analysis of how people read articles, and study the function of the lead; not obvious at all, and especially so since so many articles don't follow lead style. The real question is, even if my proposed change should be modified – even if it should say something else, shouldn't this guideline give guidance on the issue of of lead and body separation issue? Right now it provides none, and yet it it is a question that repeats. At the least, if there is some general consensus that a link in any large article a link in a summarizing lead should be repeated in the body, then the categorically announced general rule that a link should only be used once must be tempered in some way.--Fuhghettaboutit (talk) 06:28, 30 July 2012 (UTC)

The problem we're trying to solve is right here in these few words: "Generally, a link should appear only once in an article." That's either nonsense, or entirely overbroad. See discussion at "What generally should not be linked -- can we bring this to closure?" below. Boundlessly (talk) 14:22, 17 October 2012 (UTC)

What generally should not be linked -- can we bring this to closure?

Let's get the proposal up at the top of the discussion -- replace this paragraph

Generally, a link should appear only once in an article, but if helpful for readers, links may be repeated in infoboxes, tables, image captions, footnotes, and at the first occurrence after the lead.

Working text

With this (editable in situ here):

Use links when they're likely to be helpful to a reader who needs or wants more context or assistance in navigation. As a rule there is no such need to link a word more than once in a short article. On the other hand, links may be repeated whenever helpful to a reader, especially in longer articles, in subsequent sections where the link may have specific relevance, and in infoboxes, tables, image captions, footnotes, etc. Links have two main purposes, to lead onto related material of interest, or to provide a gloss, in which, apart from assisting a reader, a link may avoid the repetition of subject matter in separate articles, together with the attendant problems of maintaining, correcting and reconciling material with undocumented duplication.

Examples of helpful (or even essential) repeated links include:

* technical terms that aren't immediately known to essentially every reader, for example, heptasyllable

* terms that are key to the topic of the article and are being used as specialized jargon of the art

* a term that has multiple meanings, and is being used with each of the multiple meanings -- links may be necessary to show readers which meaning applies in which context.

* even universally familiar terms that normally would require no link at all, but may be linked legitimately when used in ambiguous senses that could be misleading or confusing. Examples include names that could refer to countries, regions, national groups, historical circumstances, or terms common in biology, such as "family". In some cases such terms may be dealt with by hatnotes, but in others links are the appropriate medium.

* terms that are used over the course of a long article where the links are far enough apart to not fall on one screen, or where they fall in multiple sections.

The "one link" rule should never be applied to leave a reader hunting for help from an isolated link floating in a sea of text—when a reader is likely to have a question or want help, the article should be helpful, and make the answer instantly and obviously accessible.

Symptoms that may indicate overlinking include:

* When the density of links is so great that links do not stand out, or run into one another.

* Links to irrelevant material or in irrelevant context are always pernicious and may suggest that the article is overlinked.

Unlinking should never be initiated by a single editor, only by consensus. It is up to the unlinker to make an affirmative showing that multiple links are confusing or otherwise functionally deleterious -- mere invocation of "WP:OVERLINK" or personal taste or aesthetics is insufficient.

Discussion

Boundlessly (talk) 21:22, 23 October 2012 (UTC); updated Boundlessly (talk) 22:37, 24 October 2012 (UTC) to reflect suggesions of JonRichfield; updated Boundlessly (talk) 01:16, 25 October 2012 (UTC) to reflect ENeville 17:19, 18 October 2012 u

Good luck, Boundlessly, but you will find that delinking is a religious belief, fanatically adhered to by the delinking faithful and no amount of reason will sway them. Believe me, many have tried for years.-- cheers, Michael C. Price ^talk 07:43, 25 October 2012 (UTC)

Boundlessly substantially edited the "What generally should not be linked" section, apparently in order to support his use of (self-)links at Common Law. I am not into discussing policy, nor into edit wars, so I hope someone will take a second (actually a fourth or more) look. Tks - Nabla (talk) 22:55, 16 October 2012 (UTC)

I agree with the additions to MOS:LINK (though such major changes are better discussed in advance); but having a section link to itself is pointless and confusing (and it's not supported by the new wording anyway). — A. di M.  07:39, 17 October 2012 (UTC)

The issue of multiple links vs. a single link per term per article has been talked to death above, but not brought to closure. Reading through the (long) transcript above, there's a lot of dissent from the "single link" guideline, but nobody has proposed s concrete alterantive guideline. Some junior folks seem to think that "once per article" is an invariant mandatory proscriptive rule, without exercise of judgment. Simply leaving the question open to unbounded "judgment" as the article suggests is not helpful.

So here's a starting point for a restatement of the discussion above.

Use links when they're helpful to a reader that needs or wants more context. In a shorter article, a commonly known word should generally only be linked once in the article. If helpful for readers (for example in a longer article, or a for a technical term that isn't immediately known to essentially every reader, or a term that is key to the topic of the article and is being used as specialized jargon of the art), links may be repeated in infoboxes, tables, image captions, footnotes, etc. Don't leave a reader hunting all over for a relevant link when the reader is likely to want help—when a reader is likely to have a question or want help, be helpful, make the answer accessable and obvious. Links are just electronic spots on a disk—they're not a conserved quantity whose "overuse" will contribute to a global shortage for our grandchildren.

f

I'm a lawyer, until there's draft language, there ain't nothing, and apparently there's been no language proposed so far. So here's a draft. Shall we work this and get some elaboration of "judgment" into the article itself? Boundlessly (talk) 14:15, 17 October 2012 (UTC)

By the way, I think this entire "Manual of Style" has gotten totally out of hand. Some people can write, some can't. Some people have judgment, some don't. The people that can't write keep drilling onto isolated statements in the Manual to make edits that frustrate those of us that can. Every page of the Manual should have a header "This is purely advisory guideline. It is not proscriptive truth. Do not get into edit wars based on this manual. Any reasonable explanation for an exception to this Manual wins." Boundlessly (talk) 14:15, 17 October 2012 (UTC)

"junior people"?! - Nabla (talk) 14:38, 17 October 2012 (UTC)

Boundlessly- your occupation is beside the point and as Nabla points out "junior people" isn't a concept here. WP works through consensus not through appeals to authority[[4]]. Bhny (talk) 14:55, 17 October 2012 (UTC)

(ec*3)As to the point, I have nothing against multiple links to the same subject. Excluding the 'snarky' last remark, Boundlessly's wording is fine. I'd say a one link for section/'page' is a good rule-of-thumb (wasn't it so sometime ago?). But I note that Boundlessly is (too?) strongly motivated to protect a specific article. My issue there is not about too many links, but about self-links pointing to the previous line. - Nabla (talk) 15:01, 17 October 2012 (UTC)

I'll have some comments on this within 24 hours. Tony (talk) 15:29, 17 October 2012 (UTC)

Per the proposed language, I would note that there's no issue of readers "hunting all over for a relevant link" because they can simply enter whatever term they're thinking of in the search box. I do propose that sections be included in the list for permissible repetitions, since often a reader may skip to a section of a longer article and not be acquainted with topics previously introduced. ENeville (talk) 17:19, 18 October 2012 (UTC)

ENeville -- "enter ... in the search box" is called "hunting all over." Especially if a term might appear as differently-inflected parts of speech (so a search won't even work), or where there are multiple synonyms that link to the same destination (again, search won't even work), or where it might not be immediately apparent whether there's a link at all...

Good writers put themselves in the frame of reference of the reader -- what does the reader know right now? What is that reader likely to want to know right now? What will that reader want to do next? A reader that's confronted with a second term that's relevant and necessary to understanding a first topic wants to get to the second topic easily, and wants the "back" button to get back to identically the same context, not have a third task of "searching" that destroys the readers' thought context. (Ironically, as I was editing this note, I had to search -- it totally destroys context, and the train of thought. "Search" is the problem, not the solution!)

The current MOS, applied without judgment, can destroy clarity and readers' ability to navigate. Thanks for demonstrating the point.Boundlessly (talk) 13:01, 23 October 2012 (UTC)

The current MOS, applied without judgment, can destroy clarity and readers' ability to navigate. My observation is that self links like this one make absolutely no sense, and it could be indicative of same faulty judgement that was being suggested, so I would be inclined to ignore/oppose this proposal above. It would be reasonable to assume that it will lead to more of the same nonsensical linking practices there and elswehere. -- Ohconfucius ^{ping / poke} 06:00, 25 October 2012 (UTC)

• What exactly is wrong with the wording as it has been until this discussion? Tony (talk) 00:33, 19 October 2012 (UTC)

I think it's that too many of us have encountered editors who insist on a prescriptive reading of "only one link to the same subject per article". It would seem to me that the original meaning of "Overlink" was to avoid linking commonly-understood words and terms; but some editors seem to believe that the judicious repeating of a few relevant links will lead to an unnavigable "sea of blue", or just believe blue links are less aesthetically pleasing than plain prose. User @N-HH also makes some, IMO, very good points upthread about links serving to bind the encyclopedia together (per the MOS). --Chaswmsday (talk) 10:15, 21 October 2012 (UTC)

So you'd like to spell out acronyms and initialism again and again through an article? Tony (talk) 14:00, 21 October 2012 (UTC)

I certainly wouldn't (and of course the comment above talks about "judicious repeating of a few relevant links", not doing so "again and again" anyway); but equally, I'm not sure that a rule that rigidly insists it can only be done once per article is that desirable either. Beyond that, the comparison/analogy with linking only works up to a point - links are not as visually disruptive or repetitive as the spelling out of acronyms. They also do more than simply explain potentially unfamiliar terms or point people to further details on a topic. Some links will be there for that reason, but there is also the navigation issue, which always gets lost in all these discussions, as if the only function of links is a pedagogic one. N-HH talk/edits 15:52, 21 October 2012 (UTC)

What is the whole thing about if not educational? Tony (talk) 01:33, 22 October 2012 (UTC)

Well, it depends how narrowly or widely you define educational and who we're trying to educate .. one point I'm trying to focus on, as ever, is the broad – and slightly artificial, depending on the initial knowledge base of the individual reader – distinction between a link that takes someone to a page on a thing they have never heard about before, without which they might not be able to understand some aspect of the initial topic, and a link that enables navigation to related topics that they may be aware of in the broadest sense but where that article as a whole may offer further enlightenment and points of interest. N-HH talk/edits 11:17, 22 October 2012 (UTC)

Tony, please don't go off topic. The proposed paragraph sets out a set of exceptions. "spell out acronyms and initialism again and again through an article" is not within the proposal. Red herring, not helpful to reaching consensus on the proposal.Boundlessly (talk) 13:01, 23 October 2012 (UTC)

I concur with Chaswmsday -- "What exactly is wrong with the wording as it has been until this discussion" is that too many folks take the current wording as a proscriptive, black-and-white rule. There's been IMMENSE discussion of this topic (is it a majority of this entire Talk page?) yet too many folks still see every duplicate link as "overlinking." Obviously the existing wording isn't right, if it's generated this much Talk!

I share the view expressed often above, "I'm actually quite annoyed that such a concept as overlinking exists to be honest." But I'll accept that somewhere, sometime, there was at least one situation of clear "overlinking." Sure enough, something needs to be said.

On the other hand, there wouldn't be thousands of lines of dissenting discussion above if the "one link" rule were sound, at least at the breadth stated.

The amount of dissent says clearly that something needs to change. And so I propose the alternative above, that scales the recommendation back to where it does more good than harm. The best way I know to overcome the psychological tendency of some people to elevate one sentence out of context over all others is juxtaposition-- put the contextual limits, counterexamples, and countervailing factors immediately next to the rule of thumb. The proximity places the single statement in its proper context, and makes clear that it's not a "one size fits all" proscriptive rule.

Boundlessly (talk) 13:01, 23 October 2012 (UTC)

Yes there has been much dissent but also much agreement with the way it is now. The page is the result of some consensus. I disagree with your proposed changes Bhny (talk) 19:45, 23 October 2012 (UTC)

By "much dissent", I think Bhny is referring to years ago, when we used to link anything in sight, and editors took a while to get used to a more strategic, rationed approach—or to dissent on this page, which involves just a few editors who would like to see denser and less specific linking in WP articles. Certainly the vast body of editors don't object to strategic linking. Tony (talk) 02:04, 24 October 2012 (UTC)

Well, there's a fairly constant stream of people coming to this page, and commenting on the pages of those who undertake semi-automated link removal sweeps across multiple pages, questioning the extent of what sometimes goes on. By contrast, it's only "just a few editors" who in fact actively undertake such removals. There may be broad agreement on/acceptance of more "strategic" and focused linking, but even that's not universal and, more pertinently, there's a range of views on what would actually constitute said strategic linking in practice. There's certainly no consensus for mass stripping of links to specific terms (again, for the 1056th time, can we please see where this was established?).

As for the wording of the the guidelines, I'm less bothered over this. Broadly I think, as written, they are about right (and that there's a bit too much wordy commentary in the specific proposal above tbh, even if I may agree more or less with the sentiment expressed). I'd favour a slight expansion on repeatlink in respect of "subsequent sections where the link may have specific relevance" or something like that, or changing "Generally, a link should only appear once" to the less prescriptive "there is no need for multiple repeat links throughout an article", but I think the bigger problem is simply the attitude. At the end of the day people editing articles need to remember this is simply a guideline, and that they should use a bit of judgment in individual cases, with linking issues being treated like any other form of more substantive copyediting (ie it's not simply typo correction that can be done via AWB etc). Equally, we need to start from the presumption – even if it is not stated explicitly in the guidelines – that a couple of extra links that one or two editors dislike having there will not bring us to The End of Days, and may yet be useful for other readers. And, more generally, remember what the guidelines actually say currently, rather than what some of us wish they said (eg the "relevance" exemption when it comes to common terms "overlinking"). N-HH talk/edits 08:20, 24 October 2012 (UTC)

N-HH, I like the distinction you made above, in theoretical terms. But don't all of our good articles "offer further enlightenment and points of interest"? Tony (talk) 10:31, 24 October 2012 (UTC)

Tony -- one would think so -- that's how they became "good articles." The problem is, as N-HH diagnoses it, attitude among a handful of zealots -- including the two that started this particular topic -- who isolate the short phrase "a link should appear only once in an article" above all other requests for "judgment" thorughout the Style page. Tony shares my observation that the overwhelming concensus (at least 20 individuals above) note that the "one link" extreme is not appropriate. A small handful of "one link" zealots think they're helping the reader when they remove all but one link between two tightly-coupled and both-very-long articles, both on topics outside the knowledge of non-specialists.

The problem can be solved by rewording the "one link" rule to limit it to the problem it was designed to solve -- obsessive linking of every occurrence of even the most common words -- and stating some of the counterexamples.

I don't claim omniscience for my proposal (at the top of this topic), and welcome improvements, but something along these lines is essential. Boundlessly (talk) 12:25, 24 October 2012 (UTC)

• I don't at all share the multiple-link idea. I do grudgingly accept duplication in the infobox, and in certain cases in sortable table columns. That's about it. To me, linking in a lead should be discouraged ... that is, a higher benchmark should be used. The lead is the least likely place you'd want to divert a reader: the least developed context for a link. Just say no. Tony (talk) 12:43, 24 October 2012 (UTC)

Now I am becoming very puzzled.

First, I come from the school that good writing is in the frame of reference and for the benefit of the reader, not what the writer wants. Are we agreed there? If there's a term that fits into one of the categories in the bullet list I extracted from JonRichmond, then--by definition--we have a reader that is confused, curious, or intrigued. Why not leave it up to the reader to decide whether or not to follow the link--if the article presents the link, there's nothing to obligate the user to click on it, and if the user has a question, voila, the link provides the answer. How is that "diverting" (stick to the context of the bullet list -- no red herrings)? Why do you want to take the choice away from the reader? Why is whether the author does/does not "want to divert a reader" even a relevant consideration?

Second, only a few hours earlier, you had written "N-HH, I like the distinction you made above, in theoretical terms." The bullet list JonRichmond proposed and that I tried to condense is (intended) as a reification of the N-HH's theoretical view. Please recitify these two statements of yours?

Boundlessly (talk) 22:59, 24 October 2012 (UTC)

Whilst I can understand why an editor on specialist subjects may to want to gloss and link more, to apply the idea more widely would be a bad idea all around. The exemptions asked for by Boundless are already there: for example, there's no prohibition on linking to a same article twice if the links are a long way apart. Even for the specialist topic which is 'Common law', I noticed there was heavy overlinking. I have had a stab at improving the linking. Instances abound of poor use of or ambiguous piping of links, or terms that are glossed several times in the space of a paragraph; there are also links that I believe are better made to wiktionary. The proposed text solves none of those issues; the guideline already deals with them adequately. -- Ohconfucius ^{ping / poke} 06:55, 25 October 2012 (UTC)

Comment: A curious unintended consequence of the rules and regulations of Wikipedia (and I'm not saying that I could have formulated them better from scratch) is the degree to which persons who do not recognise the stated or implicit intention of the facility and its function can filibuster all constructive proposals into a paralysing culture of veto. Those who fail to filibuster fall back on red herrings.

Oh well, 'twas ever thus... Still, I can hardly escape responsibility or at least complicity for continued futility if I do not at least show willing. Let's see whether just this once...

Before posting rebuttals, let us concentrate on the meat of the matter. My apologies if this seems as though I am rushing anyone in particular. Boundless proposed:

Use links when they're helpful to a reader that needs or wants more context. In a shorter article, a commonly known word should generally only be linked once in the article. If helpful for readers (for example in a longer article, or a for a technical term that isn't immediately known to essentially every reader, or a term that is key to the topic of the article and is being used as specialized jargon of the art), links may be repeated in infoboxes, tables, image captions, footnotes, etc. Don't leave a reader hunting all over for a relevant link when the reader is likely to want help—when a reader is likely to have a question or want help, be helpful, make the answer accessible and obvious. Links are just electronic spots on a disk—they're not a conserved quantity whose "overuse" will contribute to a global shortage for our grandchildren.

To me this seems like a constructive and helpful statement and in default of anything else (like about a ream of foregoing proposal and justification in this very talk page) I am inclined to support it as it stands. On first reading it I would have thought it to be unobjectionable to all sides, but as some person of great wisdom once put it (I wish I knew who) "if you try to please everyone someone will not like it." And he obviously knew what he was talking about.

Predictably at least one reader of the proposal leapt upon the author's confession of his experience in the legal profession, claiming that "occupation is beside the point". It isn't. It may not be definitive, or even cogent, but it is not an appeal to authority and it is sufficiently to the point to earn a thoughtful reading of a pronouncement along the lines of: "...until there's draft language, there ain't nothing, and apparently there's been no language proposed so far". No doubt putting down something useful might be less fun than trading fusillades of spittle and ink, but I promise you all, if we can get past this bloody counter-constructive nonsense there will be quite enough fun remaining in Wikipedia.

Before I propose any additions please remember some of the functions of links. Do note that no one (and I mean no one) with the slightest understanding of the questions at issue proposes that any text whatsoever be linked where the link cannot be justified on the basis of a reasonably expected function, any more than we would insert an icon of a gluepot at the end of each sentence that the reader might wish to click on in the event that he would like to reassure the Wikipedia authorities that he had read and digested that sentence before proceeding. Personally I think that a lot of unnecessary links might be omitted spontaneously if, before inserting an unnecessary link some of us would think: "Gluepot! Do I really need it?" Conversely, any link that does pass the gluepot test, though possibly dispensable, should deserve reasonable consideration before anyone deletes it. In short (short by my standards anyway) I think it is more important that we have standards for what should be linked rather than what should not be linked. If an editor cannot yet think of a good, positive reason for linking something, that strikes me as adequate reason not to link it — yet. There are far more valid reasons for linking than most novices realise, but they are few enough to list in a convenient support document.

Well then, what are plausible functions for links? The important thing to bear in mind is that they are manifold; there is more to a link than just whether the reader clicking on it understands the word or not. I suspect that it would be a good idea if a list of reasonable functions could be published as a guideline to where a link should be considered, and a matching list might help discourage the insertion of links that could not be expected to be useful. Without serious editing of Boundless' draft I would suggest appending a list of positive criteria for what sort of objective one might justify a link. I suggest that a link earns its place if:

• Some readers might not be familiar with the linked term Ceratopogonidae, no matter how incredible such ignorance might seem to entomologists.
• Some readers who are properly familiar with the term Ceratopogonidae might be puzzled by, or simply miss, its use in a given connection, not realising why it was referred to in context at all, a point which is fully explained in the linked article which in fact might not even be an article dealing with Ceratopogonidae, but might explain the illustration of a heptasyllable in context. In other words, linking is not only justifiable in lieu of definition of difficult words that are more properly defined in another article. (Yes, I know that you know that, but a Caesar's half of the discussion on the matter in some of the sections above, suggest a high frequency of misinterpretation by other people.)
• Some thoroughly familiar words such as bug that no one could possibly misunderstand or need any definition of, are easily misunderstood all the same, and might well need definition in context. "Bug" is a weak example in context, because apart from its precise entomological significance, it has several other meanings, but I trust the example will suffice without my labouring it.
• Even words that are not being used in narrow technical senses, but are widely understood in all their normal usages, might justify a link to avoid doubt or confusion in context. I mentioned a few in the past (uselessly of course; part of the culture of veto is to ignore what you find it inconvenient to inspect or impossible to controvert): head, head, head, carr, ooid sten, Sten , aeon mole mole mole, germ, germ, germ, bore, or bore. Anyone with the slightest experience in editing will recognise the value of such links in obviating ambiguity and reducing misunderstanding. It also can be of value when one wishes to emphasise the precise or extended meaning intended for a term that is in wide popular use, but may be misunderstood by readers who do not realise that the popular idea of a given word is not the correct one. For example infer. This last case is not as common as a non-mother-tongue Anglophone might think of course, for example it usually will be clear from context which of more than a dozen conceivable meanings of the word "set" I have in mind when I speak of the set of the hairs on your head.
• Sometimes a link is useful simply to show the reader that a block of material apparently missing from the current article does actually exist even if it is not often necessary for a reader to retrieve that material. Such a link also makes it easy for the reader to ignore the link, comfortable in the knowledge that the material was not accidentally omitted, or to read it, not for his own benefit, but to assess the basis or competence of the coverage of material. And of course, some readers actually will want or need to seek out such obiter dicta and who are we as editors, to demand that he go and explore his options, rather than simply clicking on a link? In such cases of course the name of the linked article or section need not resemble the visible text of the link, which can make it very difficult indeed to hunt for the correct article, always assuming that one exists.
• Very similarly, where in another article there is adequate coverage of material optionally of interest in passing, then linking to it in the current article might direct a user constructively, a user who had had no idea that such coverage existed. (Of course, of course! Any such reader must infallibly be extremely stupid not to have thought of it for himself, and you and I would not dream of any such naïve behaviour ourselves but it behoves us to be considerate of lesser minds as well.)
• As frequently observed as well, it is one thing to insist on linking a word wherever it appears, which is undeniably silly, in spite of the protestations of the rabid erasers of links, who seem to think it necessary to explain this over and over and over and dishonestly (I almost said stupidly, but my natural politeness forbad) attributed such intentions to anyone they disagreed with. Trolling, we call that where I come from, such discourtesies as attributing opinions such as " So you'd like to spell out acronyms and initialism again and again through an article" in lieu of honest and sensible exchanges of operative points. In a short article, such as a single page, linking to the first occurrence of the word certainly is quite sufficient as long as every use of the word is identical. Obviously, where the meaning or usage is unclearly but radically different, one might indeed link more than once, and not always to the same place. However, such cases are exceptional especially in a short article. It is however altogether different when the article is larger, even when the usage of the word is precisely the same. Although I would not demand that the author or editor apply this principle universally, one might for example elect to link any time the word is far enough away from the previous linkage that they might never be expected to appear on the same page at the same time in reasonable practice. Another, similar principle might be at least to link to the first occurrence of such a linkable word in each section. One of the favourite tunes on the harmonium among the de-linking marchers is the deadly irritation and confusion engendered among innocent readers by superfluous links. I propose that anyone who sees a particular, apparently identical, link not more than once per page or once per section, stands a fair chance of recovery without lasting trauma. Those who disagree are welcome to compile a dossier of contrary case studies.
• It is perfectly possible to have valid occasion to link more than once to the same article in the same sentence. For example the article might not have separate sections, but might deal with more than one technical term. If one is to use more than one such term in the same sentence in the linking article, then each term would appear as a link in that sentence, but the actual article linked to might not be the visible name of any of the links. A reader might elect to click on one or the other or both in turn, but since neither is an article title he would have great difficulty in finding either of them without the link.
• By this time the point should be no novelty to readers of the discussion, that it is nothing unusual for a user to enter an article far from the top. He might have entered the article at the top and then leapt to a far lower section, having consulted the table of contents. Alternatively, he might simply have linked from another article which directed him to the middle of the linked article. Such a user might find himself in real trouble, all for the lack of one evil, pernicious, destructive, redundant link in the article where he ended. Some people argue that this is gross overstatement, missing the point that even if it is gross overstatement, which occasionally it might actually not be, even a comparatively mild inconvenience among a comparatively small number of users is a high price to pay for omitting a link, let alone for refraining from removing a link that one does not recognise the need for.
• A serious no-no for link erasers is to remove a link on the assumption that you know what it is doing there, when you have not bothered to check, and in particular not bother to check that you understand just because you think you know what that link means or what some proper usage is. For example "America". Anyone who doesn't know what America is, is too stupid to read the Wikipedia, right? Well, there is room for a few remarks on that (you were way ahead of me weren't you?) The word America I would remind those who need reminding is used in many senses including geographical, historical, biological, political, and geological. And "Latin"; if you did not actually have enough Latin at school to know what Latin means, then go and read something else, right? But Latin too can mean many things in many contexts, and a link can be an efficient way, not only to clarify the intended meaning where that is open to doubt, but for any reader unfamiliar with either the material or the usage to find out as much as he needs in context.
• Red links, as already discussed, are of importance in highlighting concepts of importance, and either the absence of an article (or a redir or disambiguation) with the necessary title. I personally have deleted a few red links, but mostly I leave them when it seems that there is any reasonable hope or prospect of the deficiency being made good.
• Ref links. Almost forgot. Strictly for the benefit of naive users; if you are not naive, don't read this. Naive authors and editors regard it as absolutely necessary to supply citations for every comma at least, and they approve it an automatic hanging offence to cite another WP article. And a link to another article in support of a statement is patently in internal citation, right? Not necessarily. If the point you are supporting is not one that your article is calculated to establish, then it is altogether proper to link instead of citing as long as the linked article established the point by any approved means whatsoever, including proper citation. In fact, it even might be OK if the linked article fails in this duty; it may be your job to tidy it up or it may not, but it definitely is not this article's job to do it. Suppose you go all noble and insist on supplying citations and material on a subject alien to you, next thing you know is that nasty people with proper competence in the subject matter descend upon you with wrath and contumely, their kindest remarks being that you are a century and a half out of date and that your matrix division wouldn't even do for accountancy. So if you are not an authority on the subject, and even if you are, but your intended readership is not intended to be don't include explanations and citations that belong in the linked articles rather than your own.

Other forms of trolling include arguing that because some people do agree with some other people, that there is effectively no dissent and no improvement to the current situation either is necessary or can be countenanced; or to claim that dissent on this page involves just a few editors who would like to see denser and less specific linking in WP articles. Exactly why anyone would want to see that, anyone who could read and understand what he reads, such experts on the subject can afford to be silent upon because they are so strident about much that they really should be silent upon for fear of revealing the depths of their competence and concerns. But for such people trolling generally is easier than arguing, it would seem.

Having settled that to everybody's satisfaction, I invite anyone to supplement or amend, as constructively as may be, the list of points here foregoing, or to assist in formulating it suitably for documentation to assist novices or editors in doubt. JonRichfield (talk) 13:39, 24 October 2012 (UTC)

I updated the work-in-progress text at the top of the section to sift your big list down into small bullet points. Boundlessly (talk) 22:37, 24 October 2012 (UTC)

WP:TLDR. Can you summarise that gigantic statement in a short para, please? No one is going to read it. Any changes that are proposed to the guideline should be clearly proposed (usually with strike-throughs and some other mark-up for replacements, and discussed. Again, there's no convincing argument for any change, as far as I can tell. Tony (talk) 00:17, 25 October 2012 (UTC)

It's worth reading, and should not be dismissed in that manner. Jon makes some good points. --Ckatz^chat_spy 08:24, 25 October 2012 (UTC)

Jon, the quotation I have heard is "those who try to please everyone please no one." I am not commenting on your wall of text in detail (though I will on your talk page if you would like) but I fear that your suggestion that linking should, or even may, be used as a disambiguation method is, to say the least, unwise.

Now as to links, it is fairly clear that for a given user on a given day and a given article there is an optimal link density that corresponds to exactly those items which the reader in question would want to visit, or perhaps be made aware of the existence of - though one could argue even this to the nth degree. When however we aggregate, effectively, all users, all days and all articles it is equally clear that we have a somewhat more blurred situation. So what principles can we enunciate that help us decide where the areas of good linkage and under and over linkage occur. Well there are some things we can say.

1. While serendipity is a wonderful thing it is not the purpose of our wiki-links.
2. Wiki-links here have two main purposes, to lead onto related articles of interest, and to provide a gloss (the conflict between these has always troubled me).
3. When the density of links is so great that links do not stand out, or run into one another, there are too many links.
4. When links to completely irrelevant articles exist they are considered a detriment.
5. Decreasing link density without decreasing navigability is a good thing.
6. Conversely increasing navigability without increasing link density is a good thing.

In general someone reading an article where there are unfamiliar terms will need to read most of the article. There may be exceptions, and here I think repeated links would not be a problem. For example an article on a foot and mouth in the UK broken down by counties, might reasonably have links to DEFRA (or MAFF) within each section.

These are the reasons behind the MoS as it stands. I would not be happy, and I doubt anyone here would, with overzealous removal of links, however those who believe the MoS is over strict (it is after all a guideline) probably have not seen many of the over-linked articles that used to abound.

We should be very careful, as careless wording here could lead to all-blue tables, infoboxes (infoboxes are generally over-linked as it is) and captions.

Rich Farmbrough, 02:39, 25 October 2012 (UTC).

Long reply from Jon Richardson

Hello Rich and thanks for your comments. I am glad that our versions of the quote do not conflict. Some of our other points however do, sometimes to a disconcerting extent. For example I cannot make much sense of: "...that linking should, or even may, be used as a disambiguation method is, to say the least, unwise." I certainly never suggested anything at all like replacing redirections or disambiguations with links. I do not say that nothing of the kind could be possible or even desirable, never having given it thought, but the point of both of those facilities as well as the point of links in general is to direct users as smoothly and efficiently as may be, to material that they want, or would wish, to locate. Furthermore it should as far as might be, not direct them to items that they do not wish to locate, not in the relevant connection anyway. In that respect linkage is at the very least an inescapable disambiguation function no matter how one wishes to regard it. Furthermore, the disambiguation function in Wikipedia as she is spoke at the moment is a stamping ground for rampant wikilawyers who vie with each other in deleting anything that they regard as explanatory or as article titles that they happen to think don't need disambiguation, no matter what anyone else thinks. In short, vandalism. Exactly why in such a climate it should be regarded as heinous for a link to spare a reader a visit to a disambiguation article, is unclear to me. I certainly have said nothing that suggested to me that I might have been recommending anything so pernicious. Could you please clarify the point in case I have misunderstood you? You also said: "... for a given user on a given day and a given article there is an optimal link density that corresponds to exactly those items which the reader in question would want to visit, or perhaps be made aware of the existence of..." Actually I tend to fall off the bus even before that point. I don't like the idea of thinking of links in terms of any density whatsoever; thoughtful inclusion and crafting of links is not a matter of titration — it is a matter of articulation. A link fits or it does not fit. It may fit for one reason or it may fit for another, but it does not earn its place by being the third rather than the fourth link in an article of 400 rather than 300 words, or a sentence of eight words. There might reasonably be no links in such an article whatsoever (though I must say that I would regard the legitimacy of inclusion of an article at all with deep suspicion if it were to link to no other material). But one way or another its merit is not to be decided by the counting of links. You then remark: "When however we aggregate, effectively, all users, all days and all articles it is equally clear that we have a somewhat more blurred situation." Even referring to a thought as disconcerting as link density being a merit or demerit, it is an uncomfortable idea that we might aggregate our works according to time, user and subject. We (and I bet this applies to you as much as to anybody) do our best to produce articles individually or in combination that can serve the interests of users and subject matter in such a way that any reader can as naturally as possible find what he wants to read and skip what he wants not to read, all this in his own individual context of course. We are not discussing a matter of blending, such as paints, but of structure, such as in building arches or engines. Your very next sentence reads: "So what principles can we enunciate that help us decide where the areas of good linkage and under and over linkage occur." Is it just me, or does that really suggest that you too think that structure is what counts, rather than density? "While serendipity is a wonderful thing it is not the purpose of our wiki-links." Errr... really? How did serendipity as a purpose get into this? Chris, any worthwhile work of reference, even a halfway decent glossary, fairly exudes serendipity. This takes me back to my childhood, when I first taught myself to read a dictionary. A lot of linking is possible in reading a dictionary, trust me! Any functional link increases the scope for serendipity for any reader with a brain halfway alive. You cannot prevent it. You also have remarkably little scope to force it or fake it. Forget serendipity. It will take care of itself. Nothing in what I have said so far seems to me to have any connection with your remark. If I am overlooking something obvious, please correct me. "Wiki-links here have two main purposes, to lead onto related articles of interest, and to provide a gloss (the conflict between these has always troubled me)." Conflict? Here I must request you to explain. It seems to me that sometimes a link is little more than a lead, and on other occasions, especially, but not only, if you use the hovering facility, it can indeed be a gloss. But a conflict? Why? And if there is a conflict what does that have to do with the intelligent structure and placement of links? Furthermore, those are not the only purposes at all, as I already have pointed out in my list. Right? "When the density of links is so great that links do not stand out, or run into one another, there are too many links." Density again! Links do not have to stand out. They are not selling anything. A link that you do not encounter in reading, or do not notice, will seldom be a link that you are in the slightest interested in looking up. We link because we have encountered something that requires some sort of reference, not because we have seen a flash of blue, nor because it is already two sentences since you last linked. In almost any highly technical field (and not all Wikipedia articles are groupie gushings about their most recently discovered garage bands; try some of the physics, chemical or medical articles; or even some of the philosophical or philological articles) it is perfectly possible to encounter two or even three or more words in succession, each one relevant and necessary in context, and each one justifying its own link to its own article. It might be entertaining to see how long a sentence one could construct in which it would be justified to link every single word. The point is that one does not link, or omit to link, a word on the basis of which words happen to be its neighbours; one links because the word or phrase itself needs linking; sometimes it needs it in context and sometimes it needs it absolutely, but if you take your job as an editor seriously, you don't link it if it doesn't need linking and you do link it if it does need linking. Sometimes of course one links a whole string of words as a single link, but I assume that that point is unobjectionable in this context; I am referring to when adjacent words need separate linking. "When links to completely irrelevant articles exist they are considered a detriment." Anything completely irrelevant in an article, whether a link, or a linked item or whatever it might be, is necessarily detrimental. This is not a point at issue. Not at this end of the pen, nor in anything I have written so far anyway. "Decreasing link density without decreasing navigability is a good thing." A curious sentiment if you will forgive my saying so. If removing a link does not increase navigability, then why would you want to remove it? And if navigability is any criterion in such a consideration at all, then surely for a word that may be justifiably linked at all to appear on a page where it is not linked, the original link having scrolled off the top or the bottom, is to reduce navigability? For a word not to appear linked in a section that has been linked to from an outside article, because that word already has been linked five pages back, that does not reduce navigability? One of us will really have to do better I leave you to decide which. Then there is the question that density is largely a matter of taste at best. I have actually encountered people, highly competent, highly literate people, who actually would prefer every occurrence of a technical term to be linked if it is not dealt with in the current article. As a matter of taste I personally think it is quite enough if every unlinked occurrence will always (well... almost always!) be within clicking distance of a visible link, but bearing in mind the total innocuity of the occasional redundant link, I would not be inclined to fight that one tooth and nail. But perhaps I have misunderstood you? "Conversely increasing navigability without increasing link density is a good thing." Similarly, don't you think this statement is a bit vacuous? I think, and I am fairly sure that you agree, that increasing navigability is almost unconditionally desirable. Doing so without any cost in redundancy, though less important, certainly strikes me as desirably elegant, but not necessarily beneficial in any other way. What does link density cost anyway? If I insert a link in an article of any general interest and a single reader uses that link to any substantial benefit, then that link has been paid for in my estimation (always assuming that I am aware of so hypothetical a circumstance), even if day in and day out thousands of people read the passage without clicking on it whether they need to or not. Again, would you care to explain in material terms rather than as a matter of taste just why this point is the least important or otherwise? "In general someone reading an article where there are unfamiliar terms will need to read most of the article." By no means even nearly in general. Many people read only part of an article, some of them only the lede. More importantly, many people link into the middle of an article from another article. In such a case I am sure that it is the exception rather than the rule that when one has linked to an article that deals with a technical point mentioned in another technical article, one actually reads the entire linked article. Often it is just a bit of clarification before one continues from the point from which you had linked. Notice that such a smooth and convenient continuation is considerably hampered if you first had to hunt around for the original link. Even if readers do begin at the beginning, continue till they come to the end, and then stop, they might well skip the first few links and only do a doubletake a few pages down, saying the likes of: "Surely! That can't be what you mean can it?" And then they might click on the link, if it is available anyway. "There may be exceptions, and here I think repeated links would not be a problem. For example an article on a foot and mouth in the UK broken down by counties, might reasonably have links to DEFRA (or MAFF) within each section." Very true, and a very pertinent example. Now let's see your equally pertinent example in which, in a perfectly normal and perfectly rational and readable article, a few extra links (more than I for one would have put in) renders the article unreasonably unreadable. "These are the reasons behind the MoS as it stands. I would not be happy, and I doubt anyone here would, with overzealous removal of links" Ahem! Who says what is zealous or overzealous? I think I remarked somewhere that the mere fact that someone has seen fit to put in a link that makes any sense whatsoever implies that its removal requires positive justification. The fact that you personally do not happen to feel like using such a link is not a positive justification. The fact that someone else might be expected to wish there were a link is a positive justification for leaving it in place. "However those who believe the MoS is over strict (it is after all a guideline) probably have not seen many of the over-linked articles that used to abound." You shock me Chris! Have you no sensitivity? Even at this end of my monitor I can detect Tony thrashing around in horror, his green gone pale at the blasphemy. A guideline? Since when have guidelines been chiselled in stone like holy writs such as this one...? And anyway what has strictness to do with it? It is perfectly possible to over-link (or under-link) while adhering to the letter of the MoS. I say again: over-linking and under-linking are not matters of density and not matters of rule; they are matters of functional sense and articulation. You will forgive me if I forbear to shudder at the idea of the over-linked articles that used to abound; I have actually seen examples of over-linking, and they occasioned less effusion of blood than the last time a cat bit me. Tony would not have liked that cat; it was a lot harder to unlink than he would have approved. "We should be very careful, as careless wording here could lead to all-blue tables, infoboxes (infoboxes are generally over-linked as it is) and captions." Thin end of the wedge is it? The most important point in this respect is the appearance of running text, because anything that interferes with the rhythm of reading text may interfere with comprehension. And naive readers, meaning those who are in their first 10 minutes of reading documents containing highlighted items, might well find the highlighting distracting, though that is a trivial price to pay for such a powerful facility, especially if it could be supplemented as I already have indicated. But all-blue captions, tables, boxes etc (should they occur; I have seen few serious cases, and in every such case it was fully and painlessly justified) do not curdle my blood. They are no problem because they do not get read in continuous, progressive mode as running text does. In a table or box it would be an unusual link with any pretensions to meaning, that one could justify removing just because the table is beginning to look a little blue. For example, suppose there is a table of biological binomials pointing to the various articles dealing with the respective species, it would be a gross error to remove those links even though each item in the table would be either blue or red. For example, in a table of just a couple of hundred, the linked articles might not be the same as the binomials in the table, making it very difficult for the reader to look them up elsewhere in the WP. Of course, there are other ways of doing it, putting in links on asterisks or the like. Just think of the aesthetic relief of link-haters in not having to see blue! Much more important than the extra burden on the reader trying to guess where the link is, no? The bottom line Chris, remains the question of function and practicality, and for that matter what Dr Johnson would call good taste. (Well not actually; he called good taste whatever he bloody pleased, and I am sure that he would not regard modern English as good taste.) The problem is that there is a strong movement afoot to gratify a particular pointless whim to prune blue in favour of dark green, irrespective of meaning or function or even tolerance for the needs of the reader. It is about time that some of us started doing something about this, and I think that if enough of us stop and think about it, as opposed to trolling the whole WP body for fun, we should be able to do better than this. JonRichfield (talk) 13:32, 26 October 2012 (UTC)

Hi Jon, I have collapsed your reply partly because it is so long, but mainly because the indentation is wrong. To avoid confusion, and boring other contributors, I will reply in a collapsed box also. Rich Farmbrough, 19:58, 27 October 2012 (UTC).

Rich's reply to Jon, and Jon's rejoinders

First can I request much shorter comments. There is no way most editors are going to read that much text, as I think others have already commented, especially when it does not need to be that long. This means your efforts are going to waste. Secondly you seem very excised about this, if you have been in conflict over linking, please drop a note at my talk page and I will be happy to see if I can help resolve the conflict. Thirdly my comments were not, by and large, (as I made plain) an answer to yours, but general comments. Sorry if I made some acerbic remarks seem personal; not at all the intention. JonRichfield (talk) 08:45, 28 October 2012 (UTC) I understand now that perhaps the point you were making about "America" was perhaps that assuming a link to America was about the entity in the article "United States" is dubious. This is of course correct when you see a link such as America. However I am not sure that there is conflict over links that merely appear to be to these types of entity. Let's get some clarity about one thing: In spite of the aspersions cast by the Tony faction, I most certainly disapprove of any linking, overlinking or not, that is futile or otherwise counter-constructive. This includes the (hypothetical) likes of "the Monroe Doctrine was America for the Americans". (Ummm... actually not such a good example; bearing in mind that both of those "Amer...s" were somewhat reasonable to elaborate on if they had not been discussed suitably in the associated text. But you know the sort of thing I mean.  ;-) ) But, as I said elsewhere, at some length, there are cases where links to such words could be justified and it is up to the linker to make sure that they are linked appropriately and only appropriately. For novices this is not easy and seasoned editors naturally should assist. But assistance goes further than arbitrary removal of all but one link in one article, as you plainly do accept, but not everyone else seems to. JonRichfield (talk) 08:45, 28 October 2012 (UTC) When you say however "might justify a link to avoid doubt or confusion in context" - that is disambiguation. If you write a sentence where it is unclear in which sense you mean a term, then the resolution is to write the sentence properly. It is quite likely that the term will deserve linking, but that is an orthogonal issue, and by no means certain. You also said: "... for a given user on a given day and a given article there is an optimal link density that corresponds to exactly those items which the reader in question would want to visit, or perhaps be made aware of the existence of..." Actually I tend to fall off the bus even before that point. I don't like the idea of thinking of links in terms of any density whatsoever; thoughtful inclusion and crafting of links is not a matter of titration — it is a matter of articulation. A link fits or it does not fit. Just because you don't like to think about link density, it doesn't cease to exist. Moreover simply because articles are linked by intelligent entities with no regard for density, it does not mean that density is a wholly useless measure, in the same way that we do not write to readability indexes, but they are not wholly useless. Rich, this is not at issue. I did mention such concepts as sentences in which every word could legitimately be linked, but they were largely academic, not to say pathological, examples. I have encountered adjacent links to separate articles, and it might be argued that one might bend over backwards to avoid these, if necessary splitting sentences and references etc, but in practice the only places where there are legitimately densely crowded links are where there are legitimately densely crowded concepts. Ergo, it is not something to worry about in principle and in practice, when such occasions arise, they either can be tolerated, or if too offensive, the can be edited to satisfy both classes of interested parties. No? I still reckon that the point of links is articulation and that when dense articulation is necessary, dense links should follow. Sometimes dense articulation amount to bad writing, then the problem is to improve the writing. Improved linking should follow automatically. JonRichfield (talk) 08:45, 28 October 2012 (UTC) This is what my following comments ("When, however, we aggregate...") was explaining, that it is certainly not valid, (and as far as I know no one holds that it is) to use link density as a pure number to attempt to draw a distinction between over-linked and non-overlinked articles, though a correlation doubtless exists both in the space of all possible linkages, and in the space of all actual linkages. In terms of serendipity, this has been adduced many times over the years as a reason to allow non-germane links. (In the early days we also had a "meta-data" thang going on, which I was (wrongly) wedded to for some years.) Hm... Sounds intriguing. I hardly dare to ask... JonRichfield (talk) 08:45, 28 October 2012 (UTC) * "Wiki-links here have two main purposes, to lead onto related articles of interest, and to provide a gloss (the conflict between these has always troubled me)." Conflict? Here I must request you to explain. It seems to me that sometimes a link is little more than a lead, and on other occasions, especially, but not only, if you use the hovering facility, it can indeed be a gloss. But a conflict? Why? And if there is a conflict what does that have to do with the intelligent structure and placement of links? Furthermore, those are not the only purposes at all, as I already have pointed out in my list. Right? Because there is one link mechanism, we can link English or English, which provides the author choice of one or the other or neither, but not both, and provides the reader with only what the editor has decided is appropriate. That however is very much out-with the scope of this discussion Ah yes!!! I see. Right! This is a problem that I have encountered a few times. Thank you. I hardly dare raise the subject here and now, but my reaction is that we need the likes of a multiple link; something with the relation to a link, that a disambiguation has to a redirection. I wonder. Maybe, instead of tinkering with the link mechanism as such, we could have a new class of article (probably rarely used) A plexus article? Anyway, it would fit in when we have a need for a click to direct a troubled reader in either a choice or a sequence of links. This could raise a large topic methinks. It also could remove half the sting from the link topic if we could deal with it smoothly. Maybe a suitably formatted type of pop-up gloss note in which overlinkers get free rein if needed. Something that explains lede-style or hatnote-style what the theme is, and supplies links and introductory material that otherwise might clutter the text for people that are competent in the subject matter. JonRichfield (talk) 08:45, 28 October 2012 (UTC) Your list, unfortunately, looses much of its impact because it rambles off. However most of your suggestions for reasons for linking are not ones I would agree with. We could perhaps break that out into a separate discussion. "When the density of links is so great that links do not stand out, or run into one another, there are too many links." Density again! Links do not have to stand out. They are not selling anything. A link that you do not encounter in reading, or do not notice, will seldom be a link that you are in the slightest interested in looking up. You appear confused. If links do not stand out how do people know they are links? Maybe one of us is confusing the other. If the links are the normal links we use in WP, then the link is visible. Period. If there are several immediately adjacent links (as discussed and mildly deprecated above) they might look like one long link, but as soon as your cursor reaches a link, and before one clicks, it becomes plain which word(s) are referenced. If there merely is a clutter of non-adjacent links, one will not be concerned till one reads that text, at which point standing out is no problem. If it is blue (or something similar) then it is a link and until you are looking directly at it, you don't care whether you notice it or not. JonRichfield (talk) 08:45, 28 October 2012 (UTC) In almost any highly technical field (...) it is perfectly possible to encounter two or even three or more words in succession Of course, and this is the field I edit in widely. It is however rarer than you might think, for adjacent technical terms are likely to either be a list, where the commas separate the links, if barely, or adjective-noun groupings, which generally means there is a more specific link. Sure. This is fully compatible with what I said. I did not say that it is an everyday event. It still is a real need, and closely related to the gloss problem IMO. JonRichfield (talk) 08:45, 28 October 2012 (UTC) "Decreasing link density without decreasing navigability is a good thing." A curious sentiment if you will forgive my saying so. If removing a link does not increase navigability, then why would you want to remove it? Joe Bloggs' company was named after Bloggs, and Bloggs said in Joe Bloggs house magazine - "Here at Joe Bloggs we are delighted to use my name Joe Bloggs to show that I, Joe Bloggs stand behind all our products." I would want to remove most of these links. I'm not sure if you missed the difference between "not decreasing" and "increasing." Actually, I agree. I cannot remember seeing such an egregious example in practice of course, though it still would not matter a scrap if my proposal for link hiding had been implemented. I did not miss the distinction. At the risk of repetition: If removing a link does not increase navigability, then why would you want to remove it? Your Bloggses frankly do to my mind decrease navigibility (very slightly) because they imply a need for separate attention, as in my earlier "head head head" example. But in terms of material damage Bloggs' egocentricity is venial. I might edit it if I encountered it, but then again I might not. If I found someone removing gloss-effective links, I certainly would revert. JonRichfield (talk) 08:45, 28 October 2012 (UTC) The argument about "one link per screen" or "seventeen screens" or whatever is I think a bit of a red herring. There is no absolute prohibition on relinking. True, but we have some active and compulsive link strippers who use admonitory guidelines for self gratification. JonRichfield (talk) 08:45, 28 October 2012 (UTC) ...What does link density cost anyway? ... What indeed? JonRichfield (talk) 08:45, 28 October 2012 (UTC) * "In general someone reading an article where there are unfamiliar terms will need to read most of the article." Here I mis-spoke and should have said "Will start reading at the beginning." Of course there are many articles where we can all jump in in the middle, but they are far less likely to link to [[ * "These are the reasons behind the MoS as it stands. I would not be happy, and I doubt anyone here would, with overzealous removal of links" Ahem! Who says what is zealous or overzealous? The community. If you have an issue, take it up with the editor, on the article talk page, at 3rd opinion, at MoS, and if you fail to agree there are many other fora. I think I remarked somewhere that the mere fact that someone has seen fit to put in a link that makes any sense whatsoever implies that its removal requires positive justification. We use WP:BRD. You use BRD, I use BRD, and plainly you use it in the manner intended. Someone like you jumps in and removes a link (or otherwise alters or deletes or inserts text) and I disagree, so I rv and ask you to raise it on the talk page. Possibly I even find that you already have explained there, etc... I very often post explanations when I delete text. No problem. After due discussion we part friends with either your or my or our edits established till further notice. We are cooperative "most interested parties" or the like. In the light of the frequency with which I find people who delete substantive material without explanation or apparent reason in context, and start edit wars if anyone counter-reverts, ignore requests to talk etc, I suspect that you and I might be in the minority. One rule-protected vandal can create more harm than ten editors can deal with, so in relying on BRD it is necessary to be alert in scanning your watchlist daily. Yes? JonRichfield (talk) 08:45, 28 October 2012 (UTC) * "We should be very careful, as careless wording here could lead to all-blue tables, infoboxes (infoboxes are generally over-linked as it is) and captions." Thin end of the wedge is it? The most important point in this respect is the appearance of running text, because anything that interferes with the rhythm of reading text may interfere with comprehension. And naive readers, meaning those who are in their first 10 minutes of reading documents containing highlighted items, might well find the highlighting distracting, though that is a trivial price to pay for such a powerful facility, especially if it could be supplemented as I already have indicated. The point is if, in an infobox, every name and/or every value is linked, it is not clear that anything is linked. It is all very well to scoff at the "naive user" but most of our users are naive, and they all (we all) were once. I take your point, and in open text it would matter, as Messrs Blogss illustrate, but I think you might struggle to find a case where that is a real prospect in a caption or species table or the like. JonRichfield (talk) 08:45, 28 October 2012 (UTC) And irrelevant links (which you agree should be delinked) do indeed break the flow for non-naive readers. Moreover experienced users may well, on seeing a re-link, look for a difference between the re-linked term and the original. These things are often non-trivial. Certainly. Common cause. But these amount to errors or even actual incompetence. They resemble any other dud editing and should be treated accordingly. Gosh, how often don't we find that someone (even yours truly!) has created a link without first checking the nature of the article linked to? And sometimes One cannot even find the right article. Once or twice I even have been driven to create the article! JonRichfield (talk) 08:45, 28 October 2012 (UTC) The problem is that there is a strong movement afoot to gratify a particular pointless whim to prune blue in favour of dark green, irrespective of meaning or function or even tolerance for the needs of the reader. I don't think there is such a movement, it seems as if you have been subject to a particularly vicious unlinking incident in your formative years, if so that is the thing to address. More a matter of encountering some attitudes and agenda in this page (not yours btw!) JonRichfield (talk) 08:45, 28 October 2012 (UTC) Rich Farmbrough, 21:39, 27 October 2012 (UTC).

Comment: Perhaps rather than prescribing, proscribing, providing extensive examples, or declaring whether the current wording is lacking or "just right", we get to the heart of the issue. It's that, as mentioned above, some editors seize on one line and elevate this guideline to an inflexible policy. Too many editors seem to see WP as a work of science, not as a work of art. As a work of art, there likely can never be a one-size-fits-all answer that covers every conceivable question. Let's allow good-faith editors to create pages useful to readers, with links designed to enhance a reader's overall time spent in the encyclopedia. If a consensus of editors feel that a given article is "over-linked" or "under-linked", then that article will be modified, as per any other consensus. Not sure of proper wording within the guideline, but it probably should be deliberately vague, perhaps explicitly incorporating the sense of what I've said here. (IMHO). --Chaswmsday (talk) 16:19, 25 October 2012 (UTC)

I could go with a lot of that Cday, the main problem is that we are dealing with people who appreciate the virtues of freedom as long as it is freedom for everyone else to toe their line; and as long as there is material in the guidelines that they can twist to their taste, guess how they will twist it. I think that boundless and a few others would agree with your general idea, just as they would agree that everyone can phrase text within reason to their own taste, but some other parties are thinking hard thoughts... Only kind they think, apparently. JonRichfield (talk) 13:40, 26 October 2012 (UTC)

I think that is pretty much where we are now. As well as the disclaimer at the top this section begins with "generally". I don't see anyone being persecuted for over-linking, or indeed under-linking. Rich Farmbrough, 19:35, 27 October 2012 (UTC).

Rich --

Yes, there are a number of truly vicious unlinkers out there (I have been so persecuted -- as well as several of the articles I take care of). Unlinking is particularly destructive when a term has multiple meanings, and a vicious unlinker removes the links that help disambiguate, so it's no longer clear to a new reader which meaning is being used where.

I've updated the proposal at the top of this section to reflect JonRichfield, RichFarmbrough, and Chaswmsday. Further comment welcome.

Boundlessly (talk) 03:24, 28 October 2012 (UTC)

Please bear in mind that most editors will regard this "working text" as bloat. They won't see the point. Nor do I. Tony (talk) 05:48, 28 October 2012 (UTC)

It it were bloat, then there would have been no such long-drawn futile debate with people who can't see points but can repeat counterfactual allegations and insinuations indefinitely. Furthermore, the term "bloat" implies vacuous content. There isn't any. Nice try. Must try harder though. Bad luck! JonRichfield (talk) 09:38, 29 October 2012 (UTC)

Not to mention the contrast between two sucessive sentences up thread here, first to state a refusal to read a previous set of comments, and then a statement that "there's no convincing argument for any change" (even in the comments that were stated to be unread). Interesting way of doing business. Boundlessly (talk) 00:31, 30 October 2012 (UTC)

Some of you will be delighted to find that I have discovered an overlooked (by me anyway) reason for linking. Someone read the word "tepal" in a caption in Floral symmetry and politely replaced it with petal, suggesting that that was what was meant. I proceeded to revert, and in fact insert the link into that caption, though there also is a link elsewhere in the text. Had it been linked in the first place, the confusion could not have arisen because the link, apart from highlighting the deliberate intention, would have made it plain as soon as the editor clicked on it, as a responsible editor would naturally do if it were linked, that the term was both deliberate and correct and referred to the correct article.

Now, a point for Tony's delectation: by the time I had clicked "Save", another editor had beaten me to the exact same edit; so it isn't just me, Tony! Some other people also can see points and tell them from bloat and blue text. All that happened within about two minutes from the original edit btw! Note too that in the same article "petal" and "sepal" are not linked and I have not subsequently linked them because I doubt that they will be subject to the same confusion, but I might well have done so. I don't know how you folks would categorise such a usage. JonRichfield (talk) 09:38, 29 October 2012 (UTC)

Interesting. The very situation that started this discussion arose out of links that perform a highly similar error-checking and scaffolding function. The "vicious" unlinkers (JonRichmond's apt term) refuse to believe that links can serves such a function. I'll wait for similar breakage to occur before pouncing. Boundlessly (talk) 00:31, 30 October 2012 (UTC)

I will not pour cold water over those who feel this to be a topic upon which to base a PhD thesis. Whilst it may be true that linking is theoretically more an art than a science, real life wiki-linking is very frequently neither scientific nor artistic. There are some good examples used above, but these were probably very carefully chosen to illustrate the point of why linking is desirable. However, I feel that this is close to the edge of the 'what should not be unlinked' part of the spectrum, and are also unlikely to be unlinked even by "vicious unlinkers". OTOH, linking for the sake of glossing ought to be minimal; if a truly unfamiliar but germane term needs to be defined or clarified, we should explore the possibility of glossing in-line, as a footnote or via wiktionary links. I come across every day examples of 'glossing' would be more appropriate to a child of 6 years of age; also, I have already mentioned above there is no prohibition to linking of two or more instances of a germane word if they are sufficiently far apart, and a 'quasi-exemption' for links that appear within infoboxes. I want to keep it brief, so I'll just add that if serendipity is a quality which lacks in your life, the Metawiki software already provides you with this. I'm already in bliss being able to click on that non-stop. -- Ohconfucius ^{ping / poke} 15:09, 30 October 2012 (UTC)

Well OC, skipping fastidiously over your cold water, I congratulate you on having enough time to pursue serendipity by riding the random button. The proportion of really interesting articles is frustratingly low. I charitably assume that they are harder to write, otherwise I would have to ascribe embarrassing average levels of various enviable attributes to our sources. However, what has often been of greater value in the pursuit of serendipity, has been the "what links here" button. It is more rewarding to those with an interest in logical connections and connectivity. Good hunting if you try it! JonRichfield (talk) 19:44, 30 October 2012 (UTC)

Seeing that doesn't bring you enough or the right sort of serendipity, perhaps you should try your hand at Online gambling. ;-) -- Ohconfucius ^{ping / poke} 01:53, 1 November 2012 (UTC)

While the start up text by Boundlessly was a reasonable one - as I stated earlier - I think the current version of the working text is not as much. Namely, if a word is used with different meaning then those meanings should be clear from the text itself - this does not exclude the possibility of linking both instances, but such clarification should not be deferred to the linked articles. Also aggressive expressions like "unlinking assault [being] uleashed by..." are undesirable (and a bad sign) - Nabla (talk) 13:56, 4 November 2012 (UTC)

"if a word is used with different meaning then those meanings should be clear from the text itself" -- concur, but not always possible. The alternatives are some klunky explanation at each point a term is used, or links -- there are certainly cases where the multiple links are the less intrusive way to embed the information.Boundlessly (talk) 21:17, 7 November 2012 (UTC)

Arb Break

I see people are trying to add the above text. While I might be inclined to agree that the copy of that text has been stablized within some form of consensus, there is no consensus to outright add that text to the MOS yet. That needs to be achieved before changing from the status quo. --MASEM (t) 14:18, 12 November 2012 (UTC)

In response to this diff, where Boundlessly claims that those wishing to not add it have been "outvoted", I will remind that consensus is not just a vote - sure, if there were only one or two voices against adding out of 20, there's consensus to add, but the discussion above is far from that type of clarity. There might be an edge towards the consensus to include but the opposition to adding it is signiifcant enough to currently make this a "no consensus" situation and ergo we maintain the status quo. --MASEM (t) 14:38, 12 November 2012 (UTC)

Even if I agree with some of the changes, it's clear there's no consensus for all the proposed changes. Yet. — Arthur Rubin (talk) 15:14, 12 November 2012 (UTC)

I've stayed out of this debate so far (mostly because it's extremely difficult to follow), but it looks like it might help to have another voice of dissent. The proposed text, to me, seems very problematic; it's clear that the opinion of the Wikipedia community is divided on this issue, so the guideline should be a compromise between both viewpoints – which it currently is. Boundlessly's proposed text is decidedly pro-linking, especially that last bit about "it is up to the unlinker to make an affirmative showing that multiple links are confusing or otherwise functionally deleterious". Adding that to the guideline is essentially giving the pro-linking crowd the ammunition they need to block any attempt by anyone to fix overlinked articles, on the basis that the unlinker can't prove that the links are detrimental.

Okay, so you could argue that the text can be amended, and the bias can be dealt with. But what's the point? I don't see a problem with the guideline as it stands. I might agree that the most controversial sentence – "generally, a link should appear only once in an article" – could be replaced with something like the corresponding line from the proposed text above – "links may be repeated whenever helpful to a reader, especially in longer articles, in subsequent sections where the link may have specific relevance, and in infoboxes, tables, image captions, footnotes, etc" – but even this seems unnecessary. The problem, as I see it, is not with the guideline itself, but with the so-called vicious unlinkers (and, indeed, the vicious overlinkers) who twist the guideline to their own ends, or ignore the word "generally" and cite it as if it were cast-iron policy. That's the problem, and rewording the guideline isn't the solution.

Anyway, I agree with those directly above me. It's impossible to discern any kind of consensus on this talk page, and personally, I think that this sort of change requires very clear consensus – two or three very vocal supporters isn't good enough. DoctorKubla (talk) 17:21, 12 November 2012 (UTC)