Wikipedia talk:Manual of Style/Linking

Relax duplicate linking rule

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.

This RFC was rather well laid out. It asks a number of specific proposals:

• Status quo. vs Relax somewhat. There is rough consensus that there needs to be some relaxing of the guideline.
  

The next two questions looked at specific changes to be made.

• Once per section. There was no consensus on once per section. The arguments were good, but the responses were split pretty evenly.
• When far enough apart. There is consensus for allowing when far enough apart. Some of the oppose arguments, which was the minority opinion, even included that this is already possible with the existing wording. AlbinoFerret 17:34, 31 May 2015 (UTC)

---

Many editors feel that the rule about duplicate links is too strict as currently written:

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, hatnotes, and at the first occurrence after the lead. Duplicate links in an article can be identified by using a tool that can be found at User:Ucucha/duplinks.

I would like to poll the attitudes of the MOS community on this topic, so that I can draft a suitable proposal to change this rule. --Slashme (talk) 21:24, 21 February 2015 (UTC)

Notes: I've canvassed both proponents and opponents of duplicate links, from the discussions that I found in the archives. It's possible to support or oppose more than one of the options below, as I have done.

General question

A strict interpretation of this rule implies that a link to an obscure term that appears in section 19 of a long article should be removed if it occurs in section 2 as well. If you support this interpretation, please explain why.

Support status quo:

• I think the actual cost of this is minimal, but I feel that most articles are more threatened by overlinking (where multiple low value links create a sea of blue, making it harder to identify valuable links) than by underlinking. The "rule", such as it is, can already be interpreted to include extra links in situations where this will be helpful to a reader, but to relax it or delete it would be to encourage the continuation or worsening of overlinking. --John (talk) 21:44, 21 February 2015 (UTC)

Comment: the current rule as written does not allow a term to be linked twice in the text of an article (excluding the lead section).--Slashme (talk) 22:12, 21 February 2015 (UTC)

• Ditto John, and see my fuller response under the When far enough apart option. Cheers, Ian Rose (talk) 23:27, 21 February 2015 (UTC)
• I feel the current rule is loose enough to allow for exceptions where truly warranted, but in my experience, as John noted, overlinking is a far more prevalent problem than underlinking. Nikkimaria (talk) 02:13, 22 February 2015 (UTC)
• Ditto John. Hchc2009 (talk) 08:41, 22 February 2015 (UTC)
• I think the current guideline is already relaxed enough and leaves room for personal discretion. It does indeed allow for links to be repeated in some circumstances. As the current guideline begins, "Generally..." Jayaguru-Shishya (talk) 11:12, 22 February 2015 (UTC)
• I have been invited to participate in this discussion on my user talk page, presumably because of my participation in prior discussions on point. My current thinking is succinctly embodied by Jayaguru immediately above. I think the current phrasing already permits editors to use a modicum of common sense and to resolve particular instances of repeated links to key topics in very long articles by compromise and article-level consensus. Given that over-linking is a far bigger problem than under-linking, IMO, I suggest we leave the current guideline as is. Dirtlawyer1 (talk) 14:23, 22 February 2015 (UTC)
• It is disingenuous to argue that the existing language "can already be interpreted to include extra links...", and (comment below) that "[t]he existing guidance already allows for this", as this is at variance with the practice of some editors (and I believe the duplinks tool) of removing all duplicate links on sight, without discussion, which they justify on a strict interpretation of the existing language. Strictly speaking, the last paragraph at WP:OVERLINK does say "only once' per article", the "generally" and "common sense" qualifications rarely allowing any exceptions with those who a fear a sea of blue. ~ J. Johnson (JJ) (talk) 21:30, 22 February 2015 (UTC)
• It not only is disingenuous, it is illogical. Suppose we assumed that we had only honest editors who never interfered with the the proper function of articles that depended on links (this you understand is purely hypothetical) then what we have is a set of debatable rules based on the personal aesthetic tastes of a few people, for controlling the functional design of didactic articles??? This puts the editor or author into the position of looking guiltily over his shoulder and thinking up his justifications every time he stretched the envelope, never knowing when someone who does not understand his links might come and remove them without notice, and start a fight at an advantage if he corrects the mutilation.
  The style guidelines should be to support production of superior products, not for permitting them to be extruded painfully through a picket of stonewallers who, let it be said in case you haven't noticed, do not hesitate to interfere, and do not hesitate to do it on a basis of Zero linking as one of them asserted elsewhere on this page.
  When the rules are based on counting links rather than on the function of links, then you should not have to ask what is wrong; the system is wrong. It is nonsense to argue that the flexibility is adequate, because that flexibility does not extend as far as the functional requirement, but as far as the whim of a luddite who not only wishes to control what HE sees in an article, but what anyone else is permitted to see as well.
  Furthermore note that dirtlawyer and Jayaguru-Shishya are typical in stating flatly that the flexibility is "sufficient" -- a quantitative assertion, please note. And who determines the quantity? Not their opposition! And who justifies the argument for that quantity? Nobody. The closest we have got to a quantitative argument is that if we have two links on a page the thin end of the wall of blue will nip us all in the bud.
  And when they have finished asserting that the flexibility is adequate and have gone back to their own affairs and the old guard have continued with their old habits uninterruptedly, where are they to protest the inflexibility? Certainly nowhere that they can do WP or anyone else a scrap of good.
  Where is the integrity, let alone the flexibility then? JonRichfield (talk) 19:18, 24 February 2015 (UTC)
• In my opinion, the rule should be left as it stands. In fact, "rule" isn't the correct word, as others have pointed out; it's a guideline which is somewhat strongly encouraged. Sure, an overly-literal reading of it could be detrimental to some articles, but the responsibility for that lies with the reader.-RHM22 (talk) 22:58, 25 February 2015 (UTC)

Not merely could, but actually is detrimental. The underlying problem that drives all this that some editors do take a strict interpretaton of this "guideline" as if were an absolute rule. ~ J. Johnson (JJ) (talk) 23:21, 25 February 2015 (UTC)

I understand that, but that's the fault of the editors who misread it, and not the policy, which begins with "[g]enerally." I agree with the others who suggest above that articles are more often harmed by an overabundance of links than a lack thereof.-RHM22 (talk) 23:38, 25 February 2015 (UTC)

I am sorry to hear that some editors are acting against the guideline we have. However, like RHM22 well put it[1]: "that's the fault of the the editors who misread it". In such cases, I think the issue should be first taken to administrative attention by dispute resolution or an administrative noticeboard instead of attempting to change the established WP guidelines straight away. As I have strongly understood, the phrasing of the guideline is not the problem itself, but the misreading of the guideline by some editors. Therefore, the right way to handle this problem would be to bring the guideline violations to administrative attention, and then to impose sanctions for these sanction if seen necessary. Has there been a case about aggressive misreading of the policy on hand, could one please bring it to our attention since it'd the most helpful! Jayaguru-Shishya (talk) 21:05, 27 February 2015 (UTC)

This gets back to my comment above about disingenuousness. You claim that "generally" ameliorates the black-letter text that says "only once", and shift blame for any problems that ensue on editors who "misread" the text. I say the bolded "once" beats the snot out of "generally", and that editors who take a strict view of "once" are supported by the actual, literal, and least interpreted meaning of the text.

However, the underlying problem here is not how this "guideline" should be interpreted, it is this bogeyman of "overlinking". Until we deal with that there will be no progress on this issue. ~ J. Johnson (JJ) (talk) 23:03, 26 February 2015 (UTC)

• I support the status quo: The existing guidance already allows for common sense addition of extra links where these would be necessary. Relaxing the guidance would result in articles peppered with repetitions of links "because they are allowed". Apuldram (talk) 21:02, 22 April 2015 (UTC)

Relax somewhat:

• An over-strict interpretation of this rule gives the situation where readers need to hunt for the first occurrence of an obscure or technical term, or else use the search box, instead of having a convenient link to the article. The cost of having something linked two or three times in an article is far outweighed by the benefits. --Slashme (talk) 21:24, 21 February 2015 (UTC)
• Maybe change "Generally" to something more instructive. Perhaps something along the lines of "Unless separated by a substantial amount of text." Regardless, we should put a period after "in an article" and remove the "But" following the period. In fact, I may just boldly do that last suggestion. Butwhatdoiknow (talk) 00:21, 22 February 2015 (UTC)
• The rule is written on the assumption that articles are read as though they are novels - that is we start at the lead and go through in the exact order to the end as though this were a pleasant read. An encyclopedia, however, is a reference tool - sections allow readers to go just to those sections which matter to them. Readers may go straight to the section they require, missing out the lead, and the section where the term is first linked. We have redirects and anchors which go straight to sections, missing out on previous links in the article. Each section needs to be treated for what it is, an independent section that readers are highly likely to read independently of the rest of the article. SilkTork ^{✔Tea time} 07:14, 22 February 2015 (UTC)
• I strongly favor relaxation, but it seems futile to argue for it. Based on prior history (e.g., see Archive 16) this proposal will not get anywhere without dealing with the main concern of the opposers: overlinking. Part of that concern is this fear of immediate inundation in a sea of blue. if any link should appear in an article twice. I find this fear to be absurd, but that is beside the point. The point is that as long as they adhere to that fear we will not get anywhere.

I suspect this sensitivity to overlinking arose from the WP:Overlink crisis of 2007-2011, when "many thousands of articles" were loaded up with navboxes and infoboxes that in turn accumulated hundreds of links. So perhaps it is about time for some of you folks to WAKE UP and note: 1) the overlink crisis is long over; 2) there are three orders of magnitude difference between a single duplication of a link, and the thousands of links some articles had back then; 3) neither retaining nor relaxing the "only once" rule has any affect on that kind of massive linking because it exempts infoboxes and such. ~ J. Johnson (JJ) (talk) 21:36, 22 February 2015 (UTC)

• I also strongly favor relaxation. I am a new editor (Oct. 2014) and I was really surprised when told I can only link to another page once. This shocked me because for lengthy articles this seemed really counter-intuitive. It was good to know that I am not meant to link every time this word occurred (this is what I thought at first!), but about once per visible page would be fine in my view, or once per section if people prefer that. - I often don't read an article from start to finish but I use the table of contents to jump straight to the section that I am interested in. What I also do now is to insert this, for example:
  Further information: compost
  which is in a way a blue link to another page but a more prominent one which I can place at the beginning of a section if it is appropriate. EvM-Susana (talk) 22:54, 22 February 2015 (UTC)
• Per SilkTork - in longer articles, especially, multiple links to obscure or precise terminology would provide a useful resource to readers. Perhaps relax the "per article" rule to a "per major section" rule. – Philosopher ^{Let us reason together.} 03:24, 27 February 2015 (UTC)
• There is nothing worse than having to hunt for a link that appeared earlier in an article, especially if linking the term would make more sense in the current instance than in the first. Linking should anticipate readers, not Wikipedia guidelines. Betty Logan (talk) 07:39, 27 February 2015 (UTC)
• The first thought I had reading the question was precisely what SilkTork stated! I will add that this is especially important in articles dealing with long articles and technical subjects. After noticing a comment further down this discussion, I will add that people who follow a wikilink to a specific section of an article will have to hunt through the previous sections of the article to find an appropriate wikilink. Another problem would be if the first link is piped, obscuring what the reader is looking for. AHeneen (talk) 08:30, 27 February 2015 (UTC)
• Relax the rule - why don't we just leave this up to an editor's discretion (as we do for example with choosing British/American English)? Sometimes linking more than once is appropriate, overlinking can be annoying. Editors in my experience are generally smart enough to make up their own minds without a prescriptive rule.--Jackyd101 (talk) 17:52, 27 February 2015 (UTC)
  • Comment; we do not leave choosing British/American English to editor discretion. In most cases this is determined by WP:ENGVAR. --John (talk) 09:56, 28 February 2015 (UTC)

In some cases national connection determines the ENGVAR used but in many cases it is left to the first (major) editor as per WP:RETAIN. Peter coxhead (talk) 16:33, 28 February 2015 (UTC)

• Yup, what he said - while WP:ENGVAR makes sense for London or Barak Obama, its entirely up to editing consensus whether we use honour/honor or pavement/sidewalk in national neutral articles like onion, robotics or supernova or indeed non-English speaking national articles like Tokyo, Silvio Berlusconi or Glasgow (just kidding!). Lets do the same with linking.--Jackyd101 (talk) 22:36, 2 March 2015 (UTC)

• Ditto, Slashme and SilkTork. The current guideline and common editor practice (i.e., more than one link of the same term in an article is overlink) is counterintuitive. Previously linked terms that would aid the reader in another section should be linked as well. As mentioned by others supporting a change, the guideline does not explicitly allow for repeat links in an article as the last sentence (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, hatnotes, and at the first occurrence after the lead. Duplicate links in an article can be identified by using a tool that can be found at User:Ucucha/duplinks.) does not concern the prose. --Lapadite (talk) 17:10, 8 March 2015 (UTC)

Comment:

To point out a few misunderstanding related to this RfC (in a chronological order):

• Slashme[2]: You said that the current guideline dows not allow a term to be linked twice in the text of an article excluding the lead section. This is not true. As explained by many users, a term indeed may be linked more than once in the article (excluding the lead section, of course). Jayaguru-Shishya (talk) 20:50, 27 February 2015 (UTC)
  • The guideline actually doesn't allow double linking as currently written: let's parse the text:

Generally, a link should appear only once in an article, (general rule)

but if helpful for readers, links may be repeated (exception, valid when?)

in infoboxes, tables, image captions, footnotes, hatnotes, and at the first occurrence after the lead. (exception is limited to these cases)

So if I understand you correctly, you interpret that sentence to be equivalent to the following:

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

@Jayaguru-Shishya: Would you be comfortable with this wording (without the bolding of "for example", of course)? --Slashme (talk) 21:05, 27 February 2015 (UTC)

@Slashme: Thanks for your suggestion. At this point, I'd give a "cautious yes". The wording you are suggesting might be a good solution, but I am still thinking that many editors seemed to have problem with the wording "once" (yes, with the bolding). I was also thinking of 1) removing the bolding, or 2) giving the bolding for word "Generally" instead of "once", since some editors feel that the word "Generally" is sometimes neglected by some editors. Jayaguru-Shishya (talk) 22:02, 27 February 2015 (UTC)

• J. Johnson (#1)[3]: You said that the duplicate links tools do remove all duplicate links on sight without discussion. This isn't true either. Actually, the Ucucha script only highlights the duplicate links. All the removals must be done manually, though. And that's where every editor using the tool is responsible for using personal discretion rather than removing the highlighted links blindly. Jayaguru-Shishya (talk) 20:50, 27 February 2015 (UTC)

Not quite. I was referring to "the practice of some editors ... of removing all duplicate links on sight, without discussion". A practice assisted and even encouraged by the duplinks tool, which I mentioned parenthetically. ~ J. Johnson (JJ) (talk) 23:27, 27 February 2015 (UTC)

• SilkTork[4]: I agree with Tony1. The WP articles are structured so that the text unfolds in a logical and sequential manner. Well said. Also, talking from my personal experience, there are lots of articles where the same terms have been redundantly linked all over and over again. Jayaguru-Shishya (talk) 20:50, 27 February 2015 (UTC)
  • I don't think that that is a clear case of a misconception that you are correcting, but rather a difference of opinion. Wikipedia is full of links to article sections because, for example, someone might need to be referred to the history of Germany or the geography of Germany for a better understanding of a certain point, without needing an introduction to the country. --Slashme (talk) 21:12, 27 February 2015 (UTC)
• J. Johnson (#2)[5]: Many of the links are already apearing twice, or even more. That's perfectly in line with the current guideline, and there's nothing wrong with that. There seems to be a lot of misunderstanding over what's being voted here, I think. Jayaguru-Shishya (talk) 20:50, 27 February 2015 (UTC)

Of course there is nothing wrong ("generally") with a second ("duplicate") instance of link. That in some articles some editors have tolerated such instances misses the point: other editors strictly remove even a mere second instance as "overlinking", as if "twice" is only one step removed from the evils of thousands of links. ~ J. Johnson (JJ) (talk) 23:27, 27 February 2015 (UTC)

• EvM-Susana[6]: You said that you were "really surprised when told" that you can only link another page once. I am sorry to hear that, but that's not actually how it is: You can indeed link to a page more than once. Jayaguru-Shishya (talk) 20:50, 27 February 2015 (UTC)
• JonRichfield (#1)[7]: I think the problem concerning the excessive usage of technical terms is already covered by Wikipedia:Make technical articles understandable, according to which we should strive "to make each part of every article as understandable as possible to the widest audience of readers". This is also consistent with WP:LINKSTYLE on "nested links". Jayaguru-Shishya (talk) 20:50, 27 February 2015 (UTC)
• JonRichfield (#2)[8]: "Furthermore note that dirtlawyer and Jayaguru-Shishya are typical in stating flatly that the flexibility is "sufficient" -- a quantitative assertion, please note." Actually, that's a qualitative assertion; it emphasizes users' personal discretion, and the current guideline doesn't have any fixed quantitative rule even. Just to correct the misunderstanding. Cheers! Jayaguru-Shishya (talk) 20:50, 27 February 2015 (UTC)

@Jayaguru-Shishya My apologies; I missed your responses here.

• "to make each part of every article as understandable as possible to the widest audience of readers" is certainly a worthy objective, but in technical topics it involves conflicting constraints and there is no fence to sit on between them, only a ditch. To use technical terms as opposed to abusing them whether in technical writing or not is because they are efficient. To abuse them means using them when they serve no purpose, and that is not competent writing; it is pretentiousness. But consider say:

"...hypopharynx varies greatly with the nature of the insect. Commonly it is a tongue-like structure attached to the floor of the preoral cavity between the other external mouthparts, just below the opening of the pharynx and above or anterior to the base of the labium. The salivary duct then emerges between the labium and the hypopharynx. In this form it assists chewing and swallowing, much as a human tongue does. It forms part of the floor of the cibarium, and in sucking insects, such as the Siphunculata, sucking lice, it forms part of the muscular suction apparatus, the cibarial pump. In some insects, such as many Diptera, the hypopharynx forms a stylet through which the salivary duct passes, and..."

That I think you will agree contains a generous helping of technical terms (bolded for this demonstration) for anyone unacquainted with the subject, but which of the terms do you suggest should be omitted, and what service do you suppose it would serve the reader to omit them? Why do you suppose that he would have been reading the article in the first place? JonRichfield (talk) 18:53, 15 March 2015 (UTC)

@Jayaguru-Shishya You please note, it is a thoroughly quantitative decision in that someone must decide how much flexibility or linkage is desirable or permissible. The discretion you refer to has nothing to do with quality, just whether you will suffer torments in the afterlife if ever you put in a link too many, not which kind of link. JonRichfield (talk) 18:53, 15 March 2015 (UTC)

Once per section

Since 2006 there have been proposals to allow linking once per section, proposed in July and November 2006 (unanimously supported both times), and again in 2013 and 2014 (no consensus).

Support:

• If this is approved, it must be made clear that it is not required to link once per section, but allowed in long articles. This is a reasonable way to relax the strict rule described above, and solves the problem that one can be linked to a section of an article, and will therefore not have seen previous links. --Slashme (talk) 21:24, 21 February 2015 (UTC)
• Yes, of course, it's quite bizarre that such a rule (as opposed to a set of general suggestions) should exist, and even worse if people are thoughtlessly enforcing it, as seems sometimes to happen with Wikipedia's rules. W. P. Uzer (talk) 21:37, 21 February 2015 (UTC)
• Yes, this is sensible, for the reasons given. Andrew D. (talk) 22:14, 21 February 2015 (UTC)
• Each section needs to be treated for what it is, an independent section that readers are highly likely to read independently of the rest of the article. SilkTork ^{✔Tea time} 07:14, 22 February 2015 (UTC)
• I made a similar suggestion a while back, with allowance for more than once per section for very long sections. This has to be tempered though. The real goal s not a link every so often or per section, it is to make links as meaningful and convenient as possible. Those tow needs conflict sometimes. So links=good except where it results in a link density that is too high in the article, in a section, or in a particular run of text separate from its section structure. To me, this means there should be an overall statement of goals, with any other "rules" (guidelines) being subordinate to that statement. Dovid (talk) 16:57, 23 February 2015 (UTC)
• For long-ish articles only. – Philosopher ^{Let us reason together.} 03:27, 27 February 2015 (UTC)
• We should permit with a caveat i.e we shouldn't blindly link a term once in each section, but I can imagine a scenario where we link the first occurrence in an article for those that read from "start to finish" and maybe any section where the term is "context relevant" and might be sought by someone who has jumped straight to that section. Betty Logan (talk) 07:32, 27 February 2015 (UTC)
• Support unless and until someone shows me real objective evidence that it would harm readability or usefulness, as in something vaguely resembling a scientific study. Here at Wikipedia, unsubstantiated statements of fact are also known as opinions. I've seen too many links removed on the basis of the current language to put much stock in the "it already leaves room for discretion" argument. I try to live in the real world, and I think our guidelines should live in it, too. ―Mandruss ☎ 08:55, 27 February 2015 (UTC)

Oppose:

• The existing guidance already allows for this and this proposal is based on a misreading of the existing guidance. To relax the guidance would encourage currently overlinked articles to stay bad or to become worse. The status quo is just fine in this instance. --John (talk) 21:46, 21 February 2015 (UTC)
  • I'm reading it exactly the same way as the proposer, i.e. that the list of exceptions is a closed list, and thus we have an apparent (and ridiculous) prohibition on the use of repeat links anywhere else than the places listed. If the list is not supposed to be closed, then something should be added to indicate as much. W. P. Uzer (talk) 21:51, 21 February 2015 (UTC)
• Ditto John, and see my fuller response under the When far enough apart option. Cheers, Ian Rose (talk) 23:27, 21 February 2015 (UTC)
• Let's not replace one rule that isn't working with another that won't work. Butwhatdoiknow (talk) 00:23, 22 February 2015 (UTC)
• Per my statement above and per John. Nikkimaria (talk) 02:13, 22 February 2015 (UTC)
• Ditto John. Hchc2009 (talk) 08:41, 22 February 2015 (UTC)
• I think the current guideline is already relaxed enough and leaves room for personal discretion. It does indeed allow for links to be repeated in some circumstances. As the current guideline begins, "Generally..." Jayaguru-Shishya (talk) 11:12, 22 February 2015 (UTC)
  • @Jayaguru-Shishya a repetition of the uncalibrated quantitative argument (...relaxed enough... yes? "Enough" being how much?) The guideline " generally" amounts to handwaving, giving NO indication of when enough is enough, so some geniuses have seen fit to create tools to remove repeated links automatically. Could you even hypothetically support a "guideline" that can be interpreted as support for such vandalism? JonRichfield (talk) 08:15, 18 March 2015 (UTC)
• While I am receptive to applying common sense to repeated links to the same topic in the main body sections of a very long article, I am also mindful of the mindless over-linking that often occurs in relatively short articles of a few hundred words, where the same topic is often linked in an infobox, one or more tables, a navbox, the lead, and then each of four or five brief one- to three-sentence sections. Again, I view the problem of over-linking in the typical Wikipedia article as a far more relevant concern than that of under-linking. Dirtlawyer1 (talk) 15:00, 22 February 2015 (UTC)

Perhaps you would provide us an example or two of "mindless over-linking that often occurs in relatively short articles of a few hundred words", so we could see just what the real problems are? Keep in mind that tables, infoboxes, etc., are entirely irrelevant here, as the current rule exempts them. ~ J. Johnson (JJ) (talk) 21:47, 22 February 2015 (UTC)

JJ, see my comment below. Dirtlawyer1 (talk) 18:02, 24 February 2015 (UTC)

Comment:

• We should probably state explicitly whether (for the purpose of "once per section") a "section" means only the top level, ie (defined per MOS:SECTIONS) ==Section==, or whether it is intended to include sub-sections eg ===Sub-section===. Some articles may have many small sub-sections, so we may want to limit duplicated links to once per top-level section. Mitch Ames (talk) 07:27, 22 February 2015 (UTC)
• By overlinking, we'd lose the opportunity to display intelligently selected and rationed links to readers. Jayaguru-Shishya (talk)
• I want to read and discuss the proposed language for any "relaxation" of the present guideline; propose it, and then let's discuss it. Dirtlawyer1 (talk) 15:00, 22 February 2015 (UTC)

"When far enough apart" is not the kind of statment one can support or oppose. More appropriate would be discussions of how frequently constitutes "far enough", such as "once per section" (like the section above), or per "20,000 words" (below), or even no definite rule. ~ J. Johnson (JJ) (talk) 21:24, 22 February 2015 (UTC)

JJ, I edit a lot of American football sports bio articles, many of which consist an overly long infobox, a lead, one or two sentences about the athlete's college career, a sentence about his NFL Draft outcome, and another two or three sentences about his pro career. Links to common words and phrases like National Football League/NFL, team names, positions, touchdown, quarterback sack, tackle, etc., are often repeated in every one of the lead and three or four very short sections. In my experience, this happens in many different sports articles, but it certainly is not a sports-only problem. I keep about 1,000 of 20,000+ such NFL bio articles on my watch list; given human time constraints, it's impossible to track them all, let alone fix the over-linking in all of them. So, you do what you can, and you try to educate other editors through WikiProject discussions and the BRD process. If you really want links, I can start digging out examples. FYI, in my six years on-wiki, I've gone from linking virtually every meaningful term in articles, to being something of a link minimalist. I really believe that it is better to emphasize 25 meaningful and relevant links in a Good Article rather than three times that number which include commonly understood terms as well as those that are not really central to an understanding of the article topic. Of course, that requires a command of the subject material and some measure of editorial judgment in prioritizing what gets linked. I am sympathetic, but I am unwilling to further qualify the general rule without specific language to review. Dirtlawyer1 (talk) 18:02, 24 February 2015 (UTC)

Please link to a specific example to review, so that the rest of us can what you are talking about. ~ J. Johnson (JJ) (talk) 23:22, 24 February 2015 (UTC)

Dirtlawyer1: Again, please link to a specific example showing how merely duplicating (repeating) a link more than once in the text of the whole article (but no more than once per section) leads to any kind of "overlink" problem. ~ J. Johnson (JJ) (talk) 20:51, 2 March 2015 (UTC)

Dirtlawyer1 We all are here for the purpose of improving the function and functioning of WP. In this matter however, we have had years and years of irrationally counterfunctional organised resistance to any constructive improvement of a pernicious point of view legalistically supported by what I have previously referred to as uncalibrated quantitative guidelines. The consequence has been a disservice to users, authors and editors.

"Over" linking in most terms is no more than a minor irritation of little functional importance. Rationally real overlinking only matters when it would be reasonable to call someone in for disruptive editing or the like. Occasionally there will be a slip certainly, and occasionally a novice could be gently admonished for overdoing it, but generally an extra link does less harm than a missing link.

Note that the very concept of overlinking is ambiguous, including personal aesthetics of some editors who apparently don't like blue and are willing to cry holy war on anyone with a link that they personally think unnecessary (even in cases where they don't understand the subject matter and don't realise that they have misunderstood

• a technical term that looks everyday, but really does need linking, such as mandible or spigot or labium, or
• a place name that they have thought to refer to someplace familiar to them, such as Trinidad or Cairo). So they unlink such items, possibly even automatedly.
• One even can find occasions where the same word in different contexts needs re-linking to different meanings, such as gravity in physics or engineering
• What about where different words link to the same article, such as dorsad, distal, distal, sagittal or occusal, sometimes in the same sentence, never mind on the same page?

So you want "specific language" that will permit definitive rulings on each link in each article? No, sir, these, I protest you, are too hard for me. Such language would fill books about it and about I reckon, though I freely invite you to demonstrate me mistaken by providing counter-examples. I reckon rather that things should be the other way around and that we should permit people to write more or less as they please, gently admonishing beginners who do ridiculous things, as most of us do in our original innocence. Like linking to New York repeatedly in the same sentence or short paragraph. If it isn't ridiculous leave it for procedures similar to English/American spelling and so on.

We might have a list of hints such as:

• • Are you sure your reader needs this link? When in doubt don't link, any more than you would include other unnecessary distractions.
  • Better link than explain something that an informed reader might be expected to know already; remember that any parenthetically helpful explanation that you include might well go out of date causing confusion, will irritate most readers much worse than a link will. By reserving explanations for linked articles that are in themselves coherent, you aid in keeping resources up to date, consistent and precise. This applies especially if your explanation assists readers in matters outside your sphere of professional competence.
  • If you already have the same wording for the same concept linked where the reader still can see it immediately, whether he is already following a link or not, eg in the same sentence, or the same short paragraph, you really need not link it again, and when you do not need to, the chances are that you should not; always avoid unnecessary distractions for the reader. In fact, if the same link is still visible to most readers on the same page, avoid repeating it unless in a caption, table, box, footnote or the like.

I suggest that quite a short list of such rules of thumb would be sufficient for the average newby after a few helpful admonishments, rather than umbrella "guidelines" that are so simple, clear, adequate and adequately flexible that we still are exchanging ink bombs and spitballs after how long now???. Furthermore I reckon that quite modest technical facilities for links to be user-hideable, user-specific, ad hoc searchable by pop-up, and user-invocable ad libitum, are easily practicable, which should render this precious waste of time unnecessary as being totally obsolete. JonRichfield (talk) 08:15, 18 March 2015 (UTC)

When far enough apart

A common suggestion is to allow duplicate links where the term is far from a previous occurrence, for example in this proposal. This addresses the basic issue that we face as readers, namely that one can see a term after reading two pages of an article, and have no idea where it was linked above. The question then becomes how frequently one can link. If this position gets significant support, we can draft some options.

• There used to be a long-standing bullet point "where the later occurrence is a long way from the first". This was removed on 27 November 2011 by User:Maunus; subsequent edits restored some parts of that radical cut. Maunus' edit summary refers to this RFC which he closed against consensus. Should we simply restore the previous bullet points? I think that would reflect common practice. -- Michael Bednarek (talk) 05:35, 22 February 2015 (UTC)

Support:

• It's hard to make a fixed rule in this case, but I'd say that anything that is more than 20 000 words away is sufficiently distant. The question remains whether we need a hard rule, though. --Slashme (talk) 21:24, 21 February 2015 (UTC)

• Did you really mean 20,000 words...? We shouldn't really be producing 20,000 word articles anyway under WP:TOOBIG, let alone articles in which links are frequently 20,000 words apart. Hchc2009 (talk) 21:43, 22 February 2015 (UTC)

Thanks, User:Hchc2009 - that was in fact a typo! But I didn't mean it as a serious suggestion anyway, just an example of how one might frame a rule. --Slashme (talk) 17:14, 24 February 2015 (UTC)

Oh? I thought it was a subtle point. I.e., if 20,000 words is not sufficient distance, then how about 40,000 words? Is there any "distance" sufficient to permit replication of a link? ~ J. Johnson (JJ) (talk) 23:19, 24 February 2015 (UTC)

That's a very good point to raise, but it's not exactly the point that I was making :-] --Slashme (talk) 08:24, 25 February 2015 (UTC)

• Yes, obviously, and there should certainly not be any hard rule. It's rather disturbing that we appear to have developed "legislation" in matters like this. Editors' judgment should be trusted.W. P. Uzer (talk) 21:37, 21 February 2015 (UTC)
• Generally, this would be roughly once per section. This is rather vague, and sure, there can be dispute about it. I suggest banning anyone who edit wars over this. I think that'll be a win-win situation. Martijn Hoekstra (talk) 21:52, 21 February 2015 (UTC)
• Links seem cheap and so we should be quite flexible in accommodating this where it seems helpful. Readers should never be required to hunt for a link. Andrew D. (talk) 22:17, 21 February 2015 (UTC)
• Supporting on the basis that this is what policy does imply (acknowledging John's point below), but more formalization as to describe "far enough" should be considered. As well as the aspect of how links in infoboxes and tables affect this. --MASEM (t) 23:59, 21 February 2015 (UTC)
• More guidance is good. Butwhatdoiknow (talk) 00:23, 22 February 2015 (UTC)
• Duplicate links, when far enough apart in the article text, are permitted if needed, is sort of the idea I would prefer, and what even the opposers below say they want. Once per section is a MAXIMUM, that should hardly ever (never?) occur. Once per article body is the goal, but I would like to see it broken frequently for odd and little-used terms and long articles. I do not agree that the language we have (below, from WP:REPEATLINK) is crystal clear. If our policy is that once per article is the general goal (to avoid the dread sea of blue), however duplicate links to the same term in the body, for odd words, are permitted in certain dire situations, then repeat that caveat right after the controlling paragraph below. Those people here saying it's already clear, are wrong, or else we wouldn't be having this long discussion. It's not clear until you put ALL the exceptions right where the reader needs to see them, which is not the case here. SBHarris 02:05, 22 February 2015 (UTC)
• Agree with Sbharris. SilkTork ^{✔Tea time} 07:14, 22 February 2015 (UTC)
• I agree with the principal of "permitted when far enough apart". As has been mentioned by others, readers do not necessarily read entire articles, from the top, and nor should they have to. We probably need some specific guideline about what is "far enough" (editors will probably disagree over it otherwise), but I have no suggestions for now. Mitch Ames (talk) 07:53, 22 February 2015 (UTC)
• I made a similar proposal some time ago. I believe my suggestion was some approximation of 1x-2x per printed page for long sections. Dovid (talk) 16:53, 23 February 2015 (UTC)
• I propose that no more is demanded than a rule of thumb,and that apart from the current "captions, sections, tables etc", we should aim for about a page apart, so that one is likely to find just one hideous blue link per page. I also reckon that the whole thing could be averted by a change in technology, but that has been shouted down repeatedly by link-haters in the past. JonRichfield (talk) 07:55, 24 February 2015 (UTC)
• The bold text suggested above by Sbharris seems right to me. Peter coxhead (talk) 10:05, 25 February 2015 (UTC)
• Definitely. I feel like objections to this are rooted in the desire for wikipedia to be "proper" and look nice, at the expense of being as useful to our readers as possible (which is obviously more important). Links that aren't very relevant shouldn't need to be duplicated, and I agree that there should be a reasonable gap, but useful/relevant links should be duplicated after a while. We don't know which section of an article a reader may be looking at, e.g. they may jump straight to a "Legacy" section, where there won't be any links to that individual's important works because they will have been linked earlier in the article. The reader would have to scroll around looking for the link, which isn't right. And I disagree that the current wording is flexible enough as people are always removing/telling others to remove any duplicate links. --Loeba (talk) 21:50, 26 February 2015 (UTC)
• I like Sbharris's choice. While I think "once per section" would still be a useful guideline, it's a distant second choice to this. – Philosopher ^{Let us reason together.} 03:30, 27 February 2015 (UTC)
• I think that User:Sbharris's proposed language would best serve the purpose of making it clear that 'only one link in the text' is not an absolute rule. 24.151.10.165 (talk) 22:27, 28 February 2015 (UTC)

Oppose:

• The existing guidance already allows for this and this proposal is based on a misreading of the existing guidance. To relax the guidance would encourage currently overlinked articles to stay bad or to become worse. The status quo is just fine in this instance. --John (talk) 21:46, 21 February 2015 (UTC)

The current text is "if helpful for readers, links may be repeated in infoboxes, tables, image captions, footnotes, hatnotes, and at the first occurrence after the lead". This does not allow links to be repeated in the body of an article, no matter how far apart. --Slashme (talk) 22:14, 21 February 2015 (UTC)

• Slashme, the current guideline is just that, and per the heading does allow for "common sense" and "occasional exceptions". I'm with John on this matter. As a FAC coordinator, one of the last checks I make on an article before promoting to FA is a duplink check using Ucucha's script. If the article is detailed and the links are some distance apart, I've always given the main editors some discretion. What I think we do want to avoid is a sea of blue, and the status quo helps us achieve that while still permitting some duplicates if they can be justified. Cheers, Ian Rose (talk) 23:14, 21 February 2015 (UTC)
• Agree with Ian - exceptions can be made, but what we already have seems a good general rule. Nikkimaria (talk) 02:13, 22 February 2015 (UTC)
• Ditto John. Hchc2009 (talk) 08:41, 22 February 2015 (UTC)
• I think the current guideline is already relaxed enough and leaves room for personal discretion. It does indeed allow for links to be repeated in some circumstances. As the current guideline begins, "Generally..." Jayaguru-Shishya (talk) 11:12, 22 February 2015 (UTC)
  • @Jayaguru-Shishya, John, Nikkimaria, Hchc2009, and Ian Rose: if I understand you properly, you do not say that a link to an obscure term should be removed from a paragraph at the bottom of a long article simply because it occurs near the top, but you all believe that the current wording makes this clear. I don't think this it does: the list of exceptions to the rule only includes non-body-text examples. --Slashme (talk) 12:47, 22 February 2015 (UTC)
    • Yes, the word "Generally" right at the start already implies this. --John (talk) 13:07, 22 February 2015 (UTC)
      • It actually doesn't: let's parse the current guideline:

Generally, a link should appear only once in an article, (general rule)

but if helpful for readers, links may be repeated (exception, valid when?)

in infoboxes, tables, image captions, footnotes, hatnotes, and at the first occurrence after the lead. (exception is limited to these cases)

So if I understand you correctly, you interpret that sentence to be equivalent to the following:

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

@Jayaguru-Shishya, John, Nikkimaria, Hchc2009, and Ian Rose: Would you be comfortable with this wording (without the bolding of "for example", of course)? --Slashme (talk) 18:44, 23 February 2015 (UTC)

• Comfortable invocation of the "sea-of-blue" catch-phrase comes easily, much as the Aussies used to speak of the"yellow peril" and the NATO nations of the "red peril", but it is no substitute for practicality or responsibility. In reality "sea of blue" so far amounts to what? An article in blue punctuated with the occasional redlink or conjunction? I have never seen such, and correcting one if we found it would be trivial, probably even leading to grounding someone on the grounds of disruptive editing. What I certainly HAVE seen are articles full of obscure terminology with linked terms that occur less than once per page. I have seen sea-of-blue nazis removing links close to each other, where the linked word happened to be the same, but the articles linked to were different. I have seen editors virtuously removing links to commonplace words in ignorance of the fact that they did not refer to commonplace concepts. I have seen links removed because of the "everyone-knows-that" syndrome when what they mean is that the local usage is familiar in his town, his discipline, his patois, or even in his personal youth, but where it is not easy for others to guess what he is on about. The function of links at a density of at least one per screenful per linkworthy concept is a real and daily re-confirmed service to the reader who needs guidance, whether he is familiar with the subject or not, and who might have entered the article in the middle for reasons good, bad, or indifferent. As against that we are to be concerned with the convenience of what? A minority of people who dislike blue??? How much more of a service to our readers is it for a blue-hater to be spared skipping a visible link than for a reader who needs a link to be sent searching for one in the hope that there might be one somewhere, or to abandon an article in which he is lost in terminology without any hint that a given term could easily have been found if he had only known where to find a link? The very principle of confining a sound practice and requiring editors to go out of their way to justify links that the majority of readers at worst don't mind, and that readers in search of support actively need (the ones we write for -- remember them?) just to gratify the personal tastes of the blue-haters is wrong-headed. If as a matter of taste some people would like to see less blue, but still would condescend to permit the occasional link, then there are technological alternatives that they could agitate for to hide the blue and extend the linkage function instead of demanding the retention of guidelines that militate against practicality on the grounds that if the editors only would spend extra effort working round them, then maybe no one will descend on their work and devalue it. Similarly if naive readers only would repeatedly interrupt their reading to check for whether there is a link somewhere that might help them, then maybe they wouldn't need links anyway. And neither would we, right? It is no good speaking of benignly flexible guidelines when in the next breath one mentions duplication detectors; that bespeaks a destructive misunderstanding of function when constructive alternatives could achieve more for the same effort. "Allowing for common sense" is when you find some real sea of blue and correct matters, not when you grudgingly admit that an editor might have a point in stretching a "guideline" that wikilawyers are happy to use in justifying their habitual disruption. JonRichfield (talk) 12:43, 22 February 2015 (UTC)

Yes. Lengthy, but I fully concur. It is a bit of a farce to claim that the rule is permissive, when in actuality some editors use it non-permissively. ~ J. Johnson (JJ) (talk) 21:58, 22 February 2015 (UTC)

Strong oppose. I don't agree with Silk Tort's contrast with reading a novel. Of course it's not the same as reading a novel, but WP articles are structured in the standard way, on the basis that the text unfolds logically and sequentially. In the minority instance where someone parachutes in via a section link, they can expect to encounter terms, concepts, ideas, information that are dependent on the previous text—whether by glossing, expansion, or unfolding context. Same for acronyms and initialisms: should we start expanding them several times during an article, just in case someone has forgotten? Well no—a reader might have to revisit the previous text if they forget. Reps of "National Science Foundation (NSF)" would be irritating to many readers.

Once you start giving the green light to multiple linkings of the same item in an article, there will be no end to it. People will argue about what constitutes "enough distance", for example. If you want a link to something that is unlinked, either scroll up till you find it (which you probably should do anyway) or type it into the search box.

Please protect our linking system from dilution by repetition: the more blue patchiness, the less "selective" value each link has. That's basic reading psychology. Tony (talk) 06:51, 23 February 2015 (UTC)

Strongly oppose Strong oppose. There is no merit to "strongly opposing" (if that expresses passion). Passion is no substitute for logic,and logic in this case says that the interests of the reader who needs it most come first, and that if what the reader who needs it most is irksome to passionate parties, then the latter are the ones who should make the concessions (if we could call that a concession at all). If their aesthetic preferences cannot be satisfied by the current system,then the proper course is to request an option that will suit them, while permitting busy researchers and novices to make the best of the available tools most effectively. Telling users not to follow links into the middle of articles, or to scroll up and down until with any luck they find the link they want and trust that it does in fact link to the intended concept as the word that they had found unlinked, is the sheerest arrogance. Who on Earth has the right to tell people how to do their research, or more precisely, to do without it because you don't like the way links look? To demand that readers read entire articles because notionally "articles are structured in the standard way, on the basis that the text unfolds logically and sequentially" is a travesty. They might sometimes, and it might be nice if they did so more often, but to conclude that only lesser life forms might need to read limited sections in contexts different from those intended by the authors cannot have much sense of the nature of research -- or of browsing. The function of the encyclopaedia trumps personal tastes, and again I repeat, to impose one's personal tastes at the cost of function is unethical as well as illogical. And as for the slippery slope into the thin end of the wedge of solid blue, that argument is an embarrassment. JonRichfield (talk) 19:52, 23 February 2015 (UTC)

Editors who refer to a "sea of blue" are reminded that we have specific WP:SEAOFBLUE, aka WP:Sea of blue, which (so far as I can tell from its location) applies only to "links next to each other so that they look like a single link". This is quite distinct from WP:OVERLINKING: "an excessive number of links". Some of the contexts in which participants here have used the term "sea of blue" suggests that they are referring to "too many links" (WP:OVERLINKING), rather than "links next to each other so that they look like a single link" (WP:SEAOFBLUE). One could reasonably favour either one of those policies while disagreeing with the other, so it might be helpful to the discussion if participants are clear about which of those two distinct concepts they refer to. Mitch Ames (talk) 12:27, 23 February 2015 (UTC)

Thank you Mitch Ames, I was one of those you described and unaware of the distinction. FWIW the substance of my objection, plus the bulk of the controversy concerns fuzzy thinking about "overlinking", with "sea-of-blue" as in essence a term of abuse handy for those opposed to reform, but shorn of any basis for rational argument. The real sea-of-blue problem is not a major issue anyway, as no serious editor to my knowledge approves confusing concatenation of links. That guideline as I see it is more of a heads-up than a prescription anyway. The guideline also recognises that concatenated links are very hard to avoid in some technically dense topics. We might think about an appropriate tool for dealing with such situations, but then, they are not very common, are they, though I certainly have encountered situations of the type in my own writing. JonRichfield (talk) 07:55, 24 February 2015 (UTC)

• Oppose per Ian Rose. Less is more. -- Ohc ^¡digame! 09:46, 23 February 2015 (UTC)

"Once you start giving the green light to multiple linkings of the same item in an article, there will be no end to it"? Bullshit. If we allowed a single instance of link in each section, there would be no more links than sections, and there is the end of it. There is no slippery slope. "No end of it" is totally bogus.

Ohc: If "less is more", why don't we go for zero instances? (Haven't we had this discussion before?) ~ J. Johnson (JJ) (talk) 22:45, 23 February 2015 (UTC)

"Zero instances" is actually what we practice. We don't go linking every single word, although this is a technical possibility; we don't link every occurrence of every less common word. We link only to articles that are relevant to aid and guide readers, without causing them to leave the article prematurely, before they have understood what they came to understand. In fact, when editors link, we carefully ration linkings of those words within an article.

I think that the currently guideline is well-balanced – it is sufficiently firm, yet leaves sufficient leeway for the editor to link as required. And I don't feel that we need for it to get more prescriptive because the general frequency of linking is deemed insufficient by some editors. -- Ohc ^¡digame! 03:09, 24 February 2015 (UTC)

OHc: you have misunderstood what I said. The current language says that a link — not a duplicate link, just a link — should occur ("generally)" "only once". By your "less is more" criterion zero instances of any link should be better. And indeed, simply eliminating ALL links certainly would put an end to all overlinking and sea of blue problems. If you reallly want to go with such an inane criterion you should apply it consistently ("down will all links!"), not just in special cases. Another thing: you misstate the proposal. Your mention of "linking every single word", and "every occurrence of every less common word", is more bullshit (more precisely, those are straw man arguments), because the proposal has nothing to do with those. Please stay on topic. ~ J. Johnson (JJ) (talk) 23:29, 24 February 2015 (UTC)

I actually want the rule to be slightly less prescriptive, not more: my problem is not that the current frequency of linking is insufficient, but that some editors interpret this rule strictly as it is written, namely as a complete ban on linking a word more than once in the body text, and I would like the rule to reflect what I see as the consensus in good articles and in this discussion, namely that duplicate links in distant parts of a long article are helpful to readers, and should be allowed. --Slashme (talk) 07:52, 24 February 2015 (UTC)

However proper it may sound to claim that "We link only to articles that are relevant to aid and guide readers, without causing them to leave the article prematurely, before they have understood what they came to understand", that is unrealistic and patronising. Our job is to make sure that the readers have the tools they need to discover what they wish to know and are assisted to find and use those tools. It is not to penalise them for wanting to know anything but what authors and editors decide that they should know,and in the sequence and context that they should know it. For a reader to wish to know what an unfamiliar word or usage might mean in context, instead of waiting till he has read uncomprehendingly from the beginning, is entirely reasonable and practical; some of us may remember having actually paused in reading a paper book to consult a paper dictionary or a paper encyclopaedia for clarification on anything on spelling or the meaning or usage of a word or on any matter of technical fact, history,or opinion. Who is to claim that such practice is in any way pernicious or undesirable because it detracts from the benefits accruing from reading the book in the manner that the author had intended?

Nor does it support that "Zero instances" is actually what we practice. It might be what Ohconfucius would prefer, but in practice we link-heads are spared the occasional link as a sop no doubt, and that is different from zero instances; note that in spite of the apparent triviality of the point, if partisans insist on appealing to the concept, it becomes necessary to refute it.

Then again we are reassured that "In fact, when editors link, we carefully ration linkings of those words within an article." Pardon those of us who thought we were editors and thought that we were doing nothing less, but there would seem to be room for debate on our delusion, given that this squabbling about the concept has continued for years, with the same team stonewalling constructive practices, constructive discussion, and constructive technology, without the slightest material justification.

For example: "I think that the currently guideline is well-balanced – it is sufficiently firm, yet leaves sufficient leeway for the editor to link as required. And I don't feel that we need for it to get more prescriptive because the general frequency of linking is deemed insufficient by some editors." Someone is of the personal opinion that everything is hunky dory, so nothing must be adjusted even if other users have repeatedly demonstrated that improvement is desirable, even though someone else passionately pleads with us to "protect our linking system from dilution by repetition: the more blue patchiness, the less "selective" value each link has. That's basic reading psychology" and asserts "In the minority instance where someone parachutes in via a section link, they can expect to encounter terms, concepts, ideas, information that are dependent on the previous text—whether by glossing, expansion, or unfolding context".

"Minority instances" is an unsupported opinion in the first place; when did that participant, or anyone else quantify the instances? Secondly, how small must that minority be to matter? I am likely to read several thousands of words before having to refer to a dictionary or encyclopaedia (or a link) even once, even in a long article; does it follow that the dynamic function reference function is only a several-thousandth as important as the rest of the material?

And if the linked reference drops me into the middle of a twenty-page article (or a two-page article FTM) then I must go back to the start and read an extra few thousand words of no relevance when reading just a few dozen plus a click or two would completely satisfy my needs without wasting my time and dividing my attention?

And "reading psychology"???? Someone had better go back and learn a bit about reading psychology before making such claims in public. The "dilution" effect is over-use of a particular format, punctuation, cliched text or expression, or other demands on a reader's patience, emotions or attention, it has nothing to do with links, which are no more "dilution" of the text than necessary paging up or down might be. He would not, I hope, urge that we chop up our articles into single-page slices to render routine paging unnecessary.

Am I alone in smelling red herring in "I think that the currently guideline is well-balanced – it is sufficiently firm, yet leaves sufficient leeway for the editor to link as required. And I don't feel that we need for it to get more prescriptive because the general frequency of linking is deemed insufficient by some editors." Veeerrry subtle. Great courtroom tactics. The prescription is all on the other side. Persons agitating for more constructive linking practices are NOT the ones requesting prescription, they are the ones wanting more constructive freedom for competent editing aimed at assisting users in need of assistance. It is the blue-sea bigots who use the guidelines as a basis for wikilawywering. "The guideline says... so do it MY way, even if I don't know what you are talking about and don't care!" JonRichfield (talk) 07:32, 24 February 2015 (UTC)

Yes, lots of red herrings. Also straw man arguments, non sequiturs, and entrenched emotionalism, all of which impede rational consideration. To make any headway in this I think is is necessary to first identify the arguments and sentiments in opposition, reject the spurious ones, and then we can see what remaining objections merit consideration. ~ J. Johnson (JJ) (talk) 20:44, 25 February 2015 (UTC)

My proposal

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, hatnotes, and at the first occurrence after the lead.

My proposed text, with the change in bold:

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

Alerted WikiProjects and other related pages to this discussion

As seen, here, here, here, here, here, here, here, here and here, I alerted WP:Med, WP:Film, WP:TV, WP:LGBT, Wikipedia talk:Good articles, Wikipedia talk:Good article nominations, Wikipedia talk:Featured articles, Wikipedia talk:Featured article candidates and Wikipedia talk:Manual of Style to this discussion. I alerted those WikiProjects because those are the main WikiProjects I am involved with these days, and I alerted those other pages for obvious reasons. Flyer22 (talk) 21:36, 26 February 2015 (UTC)

Review of opposing opinions that need demonstration

Discussion (speaking loosely) on this issue has (again!) ground to a halt, showing a general tendency of editors to repeatedly declare opinions without actually grappling with the basis of those opinions. I am going to list some of the opinions that have been declared here which I think lack a demonstrated basis. (Note: I do not say they lack a basis, only that such a basis is not adequately demonstrated.) If anyone has anything useful to say about any of these please open a separate subsection at the bottom for discussion.

1- "Relaxation not needed, as already allowed."

"Generally" yields before, and the quaint notion of "no fixed quantitative rule" is contradicted by, the actual, explicit, and bolded text of "only once per article".

2- "The problem is with editor's who 'mis-read' the rule/guideline."

The fact is that those editors explicitly cite "per WP:OVERLINK", and their interpretation is a fair reading of "only once ...."

3- "A 'sea of blue' would be created that makes valuable links harder to identify."

In fact, WP:SEAOFBLUE refers to sequences of links that appear as a single long link, what would be more accurately called "dribbles of blue", and there is no credible showing that duplicate links contribute to this specific problem. The "would be created" is entirely a speculative anticipation of an undemonstrated and unlikely prospect.

4- "Needed to protect our linking system from dilution by repetition."

There is no showing that "our linking system" needs protection, of how having (say) 103 links in an article instead of only 100 dilutes anything. It also completely overlooks that the utility of the "linking systen" is diluted if the reader is to hunt for a relevant link.

5- "Relaxation would encourage the continuation or worsening of overlinking."

None of the proposed alternatives have encouraged "overlinking", nor is there any showing of exactly what would be worsened. (See the following.

"Overlinking" has been repeatedly invoked as if that phrase alone is a complete and adequate justification. As far as I can tell, that concept includes the following objections:

5- "It would bring the severs to their knees."

News flash: the WP:Overlink crisis of 2007-2011 is over. And it involved instances of articles having thousands of links - a problem three orders of magnitude greater than any mere "duplicate link" problem. Also, these involved links in navboxes, and thereby would be exempted from the current "overlink" rule. Invoking "overlink crisis" in this discussion is entirely a red herring, as it is entirely irrelevant.

6- "Loses the opportunity to display intelligently selected and rationed links."

How? It is the current formulation that is relied on the deny the "opportunity" to "display intelligently selected" links.

Certain other arguments have been made that are so inane that listing them seems tantamount to trolling; I won't address them unless they are raised.

If anyone wants to discuss these please open a new subsection below here. ~ J. Johnson (JJ) (talk) 00:52, 8 March 2015 (UTC)

P.S. A lot of the so-called discussion here has been declarations ("arm-waving") that there are (have been, will be) problems if the "guideline" is relaxed. What we could use are links to specific articles that demonstrate definite problems. ~ J. Johnson (JJ) (talk) 20:07, 8 March 2015 (UTC)

Why does this have to be a "relaxing" of the rule (which carries implications of broad new freedoms), instead of specifically approving (but not mandating, in technical RFC-speak: SHOULD, not SHALL) particular secondary links such as inside of a table or list which hosts entries with links to items that are similar in nature to the link in question. An example of this are the names of real persons without links in the list of Secondary Characters in the novel American Tabloid. I find it ridiculous to have been corrected for wanting to make a table like that one contain links for all real persons, regardless of whether the article had mentioned them earlier. It makes a wonderfully functional table into a piecemeal jumble. Spawn777 (talk) 15:01, 9 May 2015 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Once

"Generally, a link should appear only once in an article" That is annoying, especially when you are directed into the article to a specific section. Why can't it be linked in each section? --Richard Arthur Norton (1958- ) (talk) 02:36, 22 May 2015 (UTC)

The same reason style guides and standard practice is to expand and provide (initialism) only once, and subsequently to use the initialism only. The "I parachuted into a section" argument would have us insist that "National Rifle Association (NRF)" be spelled out in every section—tedious. It's generally accepted that if you do parachute into the middle of an article, you might need to scan back earlier to extract context, structural information, and links. Thank you. Tony (talk) 02:47, 22 May 2015 (UTC)

A question from an editor about linking nations

It's not here though, it's at Talk:Brian Sylvestre#WP:OVERLINK. Please comment there. Walter Görlitz (talk) 18:17, 11 April 2015 (UTC)

It's better to move discussions like that to the guideline's talk page where it'll become part of the guideline talk archives. Users are free to simply delete entire conversations off their own talk pages.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  11:22, 23 April 2015 (UTC)

Looks as if you scanned that namespace in a hurry, SMcC -- it's an article talk page, not user talk page. :) It does perhaps make sense to broach the issue here, even so, but bear in mind it is a matter which has been discussed here before (and indeed, countless other spaces across the project). You'll see when you take a look there; it's an old issue that has no broad community consensus and upon which local consensus usually varies considerably, both as a consequence of editorial perspective and the particular context of the article in question. Snow ^{let's rap} 05:47, 29 April 2015 (UTC)

D'oh! Need better glasses, I guess.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  16:20, 13 June 2015 (UTC)

Wikipedia:Village pump (policy) discussion

Started a discussion regarding this part of the MoS at the Village pump. Wikipedia:Village pump (policy)#RfC: Should the holder of a political office be linked within an infobox more than once (i.e. as the successor), when they have already been linked (e.g. as the vice president, predecessor, lieutenant, etc.)?. —Godsy^(TALK_CONT) 06:35, 21 May 2015 (UTC)

Linking in lists and references

How should the policy be interpreted when linking in lists and references?

While I see there is some consensus on relaxing the duplicate linking rule, I am still seeing it somewhat strictly enforced and it seems to me that both lists and references should be treated somewhat differently than regular prose. It's more common for people to pick out specific items to read in more detail, and especially with reference sources, it's probably relatively rare for someone to sit and read straight through the references section.

What do other editors who are more expert on this topic think? --ProtectorServant (talk) 13:18, 9 June 2015 (UTC) (signed later)

I think the question should be split:

Linking in lists

I'm a bit puzzled by this. Surely an item is not repeated in a list? Apuldram (talk) 16:48, 8 June 2015 (UTC)

The example linked to is a List of cases of police brutality in the United States and department names (such as the Los Angeles Police Department) are linked to when those departments are involved.

Another example which has been handled differently might be the list of killings, where argument along the lines of "we don't need another link to the city of Albuquerque (as the location where a killing happened) because the police there also killed somebody else at another time in the month" doesn't make sense. --ProtectorServant (talk) 13:17, 9 June 2015 (UTC)

Most times, duplicate links are not helpful in lists, but we should allow that if consensus at an article determines that they are in that context, that this is permissible. Editors do this whether we "allow" it or not anyway, per (usually unspoken) WP:IAR. Many do it outside of lists two, e.g. when several sections have gone by since the last link. This is really okay, too. Some of what we don't want to see is over-linking in lists just to make entries consistent. This argument is sometimes actually valid in a table, but I strongly question it in a list. Another is a list of redundant links in "See also". I'm utterly merciless with those. If something has been linked in the article already, I nuke it from the see-also list, and give myself a pat on the back when I can totally eliminate a see-also section by working the links in it back into the main prose of the article. Duplicate links in any but the longest and most detailed infoboxes are ponitless and browbeatingly redundant. Back to lists: I don't have any specific wording to propose at this point, but generally I wouldn't re-link unless the last link to the same ting was more than, say, 10 list entries away. Big exception: Glossaries. It's routine to cross-reference, by linking, every occurrence of every term that has an entry, except is very short glossaries. Glossaries are a special case where people are not bothered by a sea of blue, and expect terms to be maximally linked; such article are entirely about the words as words. See Glossary of cue sports terms for one of the best-developed ones (if I may say so myself; I worked on it a lot, but so did various others). Finally, the nature of the list matters in other ways. In a list expected to be read top to bottom, don't re-link at all. If it's a list we expect people to be looking for one item in before moving on to something else, link every occurrence. More usually, our expectation is between these extremes, so re-link when it's thought this will be helpful and not annoying.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  16:11, 13 June 2015 (UTC)

I think that duplicate listing in links is permissible in some cases if it significantly facilitates the reader gaining pertinent information- unlike prose, a list may not be read from start to finish, and especially in lengthy lists and/or sortable lists, such that the "first" instance of a link may not be such when sorted differently. It can be left to editors of particular lists to decide whether multiple linking in lists is a deterrent or aid to readers, but I think certainly a place where WP:IAR can apply. I also think that in some instances it is OK to repeat a link in a "See also", especially if the body link is somewhat buried: by giving additional prominence to a well chosen article in a See also, the reader is more likely to increase their understanding of a topic. --Animalparty-- (talk) 21:52, 28 June 2015 (UTC)

I would highly discourage that. For example, would it be smart to duplicate a link for 142 times only because the link keeps occurring in a list[9]? Jayaguru-Shishya (talk) 23:47, 28 June 2015 (UTC)

Linking in references

I see no problem in relaxing the rule for references. The sea of blue problem doesn't arise there. (but see below) Apuldram (talk) 16:48, 8 June 2015 (UTC)

To the contrary, I find that it applies more there, and I delink a lot of things, especially city names in |location= parameters and other such trivia. I remove all redlinks from refs, and duplicate links for authors/publishers. And links for publishers that are just general publishers (I keep them for organizational publishers tied to the content or field; no one needs a link to Houghton-Mifflin, but one to an institute or whatever may be relevant for determining bias or reliability.) When the refs section is a sea of blue, it's very hard to find which blue item actually takes you to the source you want to examine. That said, many of the citation templates just relink stuff every time they are used, e.g. ISBN. Nothing we can do about that. But this fact isn't justification for linking like mad in refs sections; that's a WP:OTHERCRAPEXISTS failure.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  16:17, 13 June 2015 (UTC)

You've changed my opinion to don't relax. I now see the problem. Thanks. Apuldram (talk) 17:12, 13 June 2015 (UTC)

Linking to drafts from articles

FYI

 – Pointer to relevant discussion elsewhere

Please see Wikipedia talk:Drafts#linking to drafts from mainspace? for a discussion that might add content to this MOS subpage. --Ahecht (TALK
PAGE) 16:58, 11 June 2015 (UTC)

Proposal to permit redlinks in nav templates

FYI

 – Pointer to relevant discussion elsewhere

Please see Wikipedia talk:Red link#Proposal to permit redlinks in navigation templates  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  20:48, 17 June 2015 (UTC)

Main points deleted?

Why have many of the main point of the subpages deleted? I think having all the main point here is much better then having to run all over to find them. I suggest a revert and a discussion of reworking of the page. -- Moxy (talk) 20:35, 29 June 2015 (UTC)

In this edit this edit I removed the statement that links should not be given special colors, because (a) there's already a general prohibition on coloring text, and (b) I've never see this happen in eight years anyway. Certainly if others think it should remain it should be put back. Other than that, can you provide one or two diffs where the "main points of the subpages [were] deleted"? Thanks. EEng (talk) 21:09, 29 June 2015 (UTC)

link colors happens all the time.... and is a main point for accessibility for our readers. All the main points on linking... the do's and don'ts should be listed. Best not to make our readers have to run around to 12 sub pages just to get the main points on linking.

Based on what you're saying the don't-color-links warning should remain. But I don't know what you mean about the 12 subpages. Can you list a couple of diffs to illustrate what you mean? EEng (talk) 22:18, 29 June 2015 (UTC)

For each proposed change, EEng, can you please open a discussion, and give your rationale and explain why these changes would be necessary? As I stated in my Edit Summaries, many of your edits removed a lot of relevant material, changed the tone of the guideline, or moreover, changed the meaning of the guideline. For bold changes like this, it'd be important that they are discussed, especially after being reverted for obvious reasons (see WP:BRD). WP:EW is no way to go, I am afraid. Jayaguru-Shishya (talk) 22:52, 29 June 2015 (UTC)

Changes don't have to be "necessary", and the idea of opening separate discussions for edits such as [10] or [11] or [12] is ridiculous. Each edit is self-explaining via the change it makes to the text itself plus its edit summary. There's nothing I can add until you point to something you specifically object to, but you've refused to do this despite multiple requests (e.g. [13]). There are 12 edits, and you say they're all inappropriate, so it should be very easy for you to supply diffs for two or three you object to, with a short reason. EEng (talk) 23:23, 29 June 2015 (UTC)

"No substantive objection"? Are you kidding me? As you can from above, there is objection to your bold edits. Instead of discussing your bold edits, trying to explain your rationale and to establish consensus, you have been WP:EDITWARring instead.

1. Why did you change the tone (and even the meaning) of text?[14] The original version was not about "accidental linking" but about "confirming that the context is helpful". This can be the case even if the linked article was from the same subject area, but if it didn't deal with the context of the wikilink at all.
2. Why did you remove this passage[15]? Self-supporting per Edit Summary again? Please see WP:REVTALK.
3. Why did you remove this one as well?[16] As explained in no. 1, those two are different things. Therefore, it is completely reasonable to emphasize this aspect as well.
4. In this edit again you keep going on about that mantra of "mistakenly linking".[17] The previous formatting was very descriptive, really well-written, had a nice tone... Now it's all changed, why?
5. Why were these step-by-step instructions removed?[18] Although you and I might not need that sort of instructions, there's a whole variety of people editing Wikipedia. Have you ever followed WP:TEAHOUSE discussion? Trust me, there's a lot of people volunteering to answer questions that might seem obvious for the most of us.
6. Although not entirely wrong[19], I don't understand the tweak as the former version was way better English.
7. Here you changed the whole meaning of the former two passages (!)[20]. The previous version clearly said "do not unnecessarily make a reader chase links" and "as far as possible do not force a reader to use that link to understand the sentence". You changed it into "should be not only linked, but briefly glossed (if possible)...", as if linking is something one is supposed to do and explaining the term is just something additional to that. No, that's not what the original version said.
8. Why did you remove this well-prosed passage and replace it with some clumsy parentheses? How does this help to capture the original essence of the guideline?
9. Why did you move this passage?[21] It was there for a reason. Please see WP:SEAOFBLUE above, something it is strictly related to.
10. See what I just wrote above[22], they might demand a higher density of links (see WP:SEAOFBLUE), not often have a higher density of links. This is a guideline, and our guideline is "might".
11. Again, this is a guideline[23], therefore "Do not" instead of "The function of".
12. And again, what's with this?[24]

EEng, this is the first and last time I will give you a detailed explanation about "WHY NOT" to keep the changes you made, instead of you explaining "WHY" to keep the changes you made. Remember, you made the bold edits, you are ought to discuss the changes and seek consensus, and the WP:BURDEN is on you. Besides, this is not just about some individual edits. Nobody is able to catch the original essence of the guidelines if you remove all explanatory material around them, for no reason. Jayaguru-Shishya (talk) 16:31, 3 July 2015 (UTC)

No, I'm not kidding. Any old vague "objection" isn't "substantive objection", and generalities such as "many of your edits removed a lot of relevant material, changed the tone of the guideline, or moreover, changed the meaning", without specifically saying where you see that happening, are nonsubstantive. And please stop linking to policies and guidelines you don't yourself understand e.g. BURDEN applies to article content, not project material.

1. The old text's image of "a physicist speaking of" something made no sense in a discussion of article editing, and "what the system offers" was a bizarre way of saying "the linked article". I see your point about the linked article being "helpful in context", but I think it's not a useful criterion: if the linked article is the one intended -- e.g. a physics article linking to barn (unit), not barn -- then it's hard to see how it could fail to be helpful to many readers in context (no matter what its actual content) because even a stub would be helpful to a reader who'd never heard of that unit. And even if one argues that Article Y, linked from Article X, omits the very thing the reader of X wants, a properly developed version of Y would include that material, and we wouldn't not link from X to Y just because Y isn't properly developed yet. I'd be interested to hear what other editors think about this.

2. This bullet point was restored (see edit summary here) as a result of discussion earlier in this thread. The fact that you don't seem to know this raises serious question about how carefully you've considered the changes you're complaining about.

3. As you must have noticed, the very next edit (i.e. #4) reinserted this material elsewhere, so your complaint here is nonsense.

4. I don't believe in mantras. Editors don't need to be told it annoys readers when a link goes to the wrong article, that article titles need to be exactly right, that there are often multiple articles with similar titles, or other obvious and/or unimportant stuff that made this bullet point twice as long without giving editors any clearer idea what they're expected to do. The important point is to encourage editors to verify that links go where they think they're going; once that's pointed out the reader will readily understand why that's important without talking down to them. I'll be interested to hear from other editors whether they think my revision does or does not do that as (or more) effectively than the old text; it certainly is less tiresome to read.

5, 6. (These two adjacent edits are best evaluated as a single diff [25].) No, it was badly written, with a tin ear for calibrating instructions to what the reader can or cannot be reasonably expected to know already.

• Editors don't need to be told to key in links "carefully" -- that's useless advice that applies to everything.
• "display mode" -- that's not a term we use on WP to my knowledge -- why not "preview mode" (though I don't recall any talk of "modes" at all)? And even assigning it this reasonable meaning, what other "mode" would you be in after clicking "Preview"? So what's the point of saying it at all?
• Telling editors to "check [links] go where you intend" is fine, but can you possibly think it necessary to then say "if they do not, fix them"??? Well, duh! Everybody knows that, and to bother saying such a thing is a beautiful example of completely unnecessary instruction bloat.
• Similarly, everyone knows that a redlink should be checked to see if adjusting it will turn it blue (assuming they know what a redlink is -- and if they don't they won't have understood 80% of anything else being said in this guideline anyway).
• What's an "internal link"?
• "When there is not yet an article about the subject, a good link will make the creation of a correctly named article much easier for subsequent writers." Blah blah blah blah. Once you remind editors to use good naming conventions, that's all that's needed. If they don't see why good naming conventions should be followed, then they're not going to understand the naming conventions anyway, so all this motivation is just so much instruction bloat. In fact, I've made a further edit [26] along those lines.

7. You seem preoccupied with whether new text says exacty what the old text said, but that's not what matters. What matters is whether the new text as effectively, or more effectively, induces the behavior in editors we want to induce. Technical or specialized terms should be linked (at least on their first appearance in any given article) plus if possible they should be glossed where used, so that the reader might be able to skip chasing the link and still generally understand the passage -- he or she can follow the link or not depending on the level of understanding desired. The new text makes it clear that glossing is in addition to, not instead of, appropriate linking. As elsewhere, I look forward to other editors' thoughts on whether this is an improvement or not.

8. (You're talking about [27]). I look forward to whether other editors agree that the old text --

Don't assume that readers will be able to access a link at all; remember that it is not always possible. For example, a reader might be working from a printed copy of an article without access to facilities for following links.

-- wasn't bloat rightly reduced to the new text:

Note that readers working from hardcopy cannot follow links at all.

BTW, the idea that parentheses are inherently clumsy is the kind of mindless rule relied upon by people who don't know how to make stylistic decisions for themselves; parenthesizing this motivating aside is completely appropriate.

9. No, SEAOFBLUE is specifically about immediately adjacent links that appear visually to be a single link. The text we're discussing is about the overall density of links in a particular kind of article as a whole. I therefore moved it to the section on over- and underlinking, which is where it belongs.

10. Saying techincal articles "might demand" more links is effectively no different from saying they "often have" more links -- one is a statement of what the article "needs" and the other states what editors typically "give them". Either way the reader understands that more links are appropriate. You have extremely rigid ideas about the way ideas should be expressed.

11. Contrary to what you imply, the new text says "do not" just like the old text does.

12. What's with it is that this text specifically regards linking in the lead, so I moved it to the section on linking in the lead (with a very slight change in wording). If you'd actually looked at the edit in context you'd know that.

EEng (talk) 04:22, 4 July 2015 (UTC)

Of course, I am also interested to hear what other editors think about the changes. You must be familiar with the WP:BRD cycle (bold, revert, discuss)? You made a series of bold edits, you got reverted, but instead of discussing your bold changes, you decided to re-revert. And not just you ignored the discussion part, you started demanding other editors to explain why your edits do not satisfy (!). So you recently made bold changes to the stable version of the guideline, and now you are defending it as it'd be the long-term consensus.

Besides, you first criticized that "the objection was vague, not substantive". In your later response, however, you admit that the bold changes did indeed change the tone and meaning, but now you argument that the changes are an improvement in your opinion. Jayaguru-Shishya (talk) 00:01, 5 July 2015 (UTC)