Wikipedia talk:Manual of Style/Linking

Section is confusing -For instance, please clarify what is meant by "contrasting examples"

The section "What generally should not be linked" as written is rather unclear.

First, "This generally includes major examples of geographic features". It's a bit confusing because the word "this" could refer to "names of subjects with which most readers will be at least somewhat familiar" OR it could refer to "unless there is a contextually important reason to link". I think it means the former, but the ambiguity makes this a bit hard to understand.

Second, what is meant by the sub-bullet "Pairs of contrasting examples:". This isn't clear at all. Can someone explain? What pairs? How is "Prussian" a pair? How are these contrasting?

Third, what is meant by "These are two ends of a spectrum"? Does the word "these" refer to the previous two sub-bullets?

It would be great if this whole section were written a bit more clearly. Coastside (talk) 20:36, 30 December 2018 (UTC)

Overlink

I'd like to get more information about WP:OVERLINK. I think the policy is very much like MOS:BOLD and similarly people can't get used to overdoing it. I think one or two links per sentence should in most cases be more than enough. Particularly I see people linking to "review aggregator Rotten Tomatoes" and it seems to me a very clear example of overlink, since if readers need an explanation they'll get it from the Rotten Tomatoes link. Similarly editors refer to Metacritic and then explain that it uses a weighted average, and I think most readers know what an average means, and the weighted average article is too generic to be really useful.

I discussed this case overlink in film articles before with the people active on WP:MOSFILM and there was a weak consensus in favor of continuing to link to weighted average. Others seem to think this discussion amounted was a consensus in favor of keeping both links, I don't see that at all. I'd like to know more about the reasoning from those who are behind the Overlink policy and if they think it should apply and if a (weak) local consensus is enough to ingore it. -- 109.76.143.176 (talk) 19:30, 31 December 2018 (UTC)

Creating a WP:SEAOFBLUE is not what OVERLINK is about. OVERLINK is about linking common terms like "film", "movie" or "actor" in your example; "singer", "guitar" or "song" in a band or album article; or generally "United States", "Himalayas or "Christianity". Walter Görlitz (talk) 21:43, 31 December 2018 (UTC)

IP, you were told there by Betty Logan and others that linking terms that readers are likely not familiar with is not overlinking. In what way do you think general readers will usually understand what "weighted average" is? As for linking "review aggregator," that's on the iffy side since Rotten Tomatoes and Metacritic will already be linked, but I don't consider it overlinking when Rotten Tomatoes and Metacritic are also linked. Flyer22 Reborn (talk) 22:30, 31 December 2018 (UTC)

Also, we've have readers not understand when we state "average rating" (instead of "rating average" or "weighted average"). Some have thought that we are simply stating that the rating is so-so/mediocre. Linking in the case of "average rating" is definitely helpful. Flyer22 Reborn (talk) 22:35, 31 December 2018 (UTC)

In reply to the well considered opinion of User:Betty Logan in the previous discussion, I said that we can reasonably expect readers understand "average" (even if it is a specific type of average). Rotten Tomatoes and Metacritic are the most relevant links possible, and those are already linked, so call it OVERLINK or SEAOFBLUE but my point is that those additional links are redundant.

Frankly I don't even think the boilerplate explanation of aggregator scores are helpful either, but linking the terminology seems even more unnecessary. Flyer22 keeps saying "helpful" links but we don't know for sure that those links are helpful, or that people click on the RT and MC links either. -- 109.76.143.176 (talk) 23:06, 31 December 2018 (UTC)

I'm not going to keep debating you on this. I am not the only editor who has stated that the links are helpful. Here at Wikipedia talk:WikiProject Film, another editor was even clear that they can be helpful. Except for WP:SEAOFBLUE issue, which is easily remedied by changing "review aggregator Rotten Tomatoes" to "Rotten Tomatoes, a review aggregator," you have not shown how they are unhelpful. You removing these links deprives some readers of understanding; we know this not only by common sense, but by the "average rating" example I gave above. And, no, I'm not going to go digging through histories or my contributions to point to examples in that regard. Here's the thing: You do not have the MOS support or the support of film editors to keep removing these links. And the benefit of keeping them outweighs the rationale to remove them. Flyer22 Reborn (talk) 00:07, 1 January 2019 (UTC)

Flyer22 disagrees with specific changes I've made, and doesn't want to discuss this further. I'm am looking for further discussion, and different opinions. So far the comment User:Walter Görlitz suggests I shouldn't worry about it, but perhaps he could clarify with a follow up comment. I don't think he was saying that the MOS doesn't discourage redundant linking, just that what I was saying wasn't specifically Overlink. I'd welcome other opinions too, I'm don't expect rapid response to any discussions at this time of year, or for the discussion to be over without a few more opinions to make things clearer.

I think a lot of editors use linking to indirectly address other problems. What this discussion has made clearer to me, is that the underlying reason this bothers me, is the tedious and pedant boilerplate text attempting explain details about aggregators that I don't believe readers even care about. Flyer22 says the link is helpful and that user need the terms explained with a wikilink, but to me that makes it looks like the boilerplate is not actually helpful and readers end up needing to look at the Metacritic article anyway and that is best to just link the key word in the sentence. The redundant linking is a scab on an underling greater irritation but that doesn't make either of those things good.

I'm reminded by MOS:BOLD, because Wikipedia used to have lot of bold text all over the place and it seemed harmless, and took a while for me get behind that guideline and to see that after people had removed so much bold that less was more. So reading the linking guidelines and seeing phrases such as "links compete with each other for user attention" it seemed crystal clear, a reminder that less is more. Maybe I'm being overly strict but it seemed like that too when editors started removing all the excess bold formatting, and only in hindsight was it obviously an improvement. -- 109.76.149.139 (talk) 02:49, 1 January 2019 (UTC)

I never stated that "the user need[s] the terms explained with a wikilink," as if all users do. I stated that linking a term such "weighted average" is helpful, and I've already noted why. Obviously, some won't need "weighted average" linked. I don't want to discuss this further because you've WP:Forum shopped this topic around to three different places now (well, two if considering the fact that I started the discussion at WP:FILM about you) looking for the support you want and haven't gotten it (except for Tony1 below), and that is due to your arguments for de-linking being weak, especially with regard to "average rating," "weighted average" and "rating average." Linking that is not redundant linking. I don't agree with your "boilerplate" reasoning, or see this as the same as unnecessary bolding. The next step is for you to start a WP:RfC on this, preferably at WP:FILM or MOS:FILM. If I see you continuing to try to enforce your style across Wikipedia, I will revert you for WP:Disruptive editing. You need to stop edit warring to get your preferred style to stick. Flyer22 Reborn (talk) 23:56, 1 January 2019 (UTC)

Looking at UNDERLINK for contrast "An article is said to be underlinked if words are not linked and are needed to aid understanding of the article" and to understand an article about film, a reader does not need to understand the score system, be it 4 out of 5 stars or 83% or 22 out of 100, so I come back to my point that the not only is the link unnecessary but the boilerplate explanatory text is also unnecessary. I want to make it clear that an article definitely wouldn't be underlinked by not including these links.

Stylistically and from my personal experience with SEO, I think more than one link per sentence clause is bad, and I would aim for less than one link per sentence or less. Again from SEO I would generally link to nouns, occasionally to verbs, and very rarely adjectives. (So link to the proper noun RT, MC, not the descriptions, review aggregation website).

I would still very much like the people who came up with the OVERLINK policy in the first place to make it much clearer and provide a much more specific metric or provide an objective measure or tool that would allow us to settle this conclusively. I wish we didn't even need to have this discussion and it could be decided less by opinions and more by facts and evidence and ultimately I find the lack of clarity in the OVERLINK guidelines severely disappointing. -- 109.76.149.91 (talk) 22:44, 30 January 2019 (UTC)

• I generally agree with the IP. (Please consider registering ... it only takes a few minutes.) Tony (talk) 11:20, 1 January 2019 (UTC)
  • So Flyer22, your reversions would not be disruptive—is that it? Tony (talk) 00:52, 2 January 2019 (UTC)

Correct. Because, as seen here, the IP would be enforcing a style that has no WP:Consensus, and actually contrasts the style that most or many of our film articles use. The IP in this recent edit points to the 2010 Wikipedia talk:Manual of Style/Film/Archive 6#Metacritic's so-called "normalized" scores discussion. That is the discussion that made it so that "normalized" was replaced with "weighted average." And we can see at Wikipedia talk:Review aggregators#Overlink that there is no WP:Consensus on the de-linking arguments that IP 109.76 favors, and that more editors lean toward at least keeping "weighted average" linked. I would revert the IP at the articles once, then take the matter to the appropriate talk page or noticeboard, just like I first took the matter to Wikipedia talk:WikiProject Film#Removing helpful wikilinks from the Critical reception section for the IP and others to discuss. In that discussion, we can clearly see Erik stating, "I do not support going around with the express purpose of de-linking them." And I've stated similarly. I've reverted a number of editors (IPs or otherwise) trying to enforce their style on Wikipedia articles and edit warring with others while doing it; I reverted for disruptive editing, and then took the matter to a talk page or reported the editor to the appropriate noticeboard. I was never the one to get reprimanded or blocked for reverting the disruptive editing; and it wouldn't happen this time either. Flyer22 Reborn (talk) 02:35, 2 January 2019 (UTC) Flyer22 Reborn (talk) 02:41, 2 January 2019 (UTC)

While I hope this discussion isn't dead, it has had even less participation than the MOSFILM discussion, and realistically I don't think anyone can claim consensus based on any discussion involving only 3 people.

I reiterate my call for OVERLINK to include more objective measures and clearer evidence based rules. -- 109.76.149.91 (talk) 22:58, 30 January 2019 (UTC)

From the lack of interest in your proposal and two opposing voices, I'd say that there is no consensus for change. Walter Görlitz (talk) 23:33, 30 January 2019 (UTC)

Reiterate .. what? You started this discussion with: "I'd like to get more information about WP:OVERLINK." Which sure looks like a simple request for information, and why should more than one or two people be involved? If out of that discussion you come up with an idea that might warrant more discussion, that is a separate discussion, which needs it's own thread, outside of an insignificant request for information. ♦ J. Johnson (JJ) (talk) 23:57, 30 January 2019 (UTC)

When to link common terms

I have been flummoxed by this as well. Usually, this comes up in the context of album articles for me: I will link all of the instruments played by musicians in the personnel section and others will remove some of the links citing this page. But if we remove links to guitar because it's a common term, then when we we ever include any links to it? I don't include links to guitar in standard running text but a list section within an article seems like a perfectly appropriate place for it. What am I missing here? ―Justin (koavf)❤T☮C☺M☯ 02:23, 18 January 2019 (UTC)

The reasons that linking to such a commonly known term such as "guitar" is discouraged are that: (i) it dilutes the linking system, which is stronger if rationed to items that readers might be slightly more likely to click on, such as technical/specialist terms and articles on much less-well-known musical instruments; (ii) if spread everywhere without editorial judgment, linking becomes visually disruptive to readers. So, no, please avoid linking to all but instruments that are uncommon. Certainly not the ubiquitous "guitar". You'd save us a lot of trouble if you followed the guideline. Thanks. Tony (talk) 02:29, 18 January 2019 (UTC)

It's not appropriate in a list section either. That it's done there more often is simply because of sloppy editing. Walter Görlitz (talk) 04:44, 18 January 2019 (UTC)

I'm opposed to overlinking, but I must admit that a list of linked uncommon terms interspersed with an unlinked common term look very odd. I think that should be left to editorial judgment without rigidly insisting on the guideline. -- Michael Bednarek (talk) 09:50, 18 January 2019 (UTC)

I understand your point, Michael. What I think we called the "patchiness" issue was discussed in some depth when en.WP got serious about smart linking practices, back last decade. It's true, the look isn't ideal, and yes, it does come down to editors' judgment to balance patchy visuals with the need to optimise link functionality. As an example, I've just unlinked my umpteenth "singer-songwriter" in an infobox, but left the intriguing author song linked. I thought it might have beena typo, but no. AND I delinked my umpteenth "vocals" and "piano", but left linked gusli. Fair enough? Koavf, don't be flummoxed: selective linking (using our editorial skills and field knowledge) is a service to readers. Tony (talk) 11:33, 18 January 2019 (UTC)

@Tony1 and Walter Görlitz: You are misunderstanding or ignoring my point (maybe I am not making it clear): If I followed this guideline, then guitar would be orphaned, no? As I asked above: when should someone link to guitar? ―Justin (koavf)❤T☮C☺M☯ 04:20, 20 January 2019 (UTC)

I did miss that point. It wouldn't be orphaned because of album personnel listings not linking it, but close associations, like with notable guitarists, would all link back to the article. But do we need to link to the instrument that Ryan Hedgecock played on Lone Justice (album) to know what a guitar is? It goes against OVERLINK. It's the same reason that American and 1985 are not linked in the lede of that article. Walter Görlitz (talk) 04:39, 20 January 2019 (UTC)

I'd link it in any article specifically on musical instruments, or when "guitar" takes on a technical or historical dimension. But in those cases, it might be best to link to a section of the article on "guitar". Tony (talk) 08:45, 20 January 2019 (UTC)

Thanks for this. ―Justin (koavf)❤T☮C☺M☯ 23:04, 23 January 2019 (UTC)

The term "enumerated"

The article sequence begins with "In mathematics, a sequence is an enumerated collection of objects in which repetitions are allowed." In this sentence, is the term "enumerated" common enough that it should not be linked? There is an article on enumeration, but it can be argued that the word can be understood in context. Shawnc (talk) 20:04, 2 February 2019 (UTC)

MOS:DUPLINK revisited

I'm aware this has been discussed at length before. My earliest memory of editing is of encountering such a discussion. But that was five years ago, and I think enough time has passed to invoke WP:CCC. An RfC would be in order, but first I wanted to test the water on this (i.e., there is no point in !voting; save that for the RfC if any).

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

The current guidance seems predicated on two assumptions: (1) That articles are read top down, rather than using the table of contents to skip to a section of interest (why make links of TOC entries, then?), and (2) that the reader will generally follow the link upon first reference (or first reference after the lead) if they don't know what it refers to. Unless both are true, upon encountering an unfamiliar term that is not linked, the reader is forced to (1) somehow know that it was linked earlier, despite not having read that part, and (2) go find that link. It's easier to just use the Search box, defeating the benefit of links—and either way, you lose your place in the article unless you open a second window or browser tab.

Even if the reader reads top-down and follows the link upon first reference (or first reference after the lead), are they expected to read the entire target article? What if they read only the first few sentences, and then want to know more when they encounter the term again later? Is that an implausible or uncommon scenario?

The guidance just doesn't make sense to me, and I suspect a large part of the reason it's ignored so much is because it doesn't make sense to other editors, either.

"First reference in a section" wouldn't be perfect (I doubt anything would be), but it would be a substantial improvement. ―Mandruss ☎ 00:15, 1 March 2019 (UTC)

I'm sure there are some issues to be ironed out (section? subsection?) but I've often thought along similar lines. EEng 04:38, 1 March 2019 (UTC)

Mandruss, you write: "It's easier to just use the search box, defeating the benefit of links—and either way, you lose your place in the article unless you open a second window or browser tab." Do you mean that clicking on a link in situ does not make a reader lose their place in the article?

In any case, the problem with changing the guideline in this respect is that there would be no end of multiple linking. Remember that surprisingly few links are clicked on, and for every one that is, probably hundreds of readers don't want or need a slightly disruptive splash of blue. As for encountering a word one is unfamiliar with via a section link ... well, why not scroll up after finishing that section? Most readers, I'm sure, wouldn't mind quickly surveying the larger context before the link-targeted section, if they really can't find it within themselves to type the word into the search box—especially to where the item is first linked, which may well avoid the need to divert to another article in the first place. Tony (talk) 12:02, 1 March 2019 (UTC)

Tony1: I can middle-click a blue link and that will open a new tab in the browser; I'm not aware of any method that uses the Wikipedia search box that allows that. // I, too, often find that I have to go back to locate a link for a term that was clear from its context before, but later baffles me. There used to be a bullet point here that allowed repeated links "where the later occurrence is a long way from the first" until it was removed by User:Maunus on 27 November 2011. I can't find any discussion that led to that edit, but it was discussed between February and May 2015. According to the closing statement there, that discussion reached "consensus for allowing when far enough apart." But I can't find any edits that implemented it. -- Michael Bednarek (talk) 13:30, 1 March 2019 (UTC)

Do you mean that clicking on a link in situ does not make a reader lose their place in the article? Yes I did, but I suppose it does. Stricken, not that I feel that tilts the scale in the other direction. ―Mandruss ☎ 15:26, 1 March 2019 (UTC)

When I encounter a sea-of-blue article, I'm generally not inclined to spend my time making it conform with a guideline that makes absolutely no sense to me. I'm a believer in respect for consensus, but not absolutely so. So I leave the sea-of-blue alone, while "first reference in section" would reduce the blue significantly. So your argument self-defeats per the Law of Unintended Consequences. In any case, surely providing some reasonable minimum ease-of-use comes before minimization of blue. No, it's not reasonable to ask a reader to go find a link when they need it, and if they are to use the Search box we can eliminate all blue very easily. ―Mandruss ☎ 15:33, 1 March 2019 (UTC)

"When I encounter a sea-of-blue article, I'm generally not inclined to spend my time making it conform with a guideline that makes absolutely no sense to me" – OK, so you're for maximal linking of WP's article text. The lobby for that usually comes from techie people who don't understand why you wouldn't maximise "utility". The argument against comes from people (like me) who (i) value reading text without undue obstruction by bright blue splashes, (ii) feel that the linking system is degraded/diluted unless it's carefully rationed, (iii) find little value in catering to readers who might article-hop, browsing superficially by links as one might when shoppping online—particularly given that clicking on a link goes against the whole notion of article cohesion and structure, by diverting suddenly in the middle of a text; (iv) are suspicious of the notion that internal links are used more than marginally by readers, and (v) wonder why it's such a pain to spend five seconds typing a target into the search box ... that's what it's there for. These issues were discussed ad nauseum in 2008–09, and community consensus was in favour of carefully limiting linking. If you want to see examples of uncontrolled, undisciplined, scattergun linking, take a look at some of the non-English WPs, where the sea-of-blue tech lobby has been too daunting for development of a more disciplined system, as en.WP has done. Tony (talk) 03:28, 2 March 2019 (UTC)

OK, so you're for maximal linking of WP's article text. No, what I said was "first reference in a section". That's less than maximal.
Most of the sea-of-blue problem comes from linking things that don't need to be linked at all, not from repeating links of things that do. There are many editors going around linking everything in sight, with virtually no sense of what's really useful, and there are a lot more of them than there are of us since that's something one can do with very little learning. If DUPLINK were changed as I suggest, that would still be the major cause of sea-of-blue. But I accept that there are competing philosophies and conflicting goals that will forever remain unresolved, resulting in very little site-wide coherence or consistency in linking. ―Mandruss ☎ 04:33, 2 March 2019 (UTC)

I'm against changing it. Flyer22 Reborn (talk) 01:20, 3 March 2019 (UTC)

Templates used as links in articles

MOS:Internal links has this point: Do not create links to user, WikiProject, essay or draft pages in articles, except in articles about Wikipedia itself (see Self-references to avoid) - what about links to templates which act as pseudo-articles, such as those linked in the season column here? --Gonnym (talk) 20:41, 1 March 2019 (UTC)

Would it hurt to give a bit of context? I have no opinion on the significance of the context, which I will leave to others, but it's surely relevant. ―Mandruss ☎ 22:42, 1 March 2019 (UTC)

Anyone of the watchers here with a comment? Would really appreciate if I don't have to open a RfC just for a simple answer. --Gonnym (talk) 23:01, 9 March 2019 (UTC)

We should never link directly into the template space as a matter of surprise, unless a) it is obvious the link goes to the template space and/or B) the link is provided for maintenance purposes. --Izno (talk) 00:46, 10 March 2019 (UTC)

What do you count as an obvious link to a template? Does it need be spelled directly as Template:Template name? If that is indeed the case, then how is that a valid option and not just transcluding it? --Gonnym (talk) 09:09, 10 March 2019 (UTC)

It either needs the exact title or it needs something like Template. I prefer the former to the latter, but neither to the other option which is transclusion; if it is indeed material to be transcluded, it should be transcluded. --Izno (talk) 14:01, 10 March 2019 (UTC)

Do you have a live example you know of a template linked the way you say is valid? I'm not quite sure I understand how that link should work like (again, from main space and not from project-space). --Gonnym (talk) 17:07, 10 March 2019 (UTC)

Do not confuse "better than bad" with "valid". :^) The latter "maintenance" case is for e.g. the VTE links Template:navbar. --Izno (talk) 17:16, 10 March 2019 (UTC)

The VTE links weren't my issue as those are valid links, was refering more to how links in here appear. --Gonnym (talk) 11:05, 11 March 2019 (UTC)

That's definitely wrong-headed linking to templates. The obvious fix is the create an article and transclude the templates with sections (or at least anchors) and link to the proper part of the article.  — SMcCandlish ☏ ¢ 😼  10:10, 11 March 2019 (UTC)

The problem with Big Sky Conference#Big Sky men's basketball and other recent TfD discussions similar to this was that editors argued that linking to template directly is ok as there isn't any place that says otherwise. --Gonnym (talk) 11:05, 11 March 2019 (UTC)

I.e., WP:WIKILAWYERing. We don't link to template pages from mainspace. It's just not done. A site-wise consensus on what we do and don't do that's this self-evident doesn't need to be written down. This is one of those WP:AJR situations. We can write it down if someone's going to light WP:Common sense on fire to get their way, but if it's not actually happening on any kind of scale, we're better without adding yet another written rule, and fixing the few instances of people being boneheads and linking to templates from articles.  — SMcCandlish ☏ ¢ 😼  14:20, 11 March 2019 (UTC)

Contrasting examples of what not to link?

I have tried hard but can make no sense of the following under "what generally should not be linked". Here is the bullet point including the (for me) opaque comment:

• Pairs of contrasting examples: Lake Xochimilco, mid-Atlantic Ridge; Aberbeeg, Tonga; Frisian languages, Middle English; Prussian, Tohono O'odham, Afro-Cuban; Zoroastrianism, Shinreikyo<!--Good mix, including temporally and across definitions of things like "nation".-->

Are these words that should not (normally) be linked? If so, was it a good idea to link them? Or should they be linked? Or are they pairs where one should be linked and the other not (the next bullet point says "These are two ends of a spectrum")? If so which way round? And why, scrutinising the punctuation, is there a threesome in there?

I've been pondering this even more before posting and I wonder if these are things that should normally be linked, in contrast to the items in the previous bullet point that should not. If so, the way this has been expressed is not at all clear. Yes, I think that is what is intended but what do others think? Thincat (talk) 22:08, 9 March 2019 (UTC)

• I've always found this section puzzling. It's especially weird that the bit you quote is introduced as names of subjects with which most readers will be at least somewhat familiar. Now, I'm gonna go out on a limb and say that the idea that "most" readers are "at least somewhat familiar" with (or have even heard of) Lake Xochimilco, Aberbeeg, Tohono O'odham, and Shinreikyo reflects a vast overestimation of most readers' breadth of experience. I have no idea what the "contrasting examples" thing is. EEng 23:44, 9 March 2019 (UTC)
• @SMcCandlish: who, I think, added this originally, @BeenAroundAWhile: who, somewhat doubtfully, changed it and @EEng: who changed it again. Does anyone have any second thoughts? My feeling is that this bullet point (and the one following} should be removed. Such obscure topics generally should be linked and yet they are so obscure that they are not helpful topics to illustrate the point. Thincat (talk) 10:01, 11 March 2019 (UTC)
• Yeah, I don't know what this is supposed to mean either. I know what I used to mean. The early version made sense (though was possible to misinterpret if you tried hard). I've reverted to that version, then clarified the iffy phrase. This had nothing to do with "pairs of contrasting examples"; rather, the entire block is "do link things like this", in contrast to "don't link things like the prior list".  — SMcCandlish ☏ ¢ 😼  10:22, 11 March 2019 (UTC)
• I am similarly confused. How did these examples ever make it into the guideline? Tony (talk) 10:42, 11 March 2019 (UTC)
  • It's a set of examples of what not to link, like "New York" and "English" and "African American" and "Buddhist". This is followed by a list of counter-examples, of places, languages, peoples, religions not so commonly known that we shouldn't link them. The examples are pretty random. Someone later came along and somehow mistook the second list for something completely different, a list of link-this-one-but-don't-link-that-one pairs, despite them all being obscure to at least some subset of readers, and not actually being in pairs. Whatever the cause of this, it's been repaired and clarified; no more issue.  — SMcCandlish ☏ ¢ 😼  14:12, 11 March 2019 (UTC)