Wikipedia talk:Only make links that are relevant to the context/Archive 7

Revert the nutshell

So what part of Bold, Revert, Discuss is unclear here? This page started without a nutshell. Someone added it. That edit was reverted. Some time later, it was readded. It was reverted again with an explicit request to discuss on Talk before readding it. It was promptly added yet again, still without any discussion or explanation.

Nutshells in general are more harmful than helpful for our readers. They give the false impression to casual readers that by reading only the nutshell, they understand all the nuance and subtlety that the rest of the page explains. They have created dangerous false impressions among our new users and pointless confusion and wiki-lawyering.

Nutshells are also completely and utterly redundant to either a) a well-written, active-voice page title and/or b) a well-written and succinct introductory paragraph. This page has both. There may be marginal value to a nutshell on other pages which have ambiguous titles or poorly written opening sections but there is zero value to the nutshell on this particular page.

Other than satisfying some slavish desire for a false sense of consistency, what value is there in a nutshell on this specific page? Rossami (talk) 04:06, 19 August 2008 (UTC)

For the record, I did check for a discussion thread on the talk page, and found none, before adding an attempted compromise nutshell. When requesting discussion, you should lead by example. The nutshell argument you make is not new, and the widespread use of nutshells suggests that it is not a widely held consensus. I think your supporting argument that the nutshell gives the impression that it conveys the full substance of the policy or guideline is completely baseless, as this would be an absurd conclusion for a new user to draw. The reason for nutshells is to help new users digest WP policy. They allow new users to take a small taste of a new dish first, before attempting to devour the whole portion. They also assist in navigation, allowing users to quickly see if this is the policy they are looking for. Several editors do object to them as being unnecessary, or clutter, but if they are kept very brief, and do not grow out of control, I think it is a reasonable compromise to keep them. Dhaluza (talk) 10:32, 23 August 2008 (UTC)

I for one don't like the "nutshell style" opening that is not actually in a nutshell -- it looks awkward. (By "nutshell style", I mean the bolded repetition of the article's title followed by two summarizing bullet points.) If the opening came in the form of prose, it would be smoother; alternatively, encapsulating it in a nutshell template would, I feel, improve its formatting and make its "short overview" qualities more pronounced.--Father Goose (talk) 05:32, 19 August 2008 (UTC)

I think repetition of the title in the nutshell is redundant. It should be repeated in the opening paragraph as is standard practice, and it already appears at the top of the page. Dhaluza (talk) 10:35, 23 August 2008 (UTC)

Yes, agreed, and it's just as bad when the repetition occurs in a {{nutshell}}-less nutshell at the top of the page. I'm fine with this layout.--Father Goose (talk) 19:48, 23 August 2008 (UTC)

I added the {{nutshell}} template for precisely the reasons that Father Goose says. You'll notice that the only difference between the version with and the version without the template is that the nutshell text is placed in the nutshell template. And that is exactly what that piece of text is: it's WP:CONTEXT in a nutshell. That is, it explains the fundamental point and goal of the policy/guideline. I don't think we can or should expect all of our new editors to go and read the complete text of all of our policies and guidelines before helping improve the encyclopedia. The commonsense approach to editing Wikipedia is to be bold, do what seems necessary, and then review the policies and guidelines in detail as issues come up. Having the nutshell template at the top gives you the point of the guideline without making you digest 25K of rule if you don't need to.--Aervanath lives in the Orphanage 06:57, 19 August 2008 (UTC)

It's no big deal, but yes, I was a little disappointed to see the nutshell go. What's wrong with it? Tony (talk) 07:19, 19 August 2008 (UTC)

Linking dates for autoformatting

I'm confused by the recent changes. Are we supposed to stop linking dates for autoformatting or not? And why can't this page just say yes or no in clear unambiguous language? Dhaluza (talk) 01:07, 26 August 2008 (UTC)

The Manual of Style now says "stop" except in rare cases. I'm still working through the history of the recent discussions but if that decision sticks over at the MoS page, this page should be updated to follow the new rule. Rossami (talk) 02:24, 26 August 2008 (UTC)

Consensus at MOSNUM, after a long, long debate, has been that "Autoformatting should not generally be used unless there is a good reason to do so." So yes, the answer is that we should all stop linking dates. Tony (talk) 02:39, 26 August 2008 (UTC)

Dhaluza has a point, though. The section as currently worded is unnecessarily long, verbose and confusing. I'm not sure of how exactly to reword it at the moment, but it could definitely be made more concise.--Aervanath lives in the Orphanage 15:52, 30 August 2008 (UTC)

Just to be clear though, the changes to this page discouraging dates have been added since that argument there. -LlywelynII (talk) 22:21, 2 October 2008 (UTC)

What generally should not be linked

The current guidance says to not link well-known geographical locations, and examples are given of major countries and cities. I'd like this guidance to be extended. There is rarely any benefit to readers in linking widely familiar concepts such as:

• the names of the continents
• the major oceans and seas (Atlantic, Pacific, Mediterranean, Caribbean, ...)
• major astronomical objects (Earth, Moon, Sun, ...)
• large-scale historical events (World War I, World War II, Industrial Revolution, Middle Ages, ...)
• common professions (actor, singer, musician, writer, author, scientist, artist, physician/doctor, politician, ...)
• names of languages [a link to these is almost never beneficial unless specifically discussing the language itself]
• major religions (Christianity, Islam, Judaism, ...)
• nationalities related to well-known countries (American, British, French, Russian, Japanese, Chinese, Spanish, Italian, ...)
• names of colours (red, white, blue, green, yellow, white, black, orange, brown, gray/grey, ...)
• common human events (birth, death, marriage, divorce, ...)
• major parts of the body (head, hair, eye, nose, mouth, ear, chin, neck, arm, leg, finger, toe, ...)

Clearly there will always be times when it is appropriate make these links, and there can be debate about exactly which items should be considered familiar enough not to need linking. I think the examples I've given are a minimum set that practically everyone could agree on. Any comments? Colonies Chris (talk) 15:47, 26 August 2008 (UTC)

I agree. Is there any simple test that can be applied? We use the phrase 'plain english'. Perhaps we should also mention that Wikipedia is not a dictionary. Lightmouse (talk) 16:14, 26 August 2008 (UTC)

Oh please yes, at long last. I was getting a head of steam up to add the languages, but this wider approach is much better. Do people really link body parts and common events? I'd have thought these might not need explicit mention, but could be covered by the "dictionary words" category. Tony (talk) 16:18, 26 August 2008 (UTC)

• I think that some of the proposed examples are overly specific. I can think of many articles where a link to Europe would be entirely appropriate - and some examples where it would be pretty pointless. A blanket rule would seem to create more room for confusion and dissent than the policy already has. I would also worry about accusations of inconsistency and bias if we decided that, for example, Christianity is a major religion that doesn't get linked but Baalism does. Or that "John Smith is an Estonian chiropractor" should be inherently different in its presentation from "Jane Smith is a US doctor". Again, I agree with the principle but see too many opportunities for exception and counter-example to see much potential for a blanket rule. Those excessive links can be quietly removed by any editor already. We don't need to expand the rule before you start. Rossami (talk) 16:22, 26 August 2008 (UTC)
  • The dilution of our high-value links by an encroaching "sea of blue" has been a problem on WP for some time. We've moved away from the original scattergun "link anything you please" approach to the notion of disciplined linking, or smart linking, in which a little care is expected in prioritising what and what not to link. The boundary between these will often vary depending on context and editor; whether to link "France" (hardly ever), "Paris" (usually not), "Bordeaux" (probably—depends), or "Gironde" (almost certainly) does require, inter alia, an assessment of how well-known these items are likely to be to the readers of the article and English-speakers in general. It's what you might call a "wiki-skill", which we should proudly encourage all WPians to take seriously as a rather special, if not unique, part of this encyclopedia.
    And no, I think the entrenched habit of link-as-you-please does need to be addressed through explicit guidance, such as the list above. However, yes, you're right in that the advice should not be cast as a straightjacket, but worded in such a way that editors can develop and use such skill in context. This is easy enough to do, through the use of the old warhorses "generally", and "unless there is a good reason to do so". That leaves the final decision in the hands of editors, but provides global advice to trigger in editors' minds where a link should be justified by good reason. Having said that, can you provide an example of where a link to "Europe" might significantly increase readers' understanding (as required by MoS main)? Tony (talk) 16:59, 26 August 2008 (UTC)
    • For an off-the-cuff example, the link to "Europe" in continent is clearly appropriate since it helps to draw the distinction between Continental Europe the contiguous landmass , Europe the larger landmass including the UK, Iceland and Turkey and Europe the political entity, to name a few. I'm sure I could find many more examples.
      I agree with the general premise that we should be more deliberate in deciding which links to create. I'm merely skeptical that expanding the examples will clarify the decision any more than what's already on the page. There are too many exceptions to any given rule. Worse, the more detail you add, the more likely new readers are to say "it's not on the list, therefore I should ..." rather than to think and apply the basic principle. It's a classic problem of instruction creep. As you say, deciding which links are appropriate in a particular context is a skill that must be learned.
      (Interestingly, though, the confusion seems to have gotten a bit worse since the removal of this example which had been in the introduction practically since the earliest days of the page. Perhaps that should be added back.) Rossami (talk) 18:26, 26 August 2008 (UTC)
      • Rossami, you have a point about the danger of creating an "it's not on the list, therefore I should link it" mindset. Perhaps it would be better cast as a more general statement, along the lines of

"You should not normally link: the names of geographical locations (continents, oceans, countries, cities), and related nationalities, that are likely to be well-known to English-speakers; major astronomical objects; large-scale historical events such as the two world wars; common professions; names of languages; major religions; and other items that are likely to be familiar to English speakers."

That gives strong guidance but leaves room for individual judgment.

And yes, Tony, I'm afraid people really do link to parts of the body (often in fan articles)

"(insert heartthrob name here) has blue eyes and black hair"

and as for common human events, there's a bit of boilerplate generated text you find in hundreds of articles about US towns:

“ There were 4,235 households out of which 28.8% had children under the age of 18 living with them, 43.5% were married couples living together, 11.2% had a female householder with no husband present, and 41.7% were non-families. 35.4% of all households were made up of individuals and 13.6% had someone living alone who was 65 years of age or older. The average household size was 2.26 and the average family size was 2.93. ”

Pick any common word such as daughter and you'll find hundreds of links, very few of them necessary or worthwhile. Also, I could have mentioned days of the week and month names but they're covered elsewhere. Colonies Chris (talk) 07:37, 27 August 2008 (UTC)

Yes, it is quite disturbing when you go to a page and look at 'What links here'. The list at Most linked-to pages contains many 'plain english' terms and that does not even measure repeated links in a page. A few of us keep doing purges on days of the week but they keep coming back. Lightmouse (talk) 21:28, 26 August 2008 (UTC)

Days of the week are a very good example of things that should not be linked in most contexts. But the articles on the days of the week should link to at least the day before and after, and articles on specific days of the week, like Good Friday and Friday the 13th should link to the generic day and vice versa. So again, it's context that's important, not the specific item. Dhaluza (talk) 11:59, 27 August 2008 (UTC)

Rossami, long ago, I removed this from the lead for two reasons: "It is not always an easy call. Linking to the number three from triangle may be helpful, while linking to the number six from Six O'Clock News would not be." because (1) "three" in that context is a good example of what not to link, and (2) the distinction between the appropriateness of the three and of the six is quite unclear. Tony (talk) 01:55, 27 August 2008 (UTC)

To the extent that it makes sense to have an encyclopedia article on a common number like three at all, there are few pages where it would make more sense to link to that page than from triangle - a geometric shape that is defined by its three-ness. If "triangle" doesn't qualify for a link to "3" in your view, I can't imagine what article would. (And if no page should link to the number, why would we have an article on an orphaned topic?) The Six O'clock news, on the other hand, has no greater connection to the number six than the chance of a corporate scheduling decision. A link there is clearly inappropriate.
Obviously, I didn't disagree with your decision enough to revert that removal at the time. But in hindsight, it seems to be the sort of clarification that you're now advocating - an thought-provoking example of an inherently gray decision about when a link is likely to be helpful and when it is not. Rossami (talk) 02:22, 27 August 2008 (UTC)

Tony, unfortunately this is one of the few times I have to disagree with you. I think Rossami makes a good point here. Matthewedwards (talk • contribs • email) 03:30, 27 August 2008 (UTC)

I agree with Rossami, except on the suggestion that we cannot have orphaned articles. We link to three from triangle because the former article referrs to the latter subject, so it covers that subject in a different context, providing some value add. But if we changed the linking policy so that we didn't link to articles like "three" for whatever reason, making it an orphan, that does not mean we should not have an article if it stands on its own. Dhaluza (talk) 03:46, 27 August 2008 (UTC)

I don't disagree with the general ideas behind the list given here, but I do not think putting a list in this guideline is a good idea. It will just fire up misguided crusaders who will go off obliterating every link to items on the list without much thought. The general principle that the link should add value is enough. If the vast majority of readers are going to have sufficient familiarity with the linked word, then there is no need for a link. But if there is additional insight on the topic at hand to be gained from the linked text, then the link is worthwile, even if most users will not opt to follow it. I think this list can be helpful in indentifying the general principles that can be used to make this giudeline more descriptive, but I would object to making it more prescriptive by including a list like this. Dhaluza (talk) 02:19, 27 August 2008 (UTC)

Don't link to anything you learned about in elementary school

We could sum up this list by simply saying: "Don't link to anything you learned about in elementary school," and that would about cover it. But then, on further consideration, we do want to write to the widest audience possible. So what about readers who are in elementary school, and have not learned about all these topics yet? I'm not saying we should link to everything to bring it down to a first grade level, but where should we draw the line?

The other issue comes with geography, which is a challenge to many. You might think it is not necessary to link to China, but 30% of college age Americans can't find it on a map, and most have misconceptions about it. So, again, where to draw the line?

Finally, we need to keep in mind that many readers will not be native English speakers. The en.wikipedia contains many more articles than the other language versions, so many readers will be coming here for info not avialable in their native language. We cannot make assumptions about their education or background either. So, to someone in the third world, Industrial Revolution may be a "foreign" concept.

-- Dhaluza (talk) 02:54, 27 August 2008 (UTC)

If someone can't find China on a map, they can so easily type c h i n a into the search box—four quick keystrokes more than clicking on a link that no one else wants and that dilutes high-value links. And the article on China doesn't immediately locate it in terms that such a person would understand, anyway; have a look. Mainland China at least provides a localised map towards the top, but I wonder whether someone who doesn't know where China is wouldn't need to see it on a much larger world map.

Stooping quite so low just to save such people a few keystrokes, and to save non-native speakers looking up a dictionary on their computer or by their side—is not a reasonable balance, given that every link is a little more dilution and colour-clutter for everyone. You have to draw the line somewhere; otherwise, why not link just about every word as the all-the-web lunies urge? Then the black words could stand out: great. WP is not a kindergarten or a learn-English facility; it's a serious repository of knowledge for the everyday anglophone. Nor is it here to try to make up for the xenophobia of a particular education system at the expense of the general reader. Tony (talk) 04:12, 27 August 2008 (UTC)

It sounds like you are suggesting that the length of the link should be a factor, which is a new concept. Does this mean that we might not link to "China" but that we might link to "United States" because it needs more typing? Or are you saying that we don't need links at all, because you can type any term in the search box (as long as it's not a piped link)? Linking every word is an extreme that has been rejected. But color clutter should not be the issue either. The issue is that the reader should not be disappointed when following a link to a blind alley in the context of the article subject. A link is only a problem if it is of no help to virtually everyone. WP is not for learning English per se, but it is for learning about things in English, even if that is not your native language. So I strongly disagree that we should only be concerned with anglophones. Dhaluza (talk) 04:38, 27 August 2008 (UTC)

PS And how on earth will a diversion to the article on three increase a reader's understanding of "triangle". It opens with this:

3 (three) is a number, numeral, and glyph. It is the natural number following 2 and preceding 4."

Hmmm, that's useful. But wait, there's more; read on:

Three is the first odd prime number, and the second smallest prime. It is both the first Fermat prime (2²ⁿ + 1) and the first Mersenne prime (2² − 1), as well as the first lucky prime. However, it's the second Sophie Germain prime, the second Mersenne prime exponent, the second factorial prime (2! + 1), the second Lucas prime, the second Stern prime.

Three is the first unique prime due to the properties of its reciprocal.

Three is the aliquot sum of one number, the square number 4 and is the base of the 3-aliquot tree.

Three is the third Heegner number.

Three is the second triangular number and it is the only prime triangular number. Three is the only prime which is one less than a perfect square. Any other number which is n² − 1 for some integer n is not prime, since it is (n − 1)(n + 1). This is true for 3 as well, but in its case one of the factors is 1.

That's really helpful to someone who's reading the article on "triangle". If anything in Three is relevant to the understanding of a triangle, it should be included directly in the article on "triangle", rather than remain an isolated bit of information for the poor reader to try to hunt down in such an general, unfocused article as "Three"; by the way, the links in the opening of "three" do appear to be high-value, by contrast. It is rare to find such a high density of valuable links.

Let's get rid of this addiction to diversionary browsing that WP has got itself into. It ends up damaging our product through dilution and reducing focus. Anyone can divert when and where they wish by using the search box. It is not up to us to provide a magic carpet to anywhere at the general expense of the appearance and readability of the text, and the highlighting of genuinely high-value links. Tony (talk) 04:25, 27 August 2008 (UTC)

It may be relevant to duplicate info in two articles, or only keep it in one or the other. But this is likely to be handled on a case-by-case basis, and not always consistently or constantly. It is not necessarily diversionary browsing (approaching clicking on the "random article" link for example). The concept of three and triangle are obviously directlty related, and each discusses the other in a different context. So it is not unreasonable for the two articles to link to each other in some way. Even if they do not have embedded links in the text, a link from the "see also" section would be in order. Dhaluza (talk) 04:45, 27 August 2008 (UTC)

No, if someone doesn't know where the United States is, or what it is, they will need to type the 13 characters into the search box, sorry, rather than force the term to be bright-blue on its hundreds of thousands of occurrences in WP. Typing into a search box shouldn't be hard for anyone who uses a computer. You haven't answered my query that China doesn't help someone to locate it on a world map, nor how Three helps a reader to understand the topic of "triangle", rather than confuses them. I do not find your case about this at all convincing. Nor do I expect the Chinese WP to link lots of common words in Mandarin just because I might be learning the language. Tony (talk) 04:53, 27 August 2008 (UTC)

If link color is truly a problem, then this should be correctable by user preference settings, so that should not be the issue. Neither should length of link, whether 5, 13, or whatever number of characters. The only issue of substance is diverting readers to articles that "are not relevant to the context" as the page name suggests. As for China, if that article does not at present answer a question, then that is a deficency in the article that should be corrected over time. We would still link to the article that we would expect to answer the question, just as we red-link to articles that should exist to describe a related topic. Dhaluza (talk) 05:08, 27 August 2008 (UTC)

There are more casual readers (one time IPs) who visit WP than there are registered users. They can't change their preferences. That's why all my settings are default. I can see what they see. Matthewedwards (talk • contribs • email) 05:41, 27 August 2008 (UTC)

I agree, my point was that if users are sensitive to link color because of visual acuity, or personal preference, that should be handled by those users. We should not discard links that would help some people, because they might offend others who could take steps to mitigate the issue on their own. For example, with technically demanding material we tend to have a greater link density. We should not artificially reduce the linking in this case because of an amorphous, subjective standard based on beauty in the eye of the beholder. We should make links whenever a reasonable argument can be made that they "are relevant to the context," regardless of how it might look. We can allow that the visual appearance is in tension with WP:BTW, so links should be justifiable on some level, but we should not raise the bar just to make articles look pretty. Dhaluza (talk) 10:57, 27 August 2008 (UTC)

There may be some validity to an argument that WP:BTW is best served by avoiding overlinking so reverse searches using the "what links here" function remains useful at some level. So in this case, we might avoid linking to United States not because "everybody knows it" or "anyone can type it" but because someone doing a reverse search will be deluged with hits, just like someone doing a web search for information about an obscure meaning of a word that also has a common meaning. So we may avoid linking to "United States" in most contexts that only refer to it in passing and dont contain any information relative to that topic. But we would link to it in each of the articles about the 50 states, because they are directlty related, and someone browsing in a geopolitical context would be served by this linking. This is why we need to be careful about overly general guidance on specific topics, without providing more nuance. Dhaluza (talk) 11:07, 27 August 2008 (UTC)

←I very much like Dhaluza's idea that "Three" belongs in the "See also" section at the bottom. This is worth building into the guideline in generic terms. Tony (talk) 11:25, 27 August 2008 (UTC)

My point was that the cross-links at least belong in the "see also", since the concepts are related. Whether to link it in the body is an editorial decision based on context, not some generic rule. Dhaluza (talk) 11:38, 27 August 2008 (UTC)

---

This thread kind of veered off from the discussion I was trying to start in this sub-thread. What the main thread seemed to be suggesting is that we should assume common sense. While that makes some sense on some level, we need to be careful how far to take it. Obviously on the English language Wikipedia, we need to assume the reader has some level of familiarity with the English language, otherwise we would need to link every term. Since we have consensus that that is not what we want to do, we have decided not to link common words. The question becomes where to draw the line on what is common. I strongly disagree with the idea that we should assume readers are native speakers or fluent in English. The definition of fluency, is, well, fluid. You could argue that most anglophones are not really fluent in English because they could not read or write a paper that would be accepted at the college level (since most people don't go to college). I think most editors would agree that assuming only a first-grade reading level would be absurd, just as assuming that readers have a college degree is absurd for the majority of general topics (it might be appropriate, or even necessary, for very deep technical discussions though). Where to draw the line between these two extremes is the issue.

I think we should go back to basics, and stick to the fram of "only make links that are relevant to the context." So in context, an article on a simple topic might make links to topics that would not be linked in a more complex topic. One of the goals of WP is to make the material accessible to as many readers as practicable. Linking terms is a way to do this, and it should not be unnecessarily discouraged. For example, when using a technical term, you generally need to assume the reader will not be familiar with it. How to handle this has to be decided on a case-by-case basis. If you are going to use it multiple times, you will want to define it on the first occurrance. But it is not necessary to fully define it--a link will allow readers to get a more in-depth definition and context on the term, while readers already familiar with it can skip the link. Terms that are only used once might just be put in context with a partial definition or even a suggestive adjective. It may even be appropriate to assume the term is familiar to most readers interested in the subject, and only link it for the few who might not.

I think the point is that that we not lose sight of the context by creating overly general lists that will be taken out of context. And we need to be careful of making bright line distinctions like assuming a particular grade level education that will not be appropriate in all contexts. I think compiling lists of overlinked terms is useful to foster discussion about why the terms are overlinked, and what nuance we should add to this guideline to reduce it. But I don't think we should take the shortcut of including the lists in the guideline as guidance in an of themselves. Dhaluza (talk) 11:36, 27 August 2008 (UTC)

Where have we got to in the discussion?

There is clear consensus that there is overlinking and that it should be discouraged somehow, and no very strong disagreement with the outline list I suggested of things that should not normally be linked, but there is disagreement over the wisdom of actually specifying that list in a guideline, and there is disagreement over whether an overlinking guideline can be written that could apply to all contexts, covering articles on simple subjects and on complex technical topics. So I have two proposals; first, to change the guidelne to strengthen it but to remain general (i.e. no lists), and to allow individual judgment and a measure of adaptation depending on the nature of the article itself:

"You should not normally link: the names of geographical locations (continents, oceans, countries, cities), and related nationalities, that are likely to be well-known to readers of the article; major astronomical objects; large-scale historical events such as the two world wars; common professions; names of languages; major religions; and other items that are likely to be familiar to most readers of the article."

and second, a proposed alternative example of when to link and when not to link:

YES - Finnish and Estonian are closely related languages of the Finno-Ugric group.

NO - Finnish became an official language of the European Union in 1995 and its close relation, Estonian, in 2004. Colonies Chris (talk) 12:15, 27 August 2008 (UTC)

I think I prefer the first option, Chris. Examples are so reliant on larger context. "Dictionary"-type words, however they can be referred to, need to feature in the list. Tony (talk) 15:20, 27 August 2008 (UTC)

I prefer the more balanced approach. Although overlinking may be a greater problem now, making a list of words not to link will just encourage people to remove all links to that term, regardless of context. I can't think of any term that should not be linked in some context. As pointed out above, that would create an orphan by definition, which is not desirable. So using the example of days of the week from above, we could say:

• Do link Friday from the directly related articles Thursday, Friday the 13th, and days of the week.
• Don't link to Friday as in "Friday, January 1, 1999" from the unrelated article Bowl Championship Series.

-- Dhaluza (talk) 22:42, 27 August 2008 (UTC)

←That is why the words "normally" and "such as" are already used in the relevant wording, and presumably will be retained. I see no invitation to go on a linking spree of any set of items that is not covered; the style guides already make it clear that scattergun linking is unacceptable. No one has ever used that argument, at FAC or elsewhere, in my experience. Tony (talk) 04:10, 28 August 2008 (UTC)

I've been a professional Web master and "knowledge engineer" for large sites with millions of visitors per year. Something to factor into the mix is the very low "click through" to other articles using links. Getting 10% of viewers to "click through" is an accomplishment. (At least for technical articles. The division of Wiki subjects into many small articles, and reader disposition to browse mitigate that for Wiki. However, the basic point stands.) Increasing the number of links, perhaps counterintuitively, does not greatly add to the amount of "click through", and after a dozen or so links, tends not to add any traffic at all. This is true even where an article's author makes great effort to direct readers to a link. (E.g, with bolding, warnings, caution symbols, etc.) It's preferable to have a few links in an article to subjects that are crucial to the topic, whose importance the readers may not be likely to guess on their own. Alpha Ralpha Boulevard (talk) 02:12, 4 September 2008 (UTC)

A very useful comment, Alpha Ralpha Boulevard. Thank you. You have expressed what I think but I have never quite been able to put it as clearly. Perhaps something like that should be noted in the guideline. Lightmouse (talk) 09:16, 4 September 2008 (UTC)

Indeed; I intend to insert a sentence to that effect: "It is preferable to have a few links in an article to subjects that are crucial to the topic than many links in the hope that their quantity will encourage the reader to click on an important one." Tony (talk) 09:52, 4 September 2008 (UTC)

An alternate view

I think that what is getting lost in this discussion is that there is really no way to know how many readers are going to be familiar with various terms in various articles. Obviously overlinking is a problem, but I do not think that it is such an overwhelming problem that we need to draconianly restrict the number of links to just "a few ... that are crucial to the topic". If you look at today's featured article, Emmy Noether, you'll find that it is quite liberally sprinkled with links, but not to the point of being aesthetically overwhelming. Maybe I'm misinterpreting the words of those above, but it seems like this article, which has been voted as one of Wikipedia's best articles, would have too many links for some editors here. For myself, I find it aesthetically displeasing to see a whole paragraph or section with no wikilinks at all, because it seems to go against the whole point of what having a wiki-encyclopedia is all about. Now, obviously, as the current guideline makes clear, there are good reasons why you will get some paragraphs like that, and I'm not arguing against that. But I think we are losing sight of the fact that we are trying to make this encyclopedia accessible and useful to the widest possible audience. The point of wikilinks is NOT to increase "click-through", it simply makes it that much easier to explore the endless knowledge-base that Wikipedia has become. Yes, I can type the word into the search box, but having the wikilink there makes it a lot easier for me to click through and browse something I don't know about. And as long as having that wikilink doesn't cross the line into turning the article into a sea of blue, then it should be left. I should emphasize again that I am not encouraging indiscriminate linking, just a more common-sense, AGF approach to the issue. I think the current guideline does a perfectly good job of discouraging overlinking, and I don't see a need to make it even stricter. I think we are all turned off by seeing a sea of blue in an article, but I am equally turned off by dead-end articles. There is a good reason we have WP:Build the web.--Aervanath lives in the Orphanage 16:13, 4 September 2008 (UTC)

The TFA opening is a rare example in which the density of linking is high, but just about every link is potentially useful to most readers; this can occur at the start of an article that is in a highly technical subject. The same is true of three, as discussed above. These articles are not good examples to use to justify high-density linking in general. And I have to disagree that there's "good reason" for that BTW page. It's a bizarre vestige of the scattergun linking we used to have. Tony (talk) 16:40, 4 September 2008 (UTC)

Actually, I wasn't referring to the intro of the featured article, but the body, because I know that intros are always going to be link-heavier than the rest of the article. I would also like to point out that you said the "scattergun linking we used to have". If the scattergun linking isn't such a big problem anymore, why are we looking for ways to restrict it even further?--Aervanath lives in the Orphanage 17:25, 4 September 2008 (UTC)

I believe dates of birth and death should be linked when they refer to the subject of the article. Deb (talk) 17:43, 4 September 2008 (UTC)

Can you please explain the benefit of such linking, possibly with an example or two? Tony (talk) 15:56, 5 September 2008 (UTC)

Actually, I don't see the point of that either. I think that IS an example of overlinking. But I think there are plenty of examples in the discussion above that it wouldn't hurt to keep in an article, as long as the article isn't overlinked.--Aervanath lives in the Orphanage 16:21, 5 September 2008 (UTC)

I think my main point is this: Links which may be less relevant to the context should be removed if there are already a sufficiency of links in the article. However, if there aren't that many links in the article, then having links of lesser relevance detracts nothing, and contributes to strengthening the interlocking web of linked articles. Cheers, --Aervanath lives in the Orphanage 16:21, 5 September 2008 (UTC)

recent insertion: linked units in infoboxes

User:Docu has just inserted a guideline that units in infoboxes may be linked and that such links should not be removed by bots. Can he please present a case for this change here? Why are units in infoboxes different from those in the main text? Tony (talk) 10:17, 30 August 2008 (UTC)

There was a lengthy discussion on this on the linked talk page. Please comment there. -- User:Docu

The wp:overlink script

There's a script gathering popularity that is being run through articles to remove a formulaic idea as to what constitutes overlinked terms. Fine, but it's a bit daft I'm afraid. For example, it removes links to England. Why? Someone just ran this through the United Kingdom article and thinks that England shouldn't be linked in that page! I suggest a rethink here - it's not wise to have this gain more popularity without understanding when and where it is appropriate. --Jza84 |  Talk  14:39, 5 September 2008 (UTC)

Why should it be linked? Shouldn't well known geographical locations unlikely to be confused with other locations generally not be linked? I don't think there is misunderstanding of the appropriateness on my part. --Elliskev 14:44, 5 September 2008 (UTC)

Jza, please explain how diverting to the article on England will significantly add to the reader's understanding of the topic at hand. I suggest that information on England that bears directly on the UK be included in the UK article (it probably already is). As a matter of interest, is there not a more focused section that could be linked to, rather than the entire, summary-style article on England? Tony (talk) 15:55, 5 September 2008 (UTC)

Are you actually arguing that England shouldn't be linked to from United Kingdom? How is that not "relevant to the context"? I think we're getting away from the point of this guideline. Should we change the name of this guideline to WP:Link as little as possible? The problem with making a blanket rule like "well-known geographic locations shouldn't be linked" is that there ARE cases where they should be linked.--Aervanath lives in the Orphanage 03:13, 6 September 2008 (UTC)

I think you're right, in this case. Tony (talk) 04:27, 6 September 2008 (UTC)

I don't disagree that there is a legitimate argument to keep England linked. I believe there should be a way to get from the UK to England. I figured that there was no need for the inline link, since there is a link to every political subdivision in the British Isles template at the bottom of the article. That article is very, very blue. --Elliskev 17:06, 8 September 2008 (UTC)

Yes, the article is very blue, as it's essentially just one long lead section, summarizing many, many, many sub-articles. It is also a very long article. Thus, linking only one instance of England in the whole article (and that one at the very end) is clearly insufficient. To quote from the current guideline, under the heading What generally should be linked: "A link that had last appeared much earlier in the article may be repeated, but generally not in the the same section." In such a long article as this, it makes perfect sense to have multiple links to the same article, as long as they are spaced far enough apart from each other throughout the article. Cheers,--Aervanath lives in the Orphanage 17:53, 8 September 2008 (UTC)

Birth dates?

A fellow editor complained on my talk page that the birth dates should be linked, as their corresponding articles show what was going on when the subject was born/died, which helps to create a deeper understanding of the subject through understanding the time period they were alive in. This makes sense. On the other side, I see some editors are massively removing these links and the birth dates are unlinked at Wikipedia:MOSNUM#Dates of birth and death (following this example I removed them also at MOS:BIO but was partially reverted). Any comment? --Eleassar ^{my talk} 18:08, 8 September 2008 (UTC)

Birth year and date doesn't really help to show anything usefull - links to decades during formative years, education, etc. would be better (not that I'm suggesting that be done). Also, having only one date formatted according to a user's preferences, but all the others in a different, e.g. middle-endian format, would be jarring. -- Jeandré, 2008-09-08t20:29z

Thanks for the clarification. --Eleassar ^{my talk} 14:30, 9 September 2008 (UTC)

I would just like to note that worrying about the date preferences is a very minor matter; the vast majority of our readers don't have an account, and therefore don't have access to that option. In my opinion, we shouldn't even bother having date formatting options at all: editors should see exactly what the readers will see. We're not just making this for ourselves, we're making it for the whole world. (Maybe I'm just too idealistic, though.) Regards,--Aervanath lives in the Orphanage 16:38, 10 September 2008 (UTC)

I agree that birth (and death) dates should stay linked, especially within the lead sentence. They deliver as much context in those instances as any linked date possibly could, in my opinion. After all, something has to link to all of those year and day articles, so it might as well be that. --Bongwarrior (talk) 09:02, 11 September 2008 (UTC)

They are linked under the previous autoformatting system, not to provide "useful" links that "significantly increase readers' understanding" of the topic at hand. Are you thinking of the day-month link or the year-link as being the useful one? Please provide and example of how either could be useful, whether at the top as birth and death dates or anywhere else in an article. Tony (talk) 13:10, 11 September 2008 (UTC)

Eleassar already did so above. Linked dates can provide a greater understanding of the time period of the subject at hand, if the reader so desires, by showing what else was happening in the world at the time. The year is probably the more useful one in this case, I guess. I don't disagree that most dates should be unlinked, but birth/death dates should stay as they are. --Bongwarrior (talk) 14:59, 11 September 2008 (UTC)

Have you actually looked at the content on those "year" pages? First, there's almost nothing useful about them nor can you find anything useful by following "what links here" because the years themselves are so massively overlinked. Second, you can't learn anything useful about my developmental influences from my birthyear because I wasn't old enough to be influenced by anything encyclopedic yet. Influences in the first year are entirely domestic. In fact, child-development specialists will tell you that influences in the first 5-8 years are almost entirely domestic - or at best, highly localized. Certainly not the kinds of influences that will get coverage in an encyclopedia. If you really want to easily link readers to the influential periods in a person's life, you need to find a way to link to the appropriate decade article covering the ages somewhere between 10 and 30.
Likewise, a link to a death year tells almost nothing about the person's life except in the rare case where the death itself was a cause for notability. I've never yet followed one of those links and learned anything useful. Rossami (talk) 23:43, 19 September 2008 (UTC)

It is a common feature of biographical writing to mention in the same breath as mentioning the birth to the person they are writing about to mention what was happening in that year. This kind of context is useful to many people. It's not just a Wikipedia thing.Dejvid (talk) 09:18, 24 September 2008 (UTC)

What? Run that past us again, please; it seems to make no sense. Tony (talk) 10:37, 24 September 2008 (UTC)

I think he's saying that it is a common feature of biographical writing to talk about the events surrounding the subject's birth. Which is true...but not when writing biographical articles for encyclopedias.--Aervanath lives in the Orphanage 13:11, 25 September 2008 (UTC)

I would like to link birth dates (especially years), but think it should be done in the Infobox? (where the template can control it). Would that serve to limit its usage?Mjquin_id (talk) 18:23, 8 November 2008 (UTC)

Geographic locations

I have reverted the changes made to the "What should commonly be linked" and "What should not commonly be linked" sections in early July, as I can find no consensus for them in discussions here. From what I can tell, the wording was changed at the MoS links guideline, then changed here 21 hours later ("Pasting in bit from MOS (links)"). It appears to represent a reversal of the intent of that part of the guideline, without sufficient discussion or consensus at either page. (Further to this, I have noticed some objection to the idea of unlinking geographic names during the DA script runs.) Regardless of whether or not the guideline is to be changed, I don't think it should have been rewritten prematurely. Thoughts? --Ckatz^chat_spy 08:58, 11 September 2008 (UTC)

I'm reinstating the change for two reasons:

1. The issue was discussed here and the the change has been in place for two months now without complaint
2. The paragraph as it now stands doesn't even hang together properly - it mixes up mention of technical terms with mention of geographic names. Colonies Chris (talk) 12:34, 11 September 2008 (UTC)

I endorse this reversion by Colonies Chris. Ckatz, you are on record as having a deep commitment to maximising the density of linking, without regard to the original purpose of wikilinking. Having gone through a long period in which there was little or no discipline in the selection of items to link, WP is now on a trajectory towards what might be called "skilled" or "smart" linking. The key objectives are to minimise the dilution of high-value links and to avoid catering for diversionary browsing at the expense of a sea of blue. Tony (talk) 13:15, 11 September 2008 (UTC)

(To Tony1) Tony1, I'd really appreciate it if you would stick to the topic at hand, and avoid making false claims about my motivations. At no time have I ever gone on "record" as "having a deep commitment to maximising the density of linking, without regard to the original purpose of wikilinking." That is utter rubbish, completely untrue, and (more to the point) serves only to distract others from looking at how the initial changes to the guideline were instituted back in July. --Ckatz^chat_spy 17:11, 11 September 2008 (UTC)

(To Colonies Chris) Chris, I've restored the original text as it is important to discuss this reversal of the guideline first. Tony1 instituted the change, now I'm reverting it so that we can discuss. From what I can see in the talk archives, there was no prior consensus to institute the change. As well, there were several objections raised after the change, as editors started to become aware of what had happened. (It often takes time for the impact of such changes to be recognized by others, as with the recent announcement in the Signpost.) Given that the change is being used by the person who wrote it to justify a mass removal of links, we need to sort this out first. --Ckatz^chat_spy 16:54, 11 September 2008 (UTC)

Looking at the difference between the new version and old version makes it clear how big a change that was. I think that the old version was a little too strongly in favor of linking geographic place names, since I'm not sure they always need to be linked. I'm not even sure that the guideline needs to mention them specifically, since I don't really see that place names are any different than any other word. We should strike the same balance with place names that we would with any other link: the first instance in the article should be linked, and then not after that, if the article is not too long; or, if the article is too long, then it's ok to link it again much farther down.--Aervanath lives in the Orphanage 13:33, 12 September 2008 (UTC)

However, the general principle of linking the first occurrence of a term only applies when that term is likely to be unfamiliar to a significant number of readers. In a sentence like "On completing their North American tour, the band set off on a three-month tour of Europe", the links are valueless. And the same applies to many other common words - for example in a typical article opening sentence like "Jennifer Lopez is an American singer and film actress." all four links are valueless. I don't think it's wise to have an explicit list of "things not to link to" in the MoS, but I do think we need to offer some guidance in general terms to discourage valueless overlinking. The current wording actively encourages valueless linking. I proposed earlier a phrasing like this:

You should not normally link: the names of geographical locations (continents, oceans, countries, cities), and related nationalities, that are likely to be well-known to readers of the article; major astronomical objects; large-scale historical events such as the two world wars; common professions; names of languages; major religions; and other items that are likely to be familiar to most readers of the article.

but tony1's wording, although not perfectly to my liking, is much better than the status quo. Colonies Chris (talk) 15:43, 12 September 2008 (UTC)

I'm fine with that wording, Chris. I agree that there were problems in specifying names, and the UK instance above is an unfortunate example. Linking several times to "United States" in every US-related article is clearly overlinking, as it is for almost all instances of well-known country names. Ckatz, you say nothing above to convince people that you don't have such a commitment, and your past statements have strongly indicated this. It is very much the topic at hand, for you're acting again in a way that is not entirely NPOV in your assertion that the text has to be reverted to the version you prefer "so we can discuss". I don't buy that kind of tactic. Unless there are reasonable objections, I believe that Chris's proposed wording should be implemented soon. Tony (talk) 16:00, 12 September 2008 (UTC)

Tony, if you have an issue with me, take it up on my talk page rather than here. Otherwise, please stop making groundless accusations which you have not (and cannot) justify. --Ckatz^chat_spy 18:40, 12 September 2008 (UTC)

I would cut Chris's suggestion down a little bit, and re-arrange it, to put the central point right up front:

Items likely to be familiar to most readers of the article should only be linked when directly related to the topic. For example, the names of major: geographic features and locations, astronomical objects, historical events, religions, languages, and common professions are unlikely to be useful.

What do you guys think?--Aervanath lives in the Orphanage 17:39, 12 September 2008 (UTC)

Simpler is better, although we should be careful not to presume what international readers may or may not be familiar with. How about:

"Use discretion when linking items that may be familiar to most readers of the article, such as the names of major geographic features and locations, astronomical objects, historical events, religions, languages, and common professions."

I'd avoid the phrase "unlikely to be useful" as it is very much a value judgement that will differ from person to person. (It is also quite easy for an editor in one region to presume that a global audience will be familiar with his or her nearby feature or location.) Beyond this, we need to make sure that we do not underlink in a mad rush to avoid what some call a "sea of blue". For example, earlier today the article on North America was processed by the date/link script; Asia, Africa, and the other continents were delinked in the lead paragraph. I would think that those were very appropriate links in that context. --Ckatz^chat_spy 18:40, 12 September 2008 (UTC)

I see a number of problems with such a caveat:

1. We can't properly give a list of what our readers may be familiar with. We don't know this.
2. Linking is determined by relevance, not familiarity. To resurrect an earlier example, a link to ear is worthless in many contexts, but it may be relevant and useful in an article dealing with human anatomy.
3. To echo User:Ckatz's concern, we don't want to encourage robo-delinking. Judgment is necessary to determine when almost any term should or should not be linked.

I think it would be useful though to weaken the last entry in What generally should be linked to clarify that place names should not reflexively be bluelinked.

Spacepotato (talk) 00:08, 13 September 2008 (UTC)

• This is the slippery slope to the sea of useless blue: "should only be linked when directly related to the topic." Anyone can argue that the United States, and United States|American are directly related to any article with cultural ties to the country. Yet why should it ever be linked for this reason? If a reader has never heard of the US, let them type two characters into the search box (as opposed to one click on a bright-blue splotch in our main text or infobox)—look, like this ... US. In what way is this consistent with the clear directive in the Manual of Style?

It is not useful and can be very distracting to mark all possible words as hyperlinks. Links should add to the user's experience; they should not detract from it by making the article harder to read. A high density of links can draw attention away from the high-value links that you would like your readers to follow up. Redundant links clutter the page and make future maintenance harder.

And now our resident editor who wants just about everything linked, Ckatz, has removed geographical names from MOSLINK, just because he doesn't like it. No discussion of what should replace it, as here. Just launch in and remove. It's becoming an aggressive campaign. Tony (talk) 04:02, 13 September 2008 (UTC)

PS And "Use discretion" replaces a fairly strong guideline with one that is as weak as water. It gives a carte blanche to anyone who wants to litter our articles with useless blue links that no one clicks on, and that degrade the look and readability of the text, and worst of all, dilute our high-value links. This is a huge step backwards, and I don't understand the fervour with which this call to blue-spatter everything is being conducted. Tony (talk) 04:06, 13 September 2008 (UTC)

• The article on North America. Possibly the names of the other continents are justifiably linked here, but why have you put back Italy and retained Earth. Why is "the Americas" linked multiple times? Way down in the article, "continent" is linked: heck, if they don't understand what that word means by then, what's the point in the first place? "Central America" is linked twice in three paragraphs. "South America", "Western Hemisphere" and many others are blue-littered multiple times through the text. Why? It looks appalling. Then, when a relatively high-value link comes along, such as "North American Plate", it's awash. Why do we need the English language to be linked several times? This is the English WP. WHO doesn't know what English is, and why is the article on it of any relevance to this article? Tony (talk) 04:22, 13 September 2008 (UTC)

Tony, for the third time I would ask that you desist in your efforts to misrepresent my actions, while completely disregarding your own role in this matter. Since you won't back up your claims about me, can we examine what you have done with these guidelines? Back in early July, having already established your personal dislike for the "sea of blue", you added the text about geographical names to MOSLINK. There was no discussion, just the edit summary "generally no trivial linking of well-known geographical names". Shortly thereafter (roughly 21 hours, to be precise), you rewrote WP:CONTEXT to reflect what you had added to MOSLINK, again without consensus, with the edit summary "Pasting in bit from MOS (links); we should consider merging this one into it". Subsequently, you have been using that same text - authored by you - to justify your script-based delinking of these "common terms".

As for your assertion above that I "removed geographical names from MOSLINK, just because [I don't] like it." Quite the contrary, in fact; I actually commented out your undiscussed addition, with the explanation "commenting out geographical; I see no discussion about it anywhere". I then proceeded to explain my edit on the talk page. How, exactly, does that translate into an "aggressive campaign"? --Ckatz^chat_spy 05:54, 13 September 2008 (UTC)

I like Aervanath's wording - it condenses down the key points into a tighter form than my original proposal. Ideally I'd like to find a slight modification, to suggest that all language links are unlikely to be useful, except in language-specifc articles, not just those to major languages. For example in a sentence like "The novels of Agatha Christie jhave been translated into (list of languages)", there's no benefit to linking any of them. Or in a sentence like "The Beatles recorded several of their early hits in German", a link to "German language" wouldn't be helpful, not because German is a fairly well-known language, but because what the reader wants to know is "why?", not about the characteristics of the German language, and that would be the case even if the language had been a much lesser known one.

And I agree with the earlier comment that CKatz's suggestion is just much too weak - it's hardly guidance at all and what we need here is a strong steer. Colonies Chris (talk) 08:27, 13 September 2008 (UTC)

I have been watching this debate for while without commenting. It is clear that geographical terms are vastly overlinked and some editors simply link them because they are there. It is reasonable that anyone with sufficient grasp of english to read Wikipedia should know common geographical terms, just as they will know a few animal terms. There is a parallel with 'common units of measurement' and it would be simple just to add a similar bullet

• In general, do not create links to the following.
  • Plain English words, including common units of measurement (particularly if a conversion is provided). [2]
  • Common geographical terms

As with units of measurement, a footnote can list placenames that are 'included but not limited to'.
Lightmouse (talk) 18:02, 13 September 2008 (UTC)

• I love the wording from both Colonies Chris (15:43, 12 September 2008 post) and that of Aervanath (17:39, 12 September 2008 post). The differences between them are subtle, and I encourage the two to collaborate to produce a guideline that is a fusion of the two. As an engineer, I would likely have ended up with more stilted wording (and more of it) to accomplish the exact same thing.

  There should be no simple rule as to what should or should not be linked; it is entirely subject-driven and the above two suggestions capture this concept just fine. There shouldn’t even be a blanket rule that United States should or shouldn’t be linked; a simple article for beginning geography might need such a link. Every experienced editor to Wikipedia should be expected to understand who their intended audience is. If it’s an introductory article, say… Atom, or Introduction to quantum mechanics, links to *familiar* things might be quite appropriate. But a properly written article on an advanced topic, like Planck's law, should not link *common* (for that audience) topics, like “frequency”.

  The whole point of links is to understand who the intended audience is and who you are writing for, and then anticipate what topics such readers might like to further explore. When properly done, links invite exploration and learning. A well-linked article should often elicit a reaction of “oh wow; I didn’t know there’d be a Wikipedia article on that nuance!” When we bombard the reader with too many blue links—or Easter egg hunt links to something unanticipated (and probably disappointing)—we just desensitize the reader to the links and turn our articles into a giant blue turd.

  Greg L (talk) 19:43, 15 September 2008 (UTC)

Perhaps those exceptions are covered by the wording "are generally/normally not linked". The linking of the names of anglophone countries is so endemic in WP, and robs valuable limelight from high-value links (with rare exceptions), that a strongly worded passage is required. Again, there's no problem in the opt-out "generally". I myself made an error of judgement (see above) in removing the link to "UK" in an article where it was probably appropriate (a rare exception). Tony (talk) 02:17, 16 September 2008 (UTC)

• Yes, “generally” covers it. And I agree, linking United States is usually unnecessary; anyone knows they can type that subject in the search field and go to the relevant article. Just like linking “frequency” in an article on Planck units, we don’t need links to subjects that are drop-dead obvious for the intended readership. But this is Wikipedia, where there is a huge variation in the experience of the volunteer contributing editors that comprises this community. I absolutely guarantee you there are new editors whose only rule for whether or not to link is this: “is there an article on Wikipedia that I can link to?” What I like about both proposals from Colonies Chris and Aervanath, is they use key wording I think is valuable to focus editors on what their mission is when writing articles: “…that may be familiar to most readers of the article…” and “…likely to be well-known to readers of the article…”. Both focus on the important issue: write appropriately and add links that should be of interest to the intended audience. (hell, I kinda like that wording). Maybe even adding example solutions in guidelines would be valuable in this case. Tony, please post below, the wording you advocate or think shouldn’t have been changed. Greg L (talk) 03:18, 16 September 2008 (UTC)

• Greg, "are of interest to" is way too broad. I'm interested in the United States as a topic, but I don't want it linked all over the place, because it's a waste of valuable link-density. Many articles link the word five or six times, either directly or through a pipe from "American". Same for the other major anglophone countries, and many languages. Linking guidelines have be concerned with the utility of the link to the likely reader (i.e., whether it's sufficiently focused on the topic, whether it's worth linking given the assumptions of general knowledge on the part of most readers, whether travelling there will significantly increase readers' understanding of the topic, and whether the destination of the link contains specific pieces of information that instead should be included in the article at hand.) Tony (talk) 09:05, 16 September 2008 (UTC)

The wording doesn't need to be complicated. It just needs stating. As I said above, a succinct bullet:

• Common geographical terms

below the plain english reference would be as effective as the bullet for common units (which I think has worked well). Lightmouse (talk) 09:34, 16 September 2008 (UTC)

• Tony, well, how about “add germane and topical links that are likely to be of interest to the intended audience”. That, and a few examples ought to cover what you are trying to accomplish. If someone is reading about the Speed of light, and it happens to mention “United States”, it ought not be linked in that article.

Linking geographic names is analogous to technical terms. To people reading the article from some faraway place, the name of that place is probably as unfamiliar as a technical term of art. We provide the link to allow the reader to quickly look up the word to see what it is about--this is important information to readers, and should not be short-changed simply for aesthetics. Note also that there are tools that allow you to review the lead section of an article by hovering on the linked term, which is not possible if you must type the term in a search box, and potentially go to the wrong page, or have to navigate a disambiguation page. Also by linking the term, we build a system of redirects with possibilities, that suggest articles that still need to be created, or help people find what information is currently available (and not available) on WP. I have restored the long-standing provision, with a footnote intended to address concerns with overlinking. Dhaluza (talk) 17:33, 20 September 2008 (UTC)

Arbitrary break

How about this: "It is generally not necessary to link items that would be familiar to most readers of the article, such as the names of major geographic features and locations, astronomical objects, historical events, religions, languages, and common professions." Thoughts? Comments? Concerns? Abuse?--Aervanath lives in the Orphanage 18:14, 16 September 2008 (UTC)

 Done As there seems to be no objection.--Aervanath lives in the Orphanage 20:01, 19 September 2008 (UTC)

Revised, as there is an objection; removing "astronomical objects" as we cannot presume this. Removing "historical events" for the same reason. --Ckatz^chat_spy 09:16, 21 September 2008 (UTC)

Actually, I've revised it to "Items that may be familiar to most readers of the article", and left it simply at that. One, we cannot presume what readers do and do not know. Two, specifying what was listed before has the unwanted side effect of editors instituting mass removals of links, as seen already. --Ckatz^chat_spy 09:21, 21 September 2008 (UTC)

That's fine with me. I prefer your wording, I'm just not sure it reflects consensus yet. My language was written specifically as a compromise to settle the issue by finding a middle ground.--Aervanath lives in the Orphanage 10:57, 21 September 2008 (UTC)

Your efforts to keep the discussion on track are certainly appreciated. -Ckatz^chat_spy 20:22, 21 September 2008 (UTC)

• No, it's not fine at all. Ckatz, just because you want to cover WP in links doesn't give you the imprimatur to rule over this style guide. And your argument that "we cannot presume what readers do and do not know" rules out your substitute text too. If you want to make major changes, you'll have to get agreement on this talk page. Tony (talk) 11:19, 21 September 2008 (UTC)

  Tony, I must insist that you cease your posting of provocative, unjustified lies about what I do and do not wish to do. You have established a pattern of making accusations and then ignoring requests to support them; it has gone beyond incivility to the point of being disruptive. If you can't justify your desires here without dishonestly misrepresenting people who disagree with you, then don't say anything at all. After all, this began when you rewrote the guidelines to suit your preferences (as documented above), without any effort to achieve consensus. The "major changes" originated from you. --Ckatz^chat_spy 20:19, 21 September 2008 (UTC)

1. User:Ckatz's wording strikes me as superior as linking will always depend on context. Editors must use judgement in determining when terms should or should not be linked. (Actually, the criterion is one of relevance, not familiarity, but this is a step in the right direction.)
2. I see no consensus for the removal of this section. Clearly, whether or not there is consensus for delinking well-known geographical names has no relevance to whether or not little-known geographical names should be linked.

Spacepotato (talk) 21:50, 24 September 2008 (UTC)

What reason is there to avoid linking in quotations?

? —Preceding unsigned comment added by 71.182.182.120 (talk) 20:25, 14 September 2008 (UTC)

It's subjective. Mainly, the argument is that linking certain phrases in a quotation places emphasis on the linked term and prescribes it a meaning that may not be the precise semantics assumed by the original author. It's also just a style thing. Dcoetzee 05:20, 16 September 2008 (UTC)

It's not subjective at all. One of WP's pillars is to be true to our sources. Linking items in quotations jeapordises this. Linked pages may add nuances that were not intended by the original text. Linked pages can change over time. Tony (talk) 07:15, 17 September 2008 (UTC)

This is an old argument and a perennial topic of discussion. See here and here for some old discussions (there were others as well). Sometimes there is a need to explain words or phrases used in a quote, but this should be done outside the quote, either in a footnote or in the adjacent text to the quote. On very rare occassions, it may be simpler and less messy to just link (eg. glossary definitions of words in poems), but generally if enough care is taken it is possible to both explain the terminology used in the quote, and leave the quote unlinked. Carcharoth (talk) 07:01, 3 October 2008 (UTC)

Title change suggestion

Considering that the guidelines here go far beyond simple relevance, I think a title like Wikipedia:Linking guidelines or Wikipedia:When to link would be more appropriate. What do you think? Dcoetzee 06:14, 16 September 2008 (UTC)

In what way do you think they go beyond simple relevance? For example?--Aervanath lives in the Orphanage 18:19, 16 September 2008 (UTC)

It discusses, among other things, circular linking, consistent link density, "one link per entry" on disambiguation pages, avoiding links within quotations, and many other things that amount to style guidelines not related to relevance. Dcoetzee 18:35, 16 September 2008 (UTC)

The controlling guideline on those issues is WP:MOSLINK. I've added a hatnote to the Other considerations section to make that clear. This guideline should not go beyond simple relevance. In fact, perhaps that section should be taken out altogether.--Aervanath lives in the Orphanage 06:01, 17 September 2008 (UTC)

Remove section which duplicates WP:MOSLINK

Currently the Other considerations section duplicates WP:MOSLINK. As per Dcoetzee's comments above, this causes this guideline to go well beyond it's original purpose. I've added a hatnote to the section to emphasize that MOSLINK is the controlling guideline, but I'm thinking that the section should just be removed altogether. WP:OVERLINK shouldn't be a style guideline, it should be a much simpler, more straightforward directive to Only make links relevant to the context with some specific examples. If it gets too specific, then it starts to encroach on MOSLINK's "territory", as it were. Thoughts?--Aervanath lives in the Orphanage 06:11, 17 September 2008 (UTC)

Or merge the two pages. Tony (talk) 07:12, 17 September 2008 (UTC)

I think merging is a good idea. Lightmouse (talk) 08:35, 17 September 2008 (UTC)

I wouldn't merge them completely. Move the more specific stuff over to MOSLINK and make OVERLINK a more general guideline along the lines of WP:Build the web. I envision OVERLINK and BUILD as setting forth the fundamental tenets of linking, while MOSLINK would be the practical, detailed guideline.--Aervanath lives in the Orphanage 17:25, 17 September 2008 (UTC)

Sorry, I see BTW as having been written by link extremists who want to scattergun-link everything in sight. I thought you said a little while ago that you felt BTW and CONTEXT were pretty close together, or some such. BTW is not written like a guide. It's more like an essay. As far as I'm concerned it can stay, but it certainly doesn't belong in MOSLINK. This guide does. MOSLINK itself already sings from the same songsheet on the need for a more skilled and selective approach to linking. Tony (talk) 14:40, 19 September 2008 (UTC)

As I said last time we discussed it, the way the guideline is written now does not promote the sort of "scattergun-link" mindset that it may have originally. And yes, BTW is not written in a specific way. I think this was done intentionally (and if not, then I certainly intend it that way), because it's really not supposed to be a specific step-by-step guideline that sets out a lot of specific examples. All BTW is there for is to remind us that we should link, period. It does not state that you should link everything, or to link anything you please. It simple puts out in writing the basic thing that makes Wikipedia, and wikis in general, such a useful tool: "we can link, so do it. How you do it is up to you."

I would like OVERLINK to be as general as BUILD. I don't think they should be completely eclipsed by MOSLINK, rather MOSLINK should be the detailed and specific guideline that implements the spirit of BUILD and OVERLINK, much as the spirit of the very general exhortation "m:Don't be a dick" is implemented by the much more specific policies of WP:CIVIL, WP:BITE, and almost all the the other behavioral guidelines on Wikipedia. I would also refer you to WP:IAR, which is also pretty darn general.

By the way, I would just like to note that while Category:Articles with too many wikilinks has exactly SIX articles in it at present, Category:Dead-end pages has 591, and Category:Orphaned articles has a whopping 33,344 articles in it. So I'm really hard-pressed to understand WHERE all this "scattershot" linking is. It seems to me that underlinking is currently a far bigger issue.--Aervanath lives in the Orphanage 19:56, 19 September 2008 (UTC)

First, I'd never heard of that "overlinked" category; I'm not surprised that it's almost empty. I could shovel thousands of article names in there right now. Give me an example of articles that are underlinked, then? Why do I never encounter these? Second, by is BTW a guideline and not an essay? Tony (talk) 05:38, 20 September 2008 (UTC)

"How you do it is up to you." Well, this is how we got into the scattergun period; people will naturally follow an urge to chunk up their text with links without thinking of the reading experience or utility—I've caught myself doing it. Editors need a fairly strong set of guidelines to stop them doing this, just as they need to apply discipline to their use of non-free content. MOSLINK already does the job of saying "there are links, therefore use them" (to which I'd add "stragetically" or "selectively"). People have to be guided into using wikis well, which is why we have lots of policies and guidelines; the plain fact is that the more we used these internal links, the more they dilute each other. That may not seem obvious to a newbie.

Under Ckatz's formula, the name of his own country, Canada, is not allowed to appear unlinked; it must be bright-blue to beckon the readers to click it in their ignorance of Canada, no matter what the topic is. It's highly unlikely that such scattergun linking will serve its intended purpose: on the contrary, it defeats the value of wikilinking—or if it doesn't defeat it, it greatly weakens it. Tony (talk) 03:32, 20 September 2008 (UTC)

I'm not sure how you came up with "not allowed to appear unlinked"; I'd ask you to justify it, but fully expect you won't (and can't) based on previous requests. That aside, it would help the discussion a lot if you could avoid these broad, overly dramatic generalizations. Not everyone agrees with your disdain for the blue links, but insinuating that we'd like to paint the whole page blue just because we don't agree with your view is unhelpful --Ckatz^chat_spy 04:29, 20 September 2008 (UTC)

Ah, there you are—I just put that in to wake you up. Tony (talk) 05:34, 20 September 2008 (UTC)

I think it's odd that you ask me to point towards some underlinked articles right after I included a link to Category:Dead-end pages, which includes over 500 examples of completely unlinked articles. Here's another one: Wikipedia:Dead-end pages.

I'm not arguing that there weren't people that were in favor of massive overlinking in the beginning, I just think that the problem is much smaller now, since we have CONTEXT and MOSLINK. If you really think that BUILD is totally against Wikipedia consensus, I invite you to try completely removing all links from a few Featured Articles, and see what sort of consensus forms. I think you may be somewhat surprised.--Aervanath lives in the Orphanage 13:10, 20 September 2008 (UTC)

It might be tempting to cast me and others who want to raise the standard of linking on WP as anti-link. Nothing is further from the truth: the whole point of disciplined, more selective linking is to make the wikilinks that are valuable work better. I'm pro-linking, in that respect. Tony (talk) 13:22, 20 September 2008 (UTC)

Exactly my point. You are pro-linking. We all are, or we wouldn't be involved in a wiki. The reason why BUILD is a guideline and not an essay is that HAVING LINKS is an integral part of what we are. And yes, the linking should be disciplined. That's why BUILD is limited by CONTEXT and MOSLINK. Removing the specifics from CONTEXT and moving them to MOSLINK would restore CONTEXT's original purpose, in my opinion, which was not to micro-manage when and when not to link, but to provide a more general exhortation to use your common sense and not link every darn thing you see, but only provide the relevant ones. CONTEXT discourages wanton overlinking, while BUILD discourages wanton removal of links. I see nothing wrong with this arrangement. Actually, I might favor a merger of CONTEXT and BUILD into one page, called, for example, WP:Common-sense linking (or similar), that would, in one page, encompass the dynamic tension that is currently split between them. However, this page, just as BUILD is now (and CONTEXT should be) would be a very general guideline, setting forth the philosophical (wiki-sophical?) foundations of wiki-linking on Wikipedia. MOSLINK could then (as it does now) implement that philosophy, providing the nitty-gritty details of when and when not to link. To put it in different terms, BUILD and CONTEXT should be worded as a general set of values, while MOSLINK specifies the actions we should take based on those values. To draw a comparison to law, BUILD and CONTEXT should be the national constitution, while MOSLINK should be the detailed legal code. (When I sat down to write this, it was only going to be 2 or 3 sentences. Which is odd, considering how much I hated writing papers in high school and college.) Cheers!--Aervanath lives in the Orphanage 18:11, 20 September 2008 (UTC)

Aha, here it is. Sorry, I saw the links to WP:BUILD over at WP:MOS and WP:MOSLINK first and didn't see the discussion, so I reverted. I'll try to grab recent discussions and put them all here. - Dan Dank55 (send/receive) 00:20, 2 October 2008 (UTC)

←As recently discussed at Wikipedia:Village_pump_(policy)/Archive_55#Redundant guidelines, there are 4 different pages giving 4 different sets of advice on what to link: Wikipedia:Build the web, Wikipedia:Only make links that are relevant to the context, Wikipedia:Wikipedia doesn't use Allwiki and Wikipedia:Manual of Style (links). I remember Aervanath was working on this a while ago, and I'm glad to see that some progress has been made: User:Mr.Z-man recently demoted WP:ALLWIKI. But we're not home yet; I see that WP:BUILD still has sentences like this one: "George Washington should be linked to from President of the United States". It is very uncommon for a page on a particular public office to link to all the individuals who have been holders of that office. The style guidelines have ignored WP:BUILD and WP:ALLWIKI for a long time, and it looks like further work and negotiation is going to have to happen if we want to start linking to WP:BUILD. Better yet, it's probably time to get rid of at least one of the 3 remaining pages, and maybe 2 of them. - Dan Dank55 (send/receive) 03:19, 2 October 2008 (UTC)

Clarification. Allwiki has never been accepted practice at Wikipedia. The section now at WP:ALLWIKI was originally in WP:BUILD but only as a sidebar discussion that defined then explicitly rejected the practice. Aervanath recently moved it out to a separate page in an attempt to make the others more readable. Nothing should link to WP:ALLWIKI except the occasional rebuttal of a new user who makes the perennial suggestion that we should change our practices and begin to link everything.
I'll also disagree with your interpretation of the example in BUILD. While you're right that a page on a particular public office should not link to all the holders of that office, it is quite common to link to the first holder of the office. That's not to say that the example can't be improved, though. If you think you can improve it, be bold. Rossami (talk) 17:12, 2 October 2008 (UTC)

Answering Itub's question from WP:MOSLINK on what's wrong with putting George Washington in the President article: nothing, but BUILD is introducing a notion of responsibility; it's saying that it's the job of the editors to build the web, to think about what other articles might be orphaned, and to use links from the article you're writing to fix that. The complete sentence in BUILD is: "Introduce links from related articles to avoid orphaning the article (George Washington should be linked to from President of the United States)." This is not a burden editors have been interested in taking up, nor one we've been mentioning in the style guidelines ... but it's not a bad idea, either; Wikipedia works better without orphaned articles, of course, so I believe there's some room for negotiation. But we can't decide among ourselves and foist it on the writers; whatever it is, they have to agree that it's a good goal and be willing to do it. - Dan Dank55 (send/receive) 13:18, 2 October 2008 (UTC)

Idea for merge

That's fine with me. I think that the issue here is that there are so many different opinions on what the ideal level of linking is and how to enforce it, so any attempt to simplify the status quo is going to get blocked by people who think it is being simplified in the wrong direction. I will restate my views on the ideal outcome here:

• BUILD and CONTEXT should be brief, very simply stated guidelines, possibly combined in one page. They should NOT be detailed, but should be general statements of principle.
• MOSLINK should be very detailed, and should implement the principles set out in BUILD and CONTEXT or whatever page eventually combines the two.--Aervanath lives in the Orphanage 21:19, 2 October 2008 (UTC)

Makes sense to me. I'll undo my reversion of the two places where you linked to BUILD, but I second your motion to combine BUILD and CONTEXT into one page; let's merge them into BUILD. The merge notice will be enough of a clue to editors that we're still working this out. - Dan Dank55 (send/receive) 21:50, 2 October 2008 (UTC)

Reversions are undone, so WP:MOS and WP:MOSLINK have Aervanath's additions of the link to BUILD. As long as those links are there, we need to at least be discussing how to reconcile BUILD and CONTEXT, so I've added merge notices to both. - Dan Dank55 (send/receive) 22:14, 2 October 2008 (UTC)

• Run past me again why we need three pages for linking and not just MOSLINK? Tony (talk) 10:09, 3 October 2008 (UTC)

• • I've asked around, and so far, we don't, which is great ... first time in the history of WP that we have a chance of getting everyone on the same page on this issue. I'll redirect the merge notices to WP:MOSLINK. - Dan Dank55 (send/receive) 15:56, 3 October 2008 (UTC)

Merging seems like a good idea to me. Lightmouse (talk) 18:29, 3 October 2008 (UTC)

• Oh well, I might as well make sure my message point has been pollenated to the key places… Links should always be topical and germane. Properly chosen links anticipate what the readership of any given article would likely be interested in further reading. As such, judiciously selected links invite exploration and learning. In many cases, a reader’s reaction should be “Cool… I didn’t expect they’d have an article on that too!”  When links are judiciously employed in articles, they become more interesting and effective. This isn’t accomplished when articles are over-linked. But…

  If an article has zero articles that link to it, then some measure of effort should be made to get a high-profile link to it. Such links don’t have to be awkward shoe-horns that make body text blue; they can be in the See also section. Whatever works best under the circumstances. If two articles are highly redundant, then merge them. That’s my 2¢. Greg L (talk) 19:16, 3 October 2008 (UTC)

That seems like a nice summary. All I would add is, if the first several links a reader clicks on from your article lead to irrelevant or poorly written pages, they are less likely to click on any future links in your article, and they'll probably expect less from Wikipedia as a whole, too. - Dan Dank55 (send/receive) 19:30, 3 October 2008 (UTC)

• Support merge. And maybe MOS:LINK should be moved to WP:WikiLinking is a science and a form of art, but that's a separate issue. Everyme 23:12, 3 October 2008 (UTC)

• Yes, merge. BUILD and CONTEXT are suffering from bipolar disorder. I understand they exist in tension, but rather than standing alone, they need to be fit together. It's probably best to resolve the conflicts where possible, and allow the conflicting goals to appear together to let readers try to understand the whole story, rather than one half at a time. Dhaluza (talk) 23:32, 3 October 2008 (UTC)

Very succinctly put.--Aervanath lives in the Orphanage 19:07, 4 October 2008 (UTC)

Rough draft of merge

I've whipped up a very rough draft of what a merge of BUILD and CONTEXT might look like at Wikipedia:Build with Context. I've specifically kept it very general, since (as I've said before) I think we should keep the detailed implementation at WP:MOSLINK.--Aervanath lives in the Orphanage 19:44, 4 October 2008 (UTC)

P.S. "Very rough" means please feel free to improve it! Cheers! --Aervanath lives in the Orphanage 19:45, 4 October 2008 (UTC)

Merge to WP:MOSLINK

• Merge. No question. Furthermore, coordinate with WP:LINK which is a redirect to Help:Contents/Links as well, confusingly since WP:OVERLINK and [{WP:MOSLINK]] are in the Wikipedia namespace, and doubly confusing because of the "WP:" preface in the shortcut taking you outside Wikipedia namespace to Help namespace rather than to WP:MOSLINK, and OVERLINK is formatted as a spin-off subpage of the Manual of Style page (so identified in its name) Wikipedia:Manual of Style (links). As it stands, it is a major project to figure out how many different pages deal with this, let alone where they are or that if you get to one you should be looking for a bunch of others too. Gene Nygaard (talk) 11:37, 8 January 2009 (UTC)
• Comment Why does the merger-suggestion "discuss" link take you to this strangely named "Break 1" section which starts out discussing something else entirely. Somebody should straighten this out. Gene Nygaard (talk)
  • Just about to sort that out. (And I support the merge as well, of course.)--Kotniski (talk) 12:02, 8 January 2009 (UTC)

When people are voting "yes" above, they're probably voting yes to the current merge proposal, which is to merge both pages into MOSLINK. Generally, that means starting with MOSLINK, and adding material from both pages that isn't already present in some form at MOSLINK, and "merging" generally carries a flavor of "without contradicting MOSLINK", although everything is subject to negotiation. Do you have time to give that a shot, Aervanath? (Feel free to keep working at the page you started; in the end, if the merge proposal passes, and that looks likely at the moment, we'll just end up with one page, WP:MOSLINK.) - Dan Dank55 (send/receive) 21:46, 4 October 2008 (UTC)

A two-step merge may not be a bad idea. Trying to deal with three things at once is not that easy. Dhaluza (talk) 22:52, 4 October 2008 (UTC)

This assumes that MOSLINK is generally right where the pages disagree. Why should we assume that? Septentrionalis PMAnderson 02:28, 7 October 2008 (UTC)

The reason I haven't tried to merge MOSLINK into it as well is that I would like one place to be more of a general outline of principle. The merger as I envision it would be a place to set out the general ideal ("Don't underlink, and don't overlink."), whereas MOSLINK would be the place where we could set out exactly what constitutes over- and under-linking and what should and should not be linked. I'd like to get some reasons on why that wouldn't be a good idea before I worry about cramming it into MOSLINK as well. I tried that before, and it just got too unwieldy.--Aervanath lives in the Orphanage 23:06, 4 October 2008 (UTC)

Only because, at the moment, I believe I'm seeing a 5 to 1 vote (assuming you're the one) on the current merge proposal. If you've got the time, working up two pages, a two-way merge and a three-way merge for comparison, would be great. - Dan Dank55 (send/receive) 00:40, 5 October 2008 (UTC)

Again, why should our poor editors have to go to two different pages to locate guidance about linking? Why this boundary between general and detailed? Tony (talk) 03:48, 5 October 2008 (UTC)

And I've now looked at this proposed "Build with context" (note sentence case). I don't like the way "underlinking" is privileged by first position (twice)—as the point of departure. Overlinking is like a cancer at WP, so encouraging people to link seems like a backward step; I've never understood Avaerneth's points about underlinking: where is it? Examples provided there are just the kind of items we've been unlinking for some time: "desert", "Africa", "18th century"—hello? Tony (talk) 03:54, 5 October 2008 (UTC)

IMO, including "underlinking" is essential to provide some balance.Dejvid (talk) 10:41, 5 October 2008 (UTC)

• Guys, if you think the current proposal doesn't look quite right, SOFIXIT. I'm pretty busy in the big scary Outside World at the moment, and don't have the time I'd wish to sit down and perfect this like it needs. I've provided a very rough draft as a starting place for people to work on. It can be ignored, improved, scrapped. expanded, shrunk or whatever.
• The reason I'd like a separation between the general and the detailed is that one would be a more stable statement of principle. I hope that the eventual Linking guideline would stay pretty consistent over time, changing only with widely established consensus, whereas MOSLINK would be the more nit-picky, detailed, battleground over the "link this, don't link that" items. If you guys don't think that it's a good idea, then, ok, I can live with that.
• Tony, I've pointed out Category:Dead-end pages at least twice to you before now. Please don't keep asking me for some articles which are underlinked. There are 500 of them in there. I've also pointed out Wikipedia:Dead-end pages, for some more examples, and Category:Orphaned articles, which has over 30,000 examples of articles that need to be linked to. Please tell me why these are not evidence of underlinking. Also, please support your contention of massive overlinking by providing links to a similar or even larger list/category of articles which suffer from overlinking. Until then, I really don't understand your point of view.--Aervanath lives in the Orphanage 11:22, 5 October 2008 (UTC)

OK, I went to this month's list. The first is a one-sentence stub, with one good link outwards. The second contains silly links such as "authors", "difficult" and "philosophy" in the first sentence. One of the sentences is this: "ldy rhFkZ xq# pj.k esa] lsok ti ri ;ksx AA opu osn ds lw= gSa] ^’kadj* gV x;k jksx AA sakala tirth guru caran+a". Um .... Tony (talk) 12:03, 5 October 2008 (UTC)

In any cleanup category, you're going to find articles which need cleanup in more than one area. However, I just went and clicked through to the first 10 articles in Category:Dead-end pages from October 2008, and found nothing that bad. (That would be Bajaj Club through Ecological Forecasting). What I did find was a lot of articles that needed improvement. They would especially be improved by adding some quality wikilinks to relevant topics.

I have yet to be pointed to this massive deluge of overlinking which you claim is drowning the encyclopedia.--Aervanath lives in the Orphanage 04:02, 7 October 2008 (UTC)

While I agree that merging here is good, I still find some faults with the current proposal.

First, on creating new links (i assume you're talk about redlinking). You should also make a statement to have the authors check to see if there previously was such a page and ifso, check why it was deleted. If they don't feel they can make a better page, or come up with a reason that conforms to WP:N and WP:V, they in such cases they should be told not to redlink such a page as it implies that a page should be there because it would meet those criteria.

Second, on overlinking, I would suggest also stating that unless there is a clear and compelling reason, anything linked in the article (except as references) should not be linked in a "See Also" and/or an "External Links" section(s).

Third, I would like there to me some note about dates because everyone seems to want to link any date in an article, whether it's relevant or not.

Finally, something should be stated in the opening about sections that by the nature of linking to the first instance of a word or phrase usage, articles should tend to be more heavily weighted to be link-heavy towards the top, specifically the lead.Jinnai (talk) 08:58, 26 October 2008 (UTC)

went ahead and edited info on 3/4 items. Not sure how to do the last one. It may be a bit rough, but it gets the points I wanted to make across.Jinnai (talk) 12:09, 27 October 2008 (UTC)

Protected

Due to the recent edit warring this page has been protected for 72 hours. Please use the time to discuss the matter here and come to a consensus on what should and shouldn't be included on the page. If an urgent edit needs to be made during the protection, please place the template {{editprotected}} here with details of the edit that needs to be made and justification for the edit, and an administrator will come by to make the edit. If you have agreed and resolved the dispute before the expiry of the protection, please make a listing at requests for unprotection. While it is also possible to make such requests on my talk page, it would be quicker for you to use those previous methods. Thank you. Stifle (talk) 12:33, 21 September 2008 (UTC)

RfC now open on linking dates of birth and death

Is it desirable or is it undesirable for dates of birth and death at the top of a biography article to be linked to corresponding "day" and "year" pages?

Further to discussions above, an RfC is now open at WT:MOSNUM#RfC: Linking of dates of birth and death -- Jheald (talk) 11:54, 29 September 2008 (UTC)

I would say that if the person was famous enough that their birth or death is noted at the linked date page, then it might be useful. Otherwise, it's just a "bridge to nowhere". Dhaluza (talk) 22:54, 4 October 2008 (UTC)

Sorry? What is a bridge to nowhere? Tony (talk) 03:46, 5 October 2008 (UTC)

I suppose Dhaluza refers to a (birth/death) date link. That is, unless we have decided to discuss Governor Palin from now on instead of boring Wikipedia stuff. Waltham, The Duke of 21:52, 6 October 2008 (UTC)

Moving links to footnotes

Would there be any interest in noting on this page that one way to add links of secondary interest is to move "editorial aside" comments to a footnote and link to the articles there. i.e. instead of putting an aside in brackets, put it in a note at the bottom of the article instead. For example, instead of linking Troy weight vs Avoirdupois weight in the main article, a footnote could be used to explicate what units are being used in the article, and to mention the different weight systems and to link to them. Carcharoth (talk) 22:08, 2 October 2008 (UTC)

I would argue that this sort of style advisory should be in WP:MOSLINK, not here. It is a good idea, though.--Aervanath lives in the Orphanage 22:51, 2 October 2008 (UTC)

Thanks. Raised here. Carcharoth (talk) 00:04, 3 October 2008 (UTC)

Yes, much better in footnotes than the main text; people who are really interested will read the link—this extra "filter" is highly desirable and makes a secondary link unobtrusive. Tony (talk) 03:45, 5 October 2008 (UTC)

most readers?

In saying what shouldn't be linked to, this guideline mentions "Items that would be familiar to most readers of the article...". I think we should replace "most readers" with "almost all readers". The beauty of links is that you simultaneously serve the beginner (by providing extra info) while serving the advanced reader (by letting them skip over unneeded reading). In a good article most readers should have no reason to click on most links, but all links should be useful to some readers. To me this guideline, seems to be overstressed about the problems of overlinking, without seeing the benefits of good linking.--Rob (talk) 07:10, 3 October 2008 (UTC)

Good observation. It sounds like you should be involved in the discussion above.--Aervanath lives in the Orphanage 19:07, 4 October 2008 (UTC)

Agree. Most readers can easily be confused for most readers like me, which brings in WP:BIAS. Almost all readers forces you to think more broadly, and consider people who may not be like you. Dhaluza (talk) 22:56, 4 October 2008 (UTC)

I don't agree with shifting the test towards linking. "Most readers" is fine by me; "almost all" is starting to give the link-everything crowd too much support in their arguments that particular items should be linked. "Most" is still sufficiently inexact to allow some leeway. Tony (talk) 03:44, 5 October 2008 (UTC)

Support the change; as an aside, I'd ask Tony to stop the divisive comments (such as the above "link-everything crowd"). --Ckatz^chat_spy 05:00, 7 October 2008 (UTC)

Should CONTEXT and BUILD be merged into MOSLINK?

The result of the discussion is pretty clearly merge all three. So, fine, I'm closing the poll, even though I disagree with the consensus. Someone please be bold and do it already. WP:Voting is evil.--Aervanath lives in the Orphanage 16:26, 7 October 2008 (UTC)

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.

Let's get some basic responses here on the proposal to merge this page (WP:CONTEXT and WP:BTW into WP:MOSLINK before proceeding on the details of the proposal.

• Support. Tony (talk) 12:28, 6 October 2008 (UTC)

• Support. Lightmouse (talk) 12:58, 6 October 2008 (UTC)
• Yep, doing things on one page is the default position; exceptions should be rare. - Dan Dank55 (send/receive) 13:03, 6 October 2008 (UTC)
• Support, seems like a generally good idea to me. Pagrashtak 14:20, 6 October 2008 (UTC)
• Fine in principle. These pages go back before section redirects, which is how they would now be configured. The alternative is to take the advice on when to link out of WP:MOSLINK, which was originally severely technical. Septentrionalis PMAnderson 16:45, 6 October 2008 (UTC)
• Oppose: Merge CONTEXT and BUILD into one statement of principle, use MOSLINK for the application of the principle.--Aervanath lives in the Orphanage 17:02, 6 October 2008 (UTC)
• Close the poll. We have to reach agreement on the proposed content before we can start arguing over trivialities like what page to file it under. Wikipedia successfully tolerates duplication in the vast majority of its policy and guideline pages. Consolidating BUILD and CONTEXT probably makes sense. Shoehorning that into MOSLINK might work but might not depending on the level of detail that we think readers need to see. MOSLINK does not good if our readers refuse to read the page because it becomes too big. We need to actually look at the content before we can make that as an informed decision, though. This proposal to decide the title before the content is exactly backwards from the way we should be making the decision. Rossami (talk) 17:54, 6 October 2008 (UTC)
• Since all these pages are style-related (and navigation-related, but not so much), it makes sense to merge them. It certainly is a good place to start reducing the sprawl characterising the Manual of Style. The current balance between warning against overlinking and underlinking should, of course, be maintained. I do think MOSLINK could expand a bit on the subject, which is what the merger should ensure (otherwise we are talking about the deletion of the two pages). Now, depending on the amount of information we wish to retain from the two pages, we might need to re-organise MOSLINK, and if it proves too much for the page to hold we need to consider the option of a page uniting "Build the web" and "CONTEXT". I suppose before deciding we should have an idea of how the merger is supposed to happen. Personally, I think not being too specific might make the guideline more effective. Waltham, The Duke of 22:16, 6 October 2008 (UTC)
• Turn CONTEXT and BUILD into essays espousing different viewpoints and merge anything useful into MOS. These pages strike me as instruction creep. We all agree that we shouldn't have "too many" or "too few" links, but we will never agree exactly how many. Leave it to editor discretion and local consensus. --Itub (talk) 06:41, 7 October 2008 (UTC)

This seems to disregard the clear evolution of opinion and practice on WP towards more selective wikilinking, if anything, to maximise the utility of high-value links. This section is intended to be simply a preliminary request for in-principle feedback, rather than opening gambits for maximum vs minimum linking policy. MOSLINK already spells out what should not be linked. Tony (talk) 08:11, 7 October 2008 (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.

Geographic example

User:Dhaluza recently added the following example as a footnote to the geographical section of this page.

<ref>The same principle applies to geographic place names. Link to [[Troy, New York]] rather than [[Troy, New York|]], [[New York]], since the more specific article should contain a link to the more general one already.</ref>

User:Spacepotato has repeatedly removed it, arguing first that it was "undiscussed", then that "there is no advantage" to it. I restored it because this example is, to me, identical in principle to the "flag of Tokelau" example in the 5th bullet of the "what generally should not be linked" section. That example has been uncontroversial on this page for over two years. Dhaluza's addition seemed a bit redundant but certainly not controversial.

Spacepotato has removed again, this time saying that "Obviously, it is not the same principle as Troy, New York contains a link to Troy, New York." I am completely confused by that comment. Troy, New York contains a link to New York just like flag of Tokelau contains a link to Tokelau. How is this different? Rossami (talk) 21:18, 6 October 2008 (UTC)

Writing [[flag of Tokelau]] instead of [[flag]] of [[Tokelau]] is superior because it provides a link to a more specific concept, flag of Tokelau, that the reader is more likely to be interested in. This is not a factor when deciding whether to write [[Troy, New York|Troy]], [[New York]] or [[Troy, New York]], because both examples contain a link to the more specific concept, viz., Troy, New York. Spacepotato (talk) 21:40, 6 October 2008 (UTC)

By that logic, our existing example should be rewritten to encourage the format the [[flag of Tokelau|flag]] of [[Tokelau]]. Tokelau and New York are general articles, both clearly linked from their respective specific articles and easily followed by any reader wanting to find the general concept. Rossami (talk) 22:35, 6 October 2008 (UTC)

The reader can find any article by some chain of links, or by using the search box, but this is not an argument against providing the link. As for linking to the [[flag of Tokelau|flag]] of [[Tokelau]], I agree. Spacepotato (talk) 22:46, 6 October 2008 (UTC)

One of the purposes of having the article title at Troy, New York (or even Ossining, New York, which is not so called for disambiguation) is to make linking to it easy and natural; the minority of readers who need a link to New York will find one in the article on Troy. Septentrionalis PMAnderson 23:18, 6 October 2008 (UTC)

We should absolutely not be switching to the format [[flag of Tokelau|flag]] of [[Tokelau]]. That would hide flag of Tokelau behind a piped link. As experienced users, we use pipes rather freely on Talk pages but the vast majority of our users don't know what that is and are confused when links take them someplace they didn't expect. The natural expectation of a reader seeing a bluelink for "flag" is that the link will take him/her to the general page on flags, not to a specific page about only one flag. The Manual of Style for piping is quite clear that we must keep the links as intuitive as possible. The change you're talking about would violate that principle. Rossami (talk) 15:46, 7 October 2008 (UTC)

This argument has some weight in the case of flag of Tokelau, but it seems not relevant for Troy, New York as, whatever format is used, clicking on Troy will lead to Troy. Spacepotato (talk) 21:16, 7 October 2008 (UTC)

Editors may find it pleasant to write [[Troy, New York]], but this does not address the question of which form is superior for the users of the encyclopedia. Spacepotato (talk) 00:04, 7 October 2008 (UTC)

But there are two sides to that. The convoluted form [[Troy, New York|Troy]], [[New York]] produces Troy, New York; but if you click on it, you may or may not get to Troy; you may get to New York without wanting to. Those readers will be inconvenienced by the form Spacepotato wants; the straightforward form will inconvenience only those readers who are led to New York and miss the link at the beginning of Troy, New York. This is why we have customs against Easter eggs in our links.

At this point, we consider that if editors are compelled to jump through hoops, they will spend time doing so rather than writing content; some of them will give up in disgust and not write any content. Neither serves the reader. Septentrionalis PMAnderson 02:26, 7 October 2008 (UTC)

Surely, readers who click on New York expect to go to New York. Also, it's a mistake to conflate the question of how articles are named (which is not consistent from region to region) with that of how to link. The issue of how to link Troy, New York is no different from that of linking Hamburg, Germany, despite the fact that the articles on the two towns are named differently. Spacepotato (talk) 08:55, 7 October 2008 (UTC)

No, Hamburg is different from Troy; English usage differs. Hamburg, Germany (except in contexts where disambiguation is necessary) is an archaic Americanism; Troy, New York is standard usage (except, again, in extreme cases, where the fact that it is the American Troy and which state we are considering has already been made clear). Since we should only be linking the first use of Troy, New York anyway, the exception rarely applies.

This is another proposal to reform the English language; I deplore it. Septentrionalis PMAnderson 14:29, 7 October 2008 (UTC)

The Troy, New York is identical in principle to flag of Tokelau, but it's even worse to link Troy, New York to two separate articles, because since the links are adjacent (except for the tiny black comma), people are likely to think that it is just one link and click in the wrong place. --Itub (talk) 06:37, 7 October 2008 (UTC)

What is wrong with Troy, New York? Focus the reader on the most unique part of the link. Tony (talk) 12:04, 7 October 2008 (UTC)

Tony, I assume you meant [[Troy, New York|Troy]], New York (you wrote [[Troy]], New York). I agree with Dhaluza that we shouldn't link to a general item when the specific article contains a link to the more general one. The link itself is about the city - the fact that the city lies in a particular state doesn't require a separate link to that state. The names of US cities are something a difficult case, because the function of disambiguation (distinguishing Portland, Oregon from Portland, Maine) is intertwined with the function of giving background information, the name of the state, in the same way as I might write that "Le Mans, France is twinned with Bolton, England"; not because disambiguation is required but simply to inform a reader who might otherwise have little idea where either town might be. I don't think that artificially squeezing the state out of the link, as Tony suggests, is a useful approach though - in everyday usage the name of the state is an intrinsic part of the city's name for many US cities, because there are a lot of duplicated names in different states. (It's fine to hide the state name completely via a pipe if it's clear from the context though.) Colonies Chris (talk) 14:01, 7 October 2008 (UTC)

And it is normal usage even when the name itself would be unambiguous, unless the state has already been established by context. Septentrionalis PMAnderson 14:29, 7 October 2008 (UTC)

Yes, my mistake—you're both right. Tony (talk) 10:10, 8 October 2008 (UTC)

I strongly prefer the form Troy, New York, partly because New York is easy to get to from that article (it's linked in the lead) and partly because it's important to think about print, where links not matching the names would naturally be rendered as cross-references. It's almost always better to have the link name match the link text if possible. Dcoetzee 21:50, 7 October 2008 (UTC)

Countries on infoboxes

Just found a country de-linked on an infobox. Can I ask where this was agreed or put forward to any community. GarethHolteDavies (talk) 09:04, 8 October 2008 (UTC)

Which country? If it was Tuvalu or Chad, the removal of the link should be reverted, since most English-speakers are unfamiliar with these countries, and a few may wish to interrupt their reading to go to those articles. The United States, the UK, Canada, Australia, France, Germany, Italy, Russia etc, (especially the anglophone ones), should not be linked unless there's a good reason to do so. There are usually much more valuable links in the vicinity, and the aim of wikilinking is to encourage readers to click on the ones that count. Dilution by low-value links has been a significant weakener of our brilliant wilinking system. Tony (talk) 10:08, 8 October 2008 (UTC)

United States? Are you so sure that all English readers are familiar with United States? There may be very good reasons for linking United States, in case people are not so familiar with the United States. Or might not realise that you mean the United States of Africa rather then the United States that is the new name for the thirteen colonies. Some of us here in Europe never bothered updating our maps much since Boston you know. Chad I know, Tuvalu, I've visited, United States is an enigma to me.--Scott MacDonald (talk) 10:31, 8 October 2008 (UTC)

In my opinion, infoboxes should be able to stand on their own, just like rows of tables. From Wikipedia:Only make links that are relevant to the context#What generally should not be linked: "A link that had last appeared much earlier in the article may be repeated, but generally not in the the same section. (Table entries are an exception to this; each row of a table should be able to stand on its own)." While I would not generally link United States more than once in an article (and definitely not at all if it was surrounded by a number of other links that were more relevant), linking it in an infobox detracts from no other links, since usually each entry is on a line by itself.--Aervanath lives in the Orphanage 16:56, 8 October 2008 (UTC)

"A link that had last appeared much earlier in the article may be repeated, but generally not in the the same section." Who put that one in? You? It's way distant from the mark, doesn't generally reflect practice, and will need to be removed, since it's an invitation to dilute other more deserving links. It does look as though we're just going to fight and fight: you to maximise link-spattering to unthinkingly ruin our linking system; I to push for selective, disciplined linking to strengthen it. You seem to come out at every turn trying to push the system towards more linking. Over my dead body. I'm afraid that I find your arguments for linking just about everything in infoboxes lacking in logical: in fact, they're feeble. Can you come up with something better to justify your claim? Tony (talk) 17:04, 8 October 2008 (UTC)

Variations of the wording that "A link that had last appeared much earlier in the article may be repeated" have been in this page for years and have been accepted practice at Wikipedia since the project began - both before Aervanath joined the project. Please don't personalize these disagreements. Rossami (talk) 18:38, 8 October 2008 (UTC)

Thank you, Rossami. Tony, I appreciate that you feel strongly that minimized linking will strengthen the encyclopedia, and that this is a reflection of your desire to make Wikipedia a better and more useful tool. Please meet me halfway by recognizing that I am also honestly advocating measures which I believe will improve the encyclopedia. The two of us disagree on what the best thing is to do in these cases. Just because someone disagrees with you doesn't mean they're trying to hurt the encyclopedia, nor does it mean that they are stupid, or blind, or malicious. I should also point out that the language I quoted has been in WP:CONTEXT since December 22, 2007, and was discussed on the talk page first.--Aervanath lives in the Orphanage 05:01, 9 October 2008 (UTC)

• I'm not willing to meet "half way", as you put it, because that will encourage people to put lots of low-value links in. When you audit articles and see "United States" linked SEVEN times, you really start to feel that the linking system is being abused. You're trying to encourage people to repeat-link in an article, This is destructive. Tony (talk) 05:53, 9 October 2008 (UTC)
  • When I said, "please meet me halfway", I was referring to my assumption of good faith on your part. Please do me the same favor. If you are unwilling to recognize that we are having an honest disagreement, and if you are unwilling to acknowledge that I am not maliciously trying to destroy Wikipedia, and you are unwilling to seek a consensus which is inclusive of opinions which may be different than yours, then I really don't know how you can be happy on Wikipedia, where Wikipedia:Consensus is one of our bedrock policies.--Aervanath lives in the Orphanage 06:56, 9 October 2008 (UTC)

• "Please do me the same favor."—I think your view, while in good faith, has evolved from a distorted foundation.
• "If you are unwilling to recognize that we are having an honest disagreement"—What is a "dishonest disagreement"?
• "if you are unwilling to acknowledge that I am not maliciously trying to destroy Wikipedia"—Have I imputed malice? Where? I just don't think you see clearly the inevitable damaging effects of your maximum-linking push, and I'll continue to fight it.
• "you are unwilling to seek a consensus which is inclusive of opinions which may be different than yours"—I will continue to fight for widespread smart linking (i.e., more selective than your notion of it) to protect the system. People are generally aware that linking went overboard a few years ago and are supportive of efforts to bring back a balance. I see you as trying to shift the pendulum back to overlinking, although you don't see it as overlinking. Just why an item should be linked twice in two adjacent sections, ever, is beyond me. You seem to have no appreciation of the disadvantages of poorly considered links. It's like a lolly shop, apparently. Tony (talk) 08:21, 9 October 2008 (UTC)

The answer to the question of why an item should be linked twice in two adjacent sections seems obvious to me: not everyone reads every word of an article from beginning to end. People are likely to jump directly to the section that interests them, skip sections, or just skim through the article. That is a well-known fact of web usability: see [1] and [2]. --Itub (talk) 09:27, 9 October 2008 (UTC)

We avoid making statements twice in an article (with a little leeway to repeat statements from the lead, hopefully in different wording); we avoid boring readers by repeating a lexical item soon after; we avoid redundant wording; we avoid wasting readers' time with these repetitions. If readers are so distracted or muddle-brained that they can't remember where a bright-blue link was on first occurrence (can't be hard: it's BRIGHT BLUE), they should put away their marijuana bong. We don't cater for stoned readers; nor should we say yes to editors who want to speckle repeated, unnecessary, redundant bright-blue through an article. No, no, no. Tony (talk) 14:38, 9 October 2008 (UTC)

I strongly disagree with linking an item in every section it is present in (even if only once). I also strongly disagree with not accepting the possibility that an item might need to be linked in adjacent sections (that an item might need to be linked more than once in an article, especially a long one, is, in my opinion, self-evident). First of all, we need to define "section": some people use the term to refer to the entire parts of the article under level-two headings, including sub-sections; for others, "section" and "sub-section" are the same thing. Furthermore, I agree with the point that certain sections in certain articles are accessed directly by people not interested in reading the rest of the article. We shouldn't inconvenience these readers and omit links they might need. We should not unduly generalise, of course; good judgement should be exercised in determining which sections are more likely to be read in isolation.

All in all, links are navigational elements at least as much as they are style elements, and while a certain degree of prescriptiveness is plausible for style issues, navigation is more dependent on the specific circumstances and requires more flexibility. We must reconcile these two approaches. Waltham, The Duke of 21:09, 9 October 2008 (UTC)

I am right with the Duke (and I'm very glad he said it first, so I can hide behind him ... I was about to offer this opinion myself). "A link that had last appeared much earlier in the article may be repeated, but generally not in the the same section." The current text seems succinct and about right to me. Wikipedia text is constantly shredded, deleted and rearranged; if we only allow one link to any particular word or phrase in an article, that guarantees that over time, we'll get text all over the place that loses its links. We don't need to link more than once in any section (I mean "==" section), normally, and for short sections, I'm in the habit of not re-linking if I see it in the section above or below. Maybe we should quantify and qualify this, I don't know. The main point is, I think it's almost universal practice to allow linking a word or term again if the last time it was linked was way above or below the current screen, and there's a limit to how much we can successfully use style guidelines to change Wikipedian practice, even if we wanted to. - Dan Dank55 (send/receive) 21:18, 9 October 2008 (UTC)

• It's a pity that you can't tell that the Duke of Waltham is talking utter, thoughtless rubbish. If someone links straight to a "section", I'd have thought they could at least skim through the rest of the article before hitting a further link, provide right IN the section (in EVERY section?) just in case someone is too lazy to read other sections. It's just poppyock that will give the green light to link-lunies who want to blue out terms every time they occur. NO. Tony (talk) 00:48, 10 October 2008 (UTC)

And it's not that I don't think this is "thoughtless rubbish", but having heavy 'flu and a bad headache probably encourages me to express it plainly, where otherwise I might try to be more subtle. Now that you've tried to make some kind of case for multiple linking within an article, I must ask you to provide examples of where it's significantly useful to the reader, when balanced against the disadvantages of dilution and scrappy appearance. Tony (talk) 01:27, 10 October 2008 (UTC)

It is your right to call my opinion "rubbish", but thoughtless it was not, I assure you. First of all, one linked to a section might want to read the lead for some context, but that doesn't mean that one would want to skim through the rest of the article, which would not necessarily be very relevant to the section in question. I do not call this "laziness"—especially when the reader concerned is in a hurry. And I didn't even have this case in mind when making my statement; many readers consciously choose to go straight to a section and ignore the rest. I know I do it at times, usually when a section mostly concerns a subject that has too little material to have its own article and so is covered in another.

Should we accuse the readers that they don't scan an article for links they might need later, in the part they'll want to read? I think not. We should not be telling our readers what to do; it is our goal to provide them what they will normally need. I agree that we must not go to extraordinary lengths to serve the readers' navigational needs, especially when other aspects of an article's good function could be compromised. Going to the other extreme is not much better, though. I fully understand your fear that we could be opening the Kerkoporta to compulsive over-linkers by not being stringent enough. But we should afford some flexibility where navigation is concerned, and leave some room for our editors with sound judgement to make good decisions on a case-by-case basis. A guideline that is not absolute lets less room for abuse, and overlinking can always be corrected and its perpertrator advised on the actual consensus. And we do not need laws to deal with editors who systematically violate consensus; the spirit is more important than the letter of our guidelines here.

You ask for examples. I've never been very good with them, but I'll make an effort here. In "Palace of Westminster", there is a link to Guy Fawkes in the list of prominent people tried at Westminster Hall, and then again at the "Incidents" section, where more details are given about the Gunpowder Conspiracy. Black Rod is linked to from the "Grounds" section—where the office is explicitly mentioned but cannot be elaborated on—and the "Security" section—where the office is properly introduced. And Giles Gilbert Scott is mentioned in the "History" and "Commons Chamber" sections, for the same reason but in different contexts. Do you think the removal of any of these six links would be an improvement?

And a general point: like statements in the lead should always be repeated in some form in the article's body, and hopefully fleshed out, so should the links in these statements. Such duplication I consider useful more often than not, because in both locations the statement is introducing something, and in either location the reader might feel they wish to read more.

I wish you a speedy recovery. Waltham, The Duke of 04:16, 10 October 2008 (UTC)

OK, so we'd better start insisting that abbreviations now given in parentheses after spelling out on only the initial occurrence (almost a universal practice) be spelt out every time during an article, for fear that someone might land in the middle of the article and not know what it stands for. This idea of peppering large amounts of navigational infrastructure through the text on the basis that the article per chance may not be read sequentially from top to bottom will open up a hornet's nest. What is wrong with casting the eyes above to the first occurrence of an abbreviation on the rare occasions that a reader parachutes into the middle of an article?

Likewise, I suppose you think it's unreasonable to expect a reader (to whom Guys Fawkes is important enough to hit a link to) is too muddle-headed or careless to either go back and locate its first, linked occurrence in the current article ("Palace of Westminster") and segue from there, which would be a smoother pathway given the context in the earlier section "Westminster Hall", where "Guy Fawkes" is rightly linked:

Westminster Hall also housed important trials, including impeachment trials and the state trials of King Charles I at the end of the English Civil War, Sir William Wallace, Sir Thomas More, John Cardinal Fisher, Guy Fawkes, the Earl of Strafford, the rebel Scottish Lords of the 1715 and 1745 uprisings, and Warren Hastings.

After all, the parachute did put the reader down to the Palace of Westminster article for a reason (not Guy Fawkes Day, which provides a quite different context.

I note that "State opening of Parliament" is linked neither in the later "Incidents" section nor on its first occurrence; this would be a good trade-off, improving the linking without adding blue-splash overall. "King James I" is another redundant link that would be better linked through its first occurrence at the top, where the context is clearer in relation to the topic. Why repeat earlier links when "House of Commons" is nowhere linked—it appears in "Incidents"? BTW, please don't rely on links to define an item: "Heywood, Greater Manchester" might be a guild or company, but no, it's "the town of ...". Why is "manure" linked? You should consider sorting out your linking priorities better. This is not what I'd call a skilled approach to linking: wasteful, more like it. It needs to be cleaned up throughout. Tony (talk) 14:20, 10 October 2008 (UTC)

I sympathise with the Duke's opinion here. I have never had the particular aversion to (individually defined) "excessive" blue-links that Tony has, nor the hatred of perceived overlinking. I simply don't find blue-links turn it into a scrappy appearance. In several articles I have edited I have seen the need to wikilink some words twice or three times, mainly because the articles are so long and blue-links are helpful to the reader. It is one of the selling points of Wikipedia, one of its USPs. If I am to be called a link-loony, so be it, but I feel very strongly about this. Editorial judgement should be shown here, we do not need to have a guidelines on these codified in concrete. Those who write the bloody articles should be given the licence to exercise their reasonable judgement. (By the way, what is poppock? ;) Woody (talk) 17:07, 10 October 2008 (UTC)

I also agree that we should not unnecessarily limit the editors; it is very clear from reading the discussions about over- and under-linking that a significant number of editors do not share Tony1's views regarding links. Tony, you are of course entitled to your opinions. However, that does not mean they are correct, and it most certainly does not mean that those who disagree with you are "wasteful", "link-lunies", "damaging" or any of the other less-that-complimentary terms you like to use. If you can't adapt your strongly-held personal views to fit those of others, perhaps it is time to step back from this particular guideline for a bit. This is especially true if you cannot discuss the subject without insulting your fellow editors. There is obvious support for using discretion in determining what should and should not be linked, but if you insist on pushing through your views while disregarding those of others, this discussion will just drag on forever. A war of attrition is neither desirable nor beneficial to the project, and only wastes energy that could otherwise be better spent actually improving the articles. --Ckatz^chat_spy 19:01, 10 October 2008 (UTC)

Katz, you insist on personalising the debate. I agree that editors should not be "unnecessarily" limited in their use of links: that is my whole thesis. What is necessary in a culture of undisciplined scattergun linking—that significantly weakens the system by dilution and ugly appearance—is quite different from what is unnecessary.

"If you can't adapt your strongly-held personal views to fit those of others, perhaps it is time to step back from this particular guideline for a bit."

Is this code for "If you don't submit to Katz'z views, go away"? That is what is sounds like. I'm not rude enough to suggest that you go away, so stop telling me to go away. It's insulting. You're stuck with me, I'm afraid, and over my dead body will you rule the roost here.Tony (talk) 02:47, 12 October 2008 (UTC)

Tony, quite frankly I don't care anymore what you say about me personally; you've tossed enough unsubstantiated accusations in my direction here and on other pages that it is now easy to ignore. All I said was that if this debate is getting you riled up - as your contributions would seem to suggest - you might wish to take a short break from it. However, if you're really so concerned about people "ruling the roost", why are you using terms such as "utter, thoughtless rubbish", and "feeble" to dismiss the opinions of other editors? At some point, we - you, me, and every other editor here - will have to reach some sort of compromise in order to resolve this. Shouting "over my dead body" and "no, no, no" doesn't help bring that any closer. --Ckatz^chat_spy 04:59, 12 October 2008 (UTC)

Oh, don't make us vomit. Have you resigned your adminship yet? If not, why not? You flagrantly abused one of the basic rules (conflict of interest) and possibly another (gaming another admin's action) elsewhere. You are an embarrassment to the position. You appeared not to understand the issue of conflict of interest, or did you rather feign a lack of understanding to avoid addressing your own abuse? You know what I'm referring to. Tony (talk) 07:09, 12 October 2008 (UTC)

As per my above note, I don't care what you say about me personally. Can you please address the questions about this guideline instead? --Ckatz^chat_spy 09:49, 12 October 2008 (UTC)

Can you two stop bickering like a bunch of schoolchildren? No one person is meant to rule the roost here or anywhere else in the myriad of MOS pages. Tony, please don't allude to past disagreements, this is not about them, this is about the conflicting ideolgies here. You say that links "significantly weakens the system by dilution and ugly appearance": that is an opinion, and it is yours. Please don't try and impose it on others which I feel you are doing here, and please tone it down. "Over my dead body helps no-one." There are conflicting ideologies at work here, you obviously feel very strongly about perceived overlinking, it is clear from this discussion that some people don't agree with you and feel that editor discrestion should be applied. Let us try and concentrate on reconciling the two approaches in a guideline acceptable to all. Woody (talk) 15:27, 12 October 2008 (UTC)

Well that rules you out as a person who can be trusted to take an even-handed stance. You clothe a statement that is basically antagonistic towards me in bookends that try to deceptively convey NPOV ("both of you"); I can no longer take you seriously. Tony (talk) 16:06, 12 October 2008 (UTC)

We all have our different stances Tony, they are called opinions, it is how we voice those opinions that defines our point. I was trying to be nice, if I wanted to insult, demean or belittle your opinion, I would have come out and said that. You were both acting like idiots trying to impose your conflicting views on everyone else. If you cannot see that, or accept that, then that is your problem. We all have opinions, just don't try and impose them on anyone else. I am not looking for your redemption or acceptance Tony, so your opinion of me is immaterial to this discussion or any other one. Regards Woody (talk) 16:20, 12 October 2008 (UTC)

Just as well, because my opinion of you is dropping by the hour. I won't return the personal attack by calling you an "idiot", no matter what I think. My complaint was your obvious personal bias, which rules you out in my mind as a trustworthy participant. Tony (talk) 11:25, 13 October 2008 (UTC)

If you consider criticism of your actions personal attacks, then I am loathe to continue this conversation. My personal bias? Look in the mirror Tony. We all have our opinions here, that is the idea of discussion, to see everyones opinions and then to discuss a compromise where they can all be reconciled. It is not a baby throwing their toys out of the pram if they people don't agree with their opinion. Woody (talk) 11:31, 13 October 2008 (UTC)

• No, I consider your name-calling the attack. I wish you'd stop personalising. By the way, when I look in the mirror, I like much better what I see than when I look at you. Tony (talk) 12:04, 13 October 2008 (UTC)

• I haven't called anyone any names actually. I have only responded in kind to your comments Tony. You made it personal, I am yet to see a reasoned argument other than dilution and they are ugly, which is in itself subjective. You have padded your comments with personal attacks and insinuations, you have made it clear that you will not compromise or listen to other opinions. That is not name-calling, that is just stating my interpretation of your comments above, and I don't see any other way of interpretating them. What else is there to say? Woody (talk) 17:37, 13 October 2008 (UTC)

I think I need to clarify a few things. I am against overlinking. I am all about choosing carefully which words should be linked and which not. The trouble is in deciding what is overlinking and what is not. Although general rules are helpful and desirable, and apply to most articles, different cases may require different arrangements, so I believe leaving a little wiggling room without forcing one to invoke "Ignore all rules" every time will prevent several undesirable things from happening, like a) having a blanket rule that people will follow to the letter even when they shouldn't; b) having a very complex guideline that will try to cover all the eventualities and will ultimately be ignored; and c) having a very rigid rule that editors in favour of unduly liberal linking will try to twist to their advantage. It is the spirit of the consensus that matters, and this consensus we are trying to reach.

I repeat that the specific subject of linking is affected by many different factors. The concept to be linked and its relevance to the subject is only one. The length of an article and its subdivision into sections also matter, as does the existence of navigational and other elements. I've always wondered, for example, about linking practices in captions, as well as the extent to which a link present in a navbox should be repeated in "See also". And I really don't think the widespread practice of repeating links from "see also" hatnotes in the same sections is right at all. But I digress.

Not all links are the same. Some links, like to "United States" and "18th century", are useless in the vast majority of cases and should be avoided entirely, except in exceptional circumstances. The introduction of terminology, units of measurement, etc. should only be accompanied by links on first mention, because one is expected to read such scientific articles (or sections) top-to-bottom or risk not understanding them. Other, more general links could be used in slightly different ways, like, for example, in their first mention and their most important one (where elaboration follows). I'm leaving the lead out for now. These two links can be very far apart in text (double linking is better-justified, or even only justified, in long articles), although that could mean that the first link would be near the beginning of a long history section and the second link in the next section, which is somehow more specialised. We are talking about level-two (==) sections, but not all operate on this assumption; that's why we should be extra-careful with such terminology in guidance.

And I now realise that I am not sure what exactly we are debating here. Theorising is nice, but this will have to be eventually translated into something material, most likely in the form of a statement in a guideline. Considering that the paperwork has been approved and this essay is supposed to be merged into MOSLINK on the first opportunity, I'd like to have some solid knowledge about what will be transferred, added, snipped, and who knows what else, before making a conclusive statement about any changes.

Concerning "Palace of Westminster", you've made a few good points. It's a B-class article, so it will need improvement (I also want to expand it in areas, but haven't got down to it yet, and I'm not very good with sources, either). Amongst other things, I have removed the "manure" link, although I cannot say it's completely redundant in an article discussing a building full of politicians.

And one last note... Please don't use bullets in threaded discussions. This is a colon-only venue. :-) Waltham, The Duke of 21:28, 11 October 2008 (UTC)

I am not one for overlinking, I understand where there is a line. We don't need to link United States in every mention in the text, we should link them in infoboxes in my opinion, or in circumstances where they are in a line and are needed for consistency: ie Chad, Guinea, Nigeria, Somalia, United States and Mexico looks inconsistent if US is not linked. That is why we don't need a concrete guideline which a bot can link to in an edit summary. Editor discretion is needed, we should be given the ability to provide reasons why it is needed, as opposed to someone using AWB and saying read WP:OVERLINK. Woody (talk) 22:03, 11 October 2008 (UTC)

Can someone (like Spacepotato) explain the rationale for the parenthetical sentence, which has been removed and reinstated in the past 24 hours:

The same link multiple times. Redundant links make future maintenance harder. A link that had last appeared much earlier in the article may be repeated, but generally not in the the same section. (Table entries are an exception to this; each row of a table should be able to stand on its own)."

In addition, I believe the time has come to add a rider into "but generally not in the same section", which appears to be an invitation to link in adjacent sections. Repeat links, I believe, should be countenanced, but only in better defined circumstances. Tony (talk) 02:38, 12 October 2008 (UTC)

I cannot speak for normal tables, but sortable ones should not follow the "only link on the first instance" rule because the first instance can suddenly become fifth of twenty-third (lost somewhere between 68 entries). Maybe we should make this clearer. In MOSLINK. Waltham, The Duke of 10:04, 12 October 2008 (UTC)

Copy-edit of MOSLINK

I think MOSLINK is a mess per se. I'm going through it gradually, making it easier to follow, better organised, and conducive to the conflation of this page, Build and MOSLINK; there are queries at the talk page here. Tony (talk) 15:10, 12 October 2008 (UTC)

WP:OVERLINK as a main article for the WP:MOSLINK section on over/underlinking

I've read all the argument above, which has yet to produce a fix. Meanwhile, OVERLINK is a fairly long and good argument dealing with both over and underlinking, but not much on anything else. This, it comes to me that (while you're working on the Ultimate Fix), it might be good to simply reference WP:OVERLINK (which also covers "UNDERLINK") as the main article for the relatively short section on under-and-overlinking, in the Manual of Style on links (MOSLINK). Thus, WP:OVERLINK becomes the main article for (just the) short WP:MOSLINK section that covers over/underlinking. So I've gone and done that. It's an improvement! I do not claim it's optimal. SBHarris 18:08, 14 October 2008 (UTC)

I think the best idea is just to delete this page once the material is added to MOSLINK. It's quite unnecessary for the information to exist in two places. Tony (talk) 11:19, 23 October 2008 (UTC)

Note: redirect not delete for GFDL reasons. I am with Tony, there is no need for unneccessary duplication, we just need to merge them as soon as possible, especially if people are linking to this in edit summaries. Woody (talk) 15:38, 23 October 2008 (UTC)

Not applicable within infoboxes

Added line to article; basically regarding that countries within infoboxes are to be treated differently than those within articles and not de-linked. If anyone wants to tweak the grammar they are welcome to do so.Londo06 09:14, 23 October 2008 (UTC)

I don't want to tweak the grammar; I'd like to know what your rationale is. I don't see why infoboxes are suddenly different from any other text WRT to trivial, low-value links. Tony (talk) 11:05, 23 October 2008 (UTC)

I have the same question. Why should infobox links be different? Duplicative links make subsequent maintenance harder. Why should we put our editors through that? Rossami (talk) 18:10, 23 October 2008 (UTC)

Whoa! yet another exception to 'link only once'?? I have been having issues with the exemptions for articles with tables, infoboxes. And hey, why not succession boxes too? While I may accept the necessity of linking once in each large table, I wonder if the person(s) who inserted that bit I coloured in red realises what scale of overlinking that wording creates. Just for example, this version of the article Billie Jean King -admittedly an excessively long one- had in excess of ten links each for Christine Truman, Ann Haydon-Jones, Evonne Goolagong, Chris Evert. Some of these links appeared in the tables on that article. I believe the current position is clearly over the top: "What generally should not be linked- the same link multiple times. Redundant links make future maintenance harder. A link that had last appeared much earlier in the article may be repeated, but generally not in the same section. (Table entries are an exception to this; each row of a table should be able to stand on its own.)" The first two sentences are in clear contradiction to the part in brackets. I intend to strike out "row of a" from the current version. Ohconfucius (talk) 13:59, 22 November 2008 (UTC)

Whoa! One persons opinion does not a discussion make. If you have a sortable table, say at List of Victoria Cross recipients of the Royal Navy, this can be sorted several different ways. The point about each row is that when sorted, the default first row is no longer so it will not be the first linked. The reader then has to go scrolling for a link, which is unacceptable. So, I am going to revert until a discussion has occurred as this has been debated, and accepted, before. Regards. Woody (talk) 14:07, 22 November 2008 (UTC)

• With all due respects, in cases like the one you pointed out, the argument for complete delinking within a table is even stronger. If it must be 'all or nothing', then 'nothing' would be a better fit with the spirit of the guideline, not creating thes monstrous exceptions. Contents of each cell and situation has undoubtedly been referred to and linked to in the preceding paragraphs and sections, so why there an insistence on linking Crimean War 2427 times within the table? And surely it would make sense where tables do not sort?? Ohconfucius (talk) 14:40, 22 November 2008 (UTC)

• (edit conflict, you changed the meaning of your para in that edit conflict) You say undoubtedly, have you read the paragraphs and the list? I take great exception to you calling it monstrous. This list is sorted as a default, by ship name; if you sort by name then you have a completely different list. The only linking first instance rule is worthless here, as such each line has to be taken as its own first instance. Linking nothing in the table is not an option, you may as well edit for Britannica. Regards. Woody (talk) 14:47, 22 November 2008 (UTC)
• The contents of each cell and situation have not been referred to elsewhere, that is the point behind the list. The links to the conflict are neccessary as it provides the background knowledge about the circumstances in which these awards were made. If a reader is looking at the list, they may want to know what happened in that conflict, they will then click on the link and find out. Why should we make the reader scroll all the way up to the top, why should they have to expend effort when it can be linked so easily? Woody (talk) 14:51, 22 November 2008 (UTC)

• Now looking at your example, Crimean War, WWI and WWII are all linked in the paragraph immediately preceding it, and to respectively link 27, 48 and 23 times within a table would appear to not too common-sensical? And what if the table is not a sorting one? Ohconfucius (talk) 03:46, 23 November 2008 (UTC)

A link does not suddenly become necessary just because it has some lines around it. Lightmouse (talk) 00:42, 23 November 2008 (UTC)

• I have already replied to this, please read them before carrying on a circular argument. Woody (talk) 10:11, 23 November 2008 (UTC)

No, it must have been deemed neccessary for someone to have put the lines around it. I have no idea what you are trying to say. Woody (talk) 00:53, 23 November 2008 (UTC)

slow merger

So, does anyone object if I copy over the sections "What should be linked" and "What should not be linked" to MOSLINK? This would be a major step in rationalising these pages (i.e., in getting rid of this one). Tony (talk) 12:12, 27 October 2008 (UTC)

Please do. Lightmouse (talk) 14:35, 27 October 2008 (UTC)

Names of common publications in references

A script run by Dispenser (talk · contribs) is, while making several useful corrections, also delinking the names of some publications in references. Dispenser feels that this is supported by WP:OVERLINK. I can see the arguments on both sides of this: on the one hand, it's probably safe to assume that the vast majority of our readers are familiar with, say, CNN, BBC or the New York Times, and providing a link to the publication in a reference probably is of limited utility. On the other hand, Dispenser's bot is also removing links to publications like The Observer, which is well known in one country but not necessarily as well known internationally. Has there been any discussion of when the names of publications should and should not be linked? Is there a consensus about this? —Josiah Rowe (talk • contribs) 18:39, 27 October 2008 (UTC)

It's never been explicitly discussed that I know of.
Having spot-checked a few of the edits made by that user, I am inclined to support the delinking. In each case that I spot-checked, there was a direct link to the article or website. That article link included the domain name of the hosting organization (CNN, BBC, The Observer, etc). An interested reader could fairly easily continue from the article-link to the publisher's site.
Checking the versions before the delinking (where the reference included links for both the article and the publisher), the reference was moderately difficult to read. It was hard to tell exactly where the article's link ended and the publisher's link began and has a fair potential for error if a reader accidentally hit the wrong link. And because many article links are redirected to the publisher's main page after a time, it was plausible that a casual reader would not recognize the mistake.
I personally did not see much value in that subset of links where the publisher linked to a Wikipedia article about the publication. The jump from Bottlenose Dolphin to BBC was too tangential to be relevant. (In those rare situations where the credibility of the reporting organization might be relevant to the article, I would seriously hope that the credibility issue is raised in the text, not buried in a hopeful footnote link.) Rossami (talk) 20:00, 27 October 2008 (UTC)

Citations are already a sea of blue, and the systematic linking of publisher names will inevitably result in multiple linkins of the same item throughout. I don't quite understand the reader's need to go directly to the site of the publisher—it's the evidence they'll be concentrating on. If there's some doubt about the authority of the publisher, it will be relatively rare and will be worth typing the publisher's name into the search box. That is what the search box is for. Tony (talk) 01:11, 28 October 2008 (UTC)

I also agree that linking the publication name is probably not useful. Since Tony and I are at opposite ends of the spectrum on linking, then you've probably got consensus that they shouldn't be linked.--Aervanath lives in the Orphanage 04:10, 28 October 2008 (UTC)

Geographic locations?

Alright, I've read (most) of the previous discussions about this, but I find inclusion of this criterion in links to avoid to be misplaced.

First, I agree that multiple links to such articles should be excised, as should links to generic concepts such as "the". That's not what I'm disputing here. My dispute is with the statements that links to "well-known" concepts are value-less, creating a "sea of blue".

As an example, consider this edit to Canadian Confederation. The first link removed was to Canada, which I think is quite wrong. In my opinion, removing a link to such a relevant article related to the subject makes no sense.

Further, the contention from those that support this new policy is that well-known concepts need no further explanation, that a reader will already be familiar with it. (And that anything learned in primary school need not be linked.) In my opinion, a superficial familiarity with a subject does not imply that a visitor to WP isn't interested in reading further about that subject, or that s/he is familiar with the complete contents of the WP article. A South African with passing familiarity of Canada will likely not know more than a few basic facts about it, and certainly wouldn't know everything presented in the article.

For another example, a recent edit to Timeline of same-sex marriage removed all links to country articles (except those in South America, such as Columbia and Argentina). I think one link to each country mentioned is acceptable, and wouldn't create that "sea of blue" mentioned earlier. (I do agree with removing multiple instances of the links though.)

The previous discussions didn't have input from the broader WP community; at most a dozen editors contributed to its inclusion. I think you'll find more editors complaining about this as articles on their watchlists are edited according to this new policy, since they were likely unaware of its discussion and implementation. I was aware of the new policy regarding date delinking (and agree with that too), but this is the first I've heard about delinking of geographic terms. Had I been aware of it at the time, I would have opposed its implementation as it currently exists. I doubt I'm the only one in this position.

To state it concisely, my reasons for objecting this new policy are:

• passing familiarity of a subject isn't the same as knowledge about that subject; readers may want to pursue more in-depth knowledge of that subject
• from a usability perspective, we shouldn't force users to type in (or copy and paste) a title into a search box just so they can view them
• assuming that some subjects are "well-known" to readers is presumptuous
• not everyone reading English WP is a native English speaker; those editors have a different set of well-known subjects
• removing such links also removes access to sub-articles linked on those pages, of which the reader may not be aware (for example, the link to Territorial evolution of Canada in the history section of the article about Canada)

Frankly, I think this needs more analysis and broader community input, because it is a major change to the way Wikipedia functions. Mindmatrix 15:01, 29 October 2008 (UTC)

Response from Tony1: I agree that the first instance of "Canada" in Canadian Confederation would have been better retained in its linked form, and possibly even the names of the other federal countries; but I see "London" has been relinked—why? Apart from "London", I've no problem with the partial reversions. Links to whole country articles (not even sections therein, if you please) in Timeline of same-sex marriage are pretty useless. I don't see why countries should be linked even once in that article: please provide an example of how this would increase readers' understanding of the topic, as required.

To respond to your bullets:

• "passing familiarity of a subject isn't the same as knowledge about that subject; readers may want to pursue more in-depth knowledge of that subject"—Readers may wish to pursue in-depth knowledge in many items in an article (or not in an article); it doesn't mean that we need to blue-bruise the text with scattergun links just in case. The unfortunate effect of that would be to dilute the high-value links with little benefit, leading to a weakening overall of the wikilinking system. Linking is quite rightly recognised by WP as a whole as an art in itself, a facility that needs to be skilfully applied, selectively, for greatest effect. It is a pity it took so many years for this to get through.
• "from a usability perspective, we shouldn't force users to type in (or copy and paste) a title into a search box just so they can view them"—Nor should we force readers to put up with a low-value speckling of bright-blue through the text. A balance is required, and WP began with a completely undisciplined approach to linking. We are only now grappling properly with this.
• "assuming that some subjects are "well-known" to readers is presumptuous"—So is assuming that our readers are extremely ignorant, or all six-year-old children.
• "not everyone reading English WP is a native English speaker; those editors have a different set of well-known subjects"—Well I'm sorry, if they want to read an English-language site, they can't expect the text to be littered with bright-blue magic carpets just for them. The whole text would be covered in links if that were the principle. Nor would I expect the Polish WP to link lots of items just in case an English-speaker wanted to consult it.
• "removing such links also removes access to sub-articles linked on those pages, of which the reader may not be aware (for example, the link to Territorial evolution of Canada in the history section of the article about Canada". Yes, idea for insertion into the "See also" section, where the readers are more likely to catch it and follow it up.

I think you assume far too great a willingness by readers to hit links. The suspicion that it's really rather hard to get them to hit links partly motivates many WPs to be more selective in their choice of bluing. Tony (talk) 10:46, 31 October 2008 (UTC)

(My apologies for the delay in response - I had forgotten about this discussion.)

First, could you please not state as fact what is clearly your opinion; in my comments above, I noted what I was expressing my opinion, not points of fact. (I'm specifically referring to statements such as "to blue-bruise the text with scattergun links", "low-value speckling of bright-blue through the text" and "expect the text to be littered with bright-blue magic carpets", which in my opinion add no value to the discussion.)

To respond to your points:

• In response to my point about removing access to sub-articles (eg - Territorial evolution of Canada, found on Canada article), I think you missed my general point. I wasn't addressing the issue for only the Canadian Confederation article, but rather for any article on which a link to Canada exists (or would be removed). Consider an article such as Rush (band). I'd like to retain the link to the well-known subject Canada, so that readers may have access to these sub-articles. Are you suggesting we remove the link to Canada (which you already did), and add links to all its sub-articles in the "See also" section of the Rush article? If not, in what way do readers have access to the Canada sub-articles if you remove the link to Canada?
• I don't assume that "readers are extremely ignorant, or all six-year-old children". In fact, I assume nothing about the knowledge of readers, either as a collective readership or as individuals. In my opinion, Wikipedia should ensure that the needs of a broad set of readers are addressed.
• I believe my point about "not everyone reading English WP is a native English speaker" is valid. English has become the de facto language of international communication (see, for example, English language#English as a global language). Moreover, English Wikipedia is far more active than many of the smaller WP projects, so readers tend to use it when their own language version of Wikipedia does not (adequately) cover a subject. (I don't have stats for this, but I'm certain some basic research about this exists somewhere.) What's done on Polish or other WP projects is immaterial to this discussion.
• "you assume far too great a willingness by readers to hit links"; likewise, you assume a lack of willingness of readers to follow links. There's an easy way to determine the truth in this matter, given that WP access logs are massive and contain detailed information. Let's ask one of the developers to analyse these logs (they're not open to the public for privacy reasons) and determine article click-through rates. This will completely remove any opinion from the matter and provide solid quantitative evidence.
• "Linking is quite rightly recognised by WP as a whole as an art in itself" - where, and by whom? - "a facility that needs to be skilfully applied, selectively, for greatest effect" - sure. Now, define the ambiguous term "selectively" in a manner that's acceptable to all editors. Linking itself is not an art, and requires little skill, as it simply requires the addition of a few brackets, and possibly pipe text. That leaves the choice of articles to link as the skill - and we clearly disagree on this.
• "please provide an example of how this would increase readers' understanding of the topic, as required" - the purpose of a link is to provide further information about a related or relevant subject, not about the topic of the article from which the link is made. By the way, MOSLINK was edited by you to contain the phrase "Items in Wikipedia articles can be linked to other Wikipedia articles that provide information that significantly adds to readers' understanding of the topic." I can't find discussion from that time which addressed this issue, and your edit summary didn't indicate that change. Before your change, it stated "The use of links to other Wikipedia articles, for example, Ant (resulting in Ant), is encouraged". Yes, the change was made in 2007, but it seems to me that one editor is making significant changes to the style guide without broad discussion, and then using those changes to justify further changes in other parts of the style guide.
• "The unfortunate effect of that would be to dilute the high-value links with little benefit, leading to a weakening overall of the wikilinking system" - that's your opinion. I see the removal of some (but not all) links as decreasing the utility and value of Wikipedia. This is why I asked for broader input on this before making widepspread changes throughout Wikipedia. I think that Build the web is an apt guideline: "Remember that what may seem like an irrelevant link to you may actually be useful to other readers."

Anyway, I'd like to solicit input from other editors, so from where should we recruit them? Clearly, this won't be resolved with just the two of us involved in the discussion. We're certainly in agreement on a few things, for example that there is overlinking on some pages, but we differ on which or how many links to remove. Mindmatrix 16:46, 7 November 2008 (UTC)

There's a balance to be struck. What it comes down to is that in a particular context, some links are more valuable than others. There's little contention over linking the high-value links, or not linking the low-value links like the, but there is understandably no sharp dividing line in the middle, nor is there an objective way of pinpointing link value. It may be informed by things like link density and avoiding redundant links, but at the end of the day it's a matter of personal style and consistency within an article is the most important thing.

This problem isn't just one with geographic locations, but in any circumstance where the linked topic is very general. For example, if a person studied biology at university, but is not a biologist, should we link biology in their article? What if it was something more specific like entomology? Should an article about a sporting event link to its sponsor companies? Should an article about a race of people link to anatomical articles like skin? It's very difficult to create guidelines that will apply across all articles. Dcoetzee 20:44, 7 November 2008 (UTC)

Agreed that it's not just about geographic locations - it just happened to come to my attention when articles on my watchlist were edited under this particular criterion. I also agree that there isn't a policy that can be crafted that could be both sufficiently broad and sufficiently specific to cover all situations. My concern is that edits in the past few months have used the current policy to eliminate links that some consider to have value. I suppose my point is that the current focus should be to eliminate non-contentious links (duplicates etc.), and leave the others until a better policy is defined.

BTW: I know some researchers have had access to the Wikipedia access logs. Have any papers been published regarding any aspect of usage patterns, links, etc.? I didn't find anything doing a quick web search, but I'm sure there must be something out in the wild by now. Mindmatrix 00:59, 8 November 2008 (UTC)

To answer my own question, I found these, which I have yet to read, and which may or may not be useful to this discussion:

• Extracting Semantic Relationships between Wikipedia Categories (PDF)
• Discovering Missing Links in Wikipedia (PDF)
• Link density in wikis wiki-research mailing list discussion
• Preferential attachment in the growth of social networks: The internet encyclopedia Wikipedia (PDF)
• Extraction and analysis of tripartite relationships from Wikipedia
• Experiments and Evaluation of Link Discovery in the Wikipedia (PDF)
• Wikis as a Technology Fostering Knowledge Maturing: What we can learn from Wikipedia (PDF)
• Autonomously Semantifying Wikipedia (PDF)
• Average Path Length in Complex Networks: Patterns and Predictions (PDF)
• Use of Wikipedia Categories in Entity Ranking (PDF)

The above were found doing an internet search for "wikipedia link density". I'll peruse a few in the near-future, but if someone finds useful tidbits in the meantime, please post them. Mindmatrix 01:25, 8 November 2008 (UTC)

• " Are you suggesting we remove the link to Canada (which you already did), and add links to all its sub-articles in the "See also" section of the Rush article?" A "See also" link to the culture section of "Canada", or something more specific, like "List of Canadian bands in the 2000s", or "Rock music in Canada" might be appropriate. Not the whole of the "Canada" article: that is not helpful.
• "In fact, I assume nothing about the knowledge of readers, either as a collective readership or as individuals"—what, not even that they read English? There's a "Simple English" WP for children and non-native speakers. Cast the net too widely and you'll flatten it, dumb it down.
• "Selectively" is an art, yes, but needs guidelines. I will continue to ensure that links to huge or general articles are made only where they are useful, and that editors are encouraged to link to more specific locations where they have the urge to direct our readers to a sea of largely irrelevant information. We're not here to facilitate discretionary browsing at the expense of high-value links and the look and readability of the text.
• "that's your opinion"—everything that drops from my lips is my opinion.
• Build the web is a bizarre page that has somehow become a guideline (without discussion or consensus, I might say). It is an embarrassment. A few statements could be drawn from it in the merger into MOSLINK—indeed, I've drawn on its spirit in my recent recasting of the lead to MOSLINK.
• "Ant"? Um ... well, I rest my case. Just where would a link to "and" be appropriate? In "Timeline of same-sex marriage"?Tony (talk) 11:17, 8 November 2008 (UTC)

I don't have the time to address this now, but I'll answer your last point: your interpretation that the statement I listed above requires or necessitates a link to "ant" (or "and" - was this a typo?) in articles is wrong. It was an example demonstrating that a link to ant could be made if the text contains the word. Clearly it wouldn't exist in the Timeline of same-sex marriage article. Perhaps I should have included the subsequent statement, which was "Use the links for all words and terms that are relevant to the article". Mindmatrix 14:20, 8 November 2008 (UTC)

The suspension of the removal of geographical locations within infoboxes probably wants a mention somewhere.Londo06 01:11, 10 November 2008 (UTC)

What suspension? Tony (talk) 10:18, 10 November 2008 (UTC)

There was a hold placed on the removal of geographical locations after a number of users objected to the de-linking. We have conversed on the matter, some of the talk took place at your talk page.Londo06 11:11, 10 November 2008 (UTC)

News to me. Tony (talk) 13:58, 10 November 2008 (UTC)

Please take a gander through your talk pages; jist of it was that I was a supporter of de-linking, just not within infoboxes. A number of other editors were either in strong support or staunch opposition of the de-linking within the infoboxes. This was around late October.Londo06 17:24, 10 November 2008 (UTC)

suggestion re year-in-X links

the current phrasing in WP:CONTEXT#Dates regarding year-in-X links seems unclear/confusing:

Piped links to pages that are more focused on a topic than stand-alone chronological pages (1999) are possible (1997), but have the disadvantage that readers do not, prima facie, recognize that the link is piped. If an explicit link is provided, preferably in the lead, it alone can be the gateway for the reader to access the available sibling articles for other years (e.g., 1998), making multiple links throughout the article unnecessary.

first: i thought that current policy recommends not piping year-in-X links; if that's what that statement is trying to say, it needs to be made a lot clearer. second: why does it suggest providing an explicit link in the lead, and then exemplify that with an aliased link? to me "explicit link" means "not aliased" - ie, 1998 in South African sport.

i suggest replacing that statement with something like:

Links to articles on a topic-specific chronological period - for example 1441 in art, 1982 in film and 18th century in United States history - should be kept explicit, so that readers can see what they lead to. These links (often referred to as "year-in-X links") provide access to sibling articles about other years, so multiple year-in-X links throughout the article are not necessary. Explicit year-in-X links can be included in the main text of the article (for example: "For more context, see 1964 in music") or in the article's "see also" section.

Sssoul (talk) 09:21, 31 October 2008 (UTC)

I've made a slight improvement (depiping that 1998 link, which obviously didn't make sense), but have no objections to further improvements along your suggested lines.--Kotniski (talk) 09:33, 31 October 2008 (UTC)

• I believe that once we have reviewed the links to the year-in-X articles, we may come to realise that these are as extensively and indiscriminately linked as the date articles. I am certainly finding a lot of these on my travels through wikiland. To be continued... Ohconfucius (talk) 10:26, 31 October 2008 (UTC)

you may be right, Ohconfucius, but in the meantime it would still be reasonable to clarify the bit that i quoted above. Kotniski, thanks for the change you made - it's an improvement for sure, but i feel it could be made clearer still. the wording i suggested is not necessarily the greatest; it would be good to see suggestions for fine-tuning it. Sssoul (talk) 10:48, 31 October 2008 (UTC)

My conclusion is the same as Ohconfucius. Some people think 'year-in-X' links are targetted but they aren't. Many articles even have 'autoformatted' dates that include 'year-in-X' which proves that many 'year-in-X' links have been added by people that don't know what they are doing. I support the revised text as it is. Lightmouse (talk) 11:15, 31 October 2008 (UTC)

Lightmouse, do you mean you support the revision that Kotniski made, or the wording i suggested above (which i've altered a bit in the meantime)? Sssoul (talk) 12:02, 31 October 2008 (UTC)

It seems to be on the right lines. I have just added a specific clause that says:

• Piped links break autoformatting. Therefore formats such as [[August 1]], [[1997 in radio|1997]] must not be used.

This is a common error. It is sometimes difficult to stop people adding such links and sometimes difficult to persuade them not to revert a correction. I am not wedded to the precise wording as long as this common error is mentioned. Lightmouse (talk) 14:32, 11 November 2008 (UTC)

I've removed this addition, as I believe this requires more discussion, and this is perhaps the better place for it to take place than on your talkpage, Lightmouse. Since the paragraph immediately following explicitly states that dates should no longer be linked for the purpose of autoformatting, it doesn't make a lot of sense to add a proscription against a link that breaks autoformatting.

I think there's some logic in the rewording you all are trying to achieve above. I'd agree that multiple links to 'year-in-X' articles within the same article don't make a lot of sense - that linking once ought to do the job. The case that I think still needs to be considered is how to best handle these in infoboxes, where the date a radio station is established links to a 'year-in-radio' article or the date that a baseball team first joins a league links to a 'year-in-baseball' article. In these cases, the "see also" solution isn't necessarily the most elegant. Mlaffs (talk) 14:54, 11 November 2008 (UTC)

Thanks. The basic point I have been trying to convey is that you can't have autoformatting AND concealed links. You can only have one. If you want one, you must remove the other. Can we find some words that convey that? To respond to your point about infoboxes, I think categories are better than concealed links. Lightmouse (talk) 15:07, 11 November 2008 (UTC)

Since autoformatting is deprecated and concealed links are not would not the logical solution be to remove the link from the month-day pair and leave the concealed link alone? That removes the deprecated autoformatting, broken or not, problem solved. - Dravecky (talk) 17:48, 11 November 2008 (UTC)

• It should always be perfectly clear what a reader will be taken to if they click on a link. They should not have to dwell their cursor over a link to discover that 1997 actually links to 1997 in radio (or 1997 in architecture or 1997 in science, or 1997 in music). Aliasing a link to appear as “1997” is a sure-fire way to get readers to not  click on links right now due to our long practice of having linked to those “January 1”-type trivia articles, which are virtually never germane to the subject matter of the article containing the link.

  If editors find that providing proper disclosure (either without aliasing at all, or with aliasing that is just as descriptive but more appropriate) still results in cumbersome-looking links in body text, they can always put a bullet in the article’s See also section.

  We simply have to resist the temptation to make wonderful and surprising Easter eggs that make readers go up and down in their chairs and clap their hands in an “oh, goody goody”-fashion when they click on 1795 only to discover they are taken to 1795 in science; just let ‘em know where they’re going before they click. Greg L (talk) 21:40, 11 November 2008 (UTC)

Concealed links are deprecated: WP:EGG and what Greg just said. --NE2 21:41, 11 November 2008 (UTC)

Deprecated "in most cases" is what WP:EGG actually says, with infoboxes being one of the specific cases where it's envisioned that they might properly be used. Mlaffs (talk) 21:57, 11 November 2008 (UTC)

The practical difficulty that arises in working 1908 in physics into a date in prose should be addressed with some suggestions on pipelinking, e.g. "Consider piped links such as (Context: 1908 in physics) or (more on 1908 in physics) to clarify linked topics." -LeadSongDog (talk) 15:10, 14 November 2008 (UTC)

actually "Context: 1968 in music" or "For more context, see 1968 in music" works at least as well as a piped "(Context: 1968 in music)". but sure, specific suggestions are helpful. meanwhile, does anyone besides me still like the amendment i proposed above? this one, i mean:

Links to articles on a topic-specific chronological period - for example 1441 in art, 1982 in film and 18th century in United States history - should be kept explicit, so that readers can see what they lead to. These links (often referred to as "year-in-X links") provide access to sibling articles about other years, so multiple year-in-X links throughout the article are not necessary. Explicit year-in-X links can be included in the main text of the article (for example: "For more context, see 1964 in music") or in the article's "see also" section.

Sssoul (talk) 19:03, 14 November 2008 (UTC)

I have decided that Lightbot will not help fix these errors. I was motivated to trace and fix these errors but I am encountering too much hostility for my efforts, even block threats. It's all very silly. See discussion at Wikipedia:Administrators'_noticeboard#Blocking_a_bot. Lightmouse (talk) 00:43, 12 November 2008 (UTC)

• Oh, don’t get all discouraged and pissy here, Lightmouse. It’s true that one can take WP:BOLD and WP:Ignore all rules to excess, but, I believe it was Jimbo himself who was responsible for WP:IAR. Just do your best to identify whether A) there is a general consensus for something, B) that your bot properly implements the gist of the consensus, and C) that you truly believe what you are doing is good for Wikipedia. Your bots, IMO, always help Wikipedia and it appears to me that you are the most prolific creator of these. If someone threatens you with a block or other such nonsense, others here (including me) will go to bat for you and simply point to A, B, and C above; that’s all you need. Besides, this isn’t like loosing all the data on your hard drive when you’re on the verge of discovering The Cure to Type 2 diabetes or something. F*ck ‘em if you’ve got a piss ant on your ass. Just do what you think is the right thing and ignore the whinners. Greg L (talk) 02:05, 12 November 2008 (UTC)

Yeah, "general consensus" for a smarter, more skilled approach to linking, and for the total removal of DA, is at WP:FAC in utter spadefuls. Since the community's decision on DA back in August, for example, I'm not aware of one objection by any nominator at FAC (or FLC) about the issue—and FA nominators can be quite wilful (ask any of the reviewers). There was argy-bargy with Tony the Tiger earlier in the year about his practice of linking dictionary-type words, but I think he's toned down that practice since reflecting more about it, and is on record as strongly supporting the removal of DA. But that is an exception. I'm surprised that even "concealed" links are rare in FAs, and that nominators appear to be OK about making them more explicit (but this is still a big problem, not yet properly raised and debated, at FLC).

So I, too, am most disappointed that Lightbot seems to have been browbeaten into excluding the removal of Easter-egg concealed links. I'm concerned that we provide optimal ways for alerting our readers to gateways into the relevant "year-in-X" pages, and making them look like plain year-links is clearly doing the opposite. Tony (talk) 08:46, 12 November 2008 (UTC)

Actually, Tony, I too am disappointed that Lightmouse has abandoned these particular efforts, although for different reasons. But I think that characterizing the valid concerns that I and others raised during this discussion (which now spreads across this page, Lightmouse's talk page, and AN) as browbeating, or describing those of us raising the concerns as "piss ants" (?) and "whinners" as Greg L did above, is an unfortunate choice of words. This is not just another in the seemingly endless stream of DA threads, although I fear it's being received in much the same way - it's a lot more nuanced than that.

I'll try to articulate my concerns one more time. First, should Lightbot have been fixing links such as [[October 6]], [[1995 in radio|1995]] that "break" autoformatting? Despite the fact that such edits are within the scope of the bot's approval, I would argue no - autoformatting is deprecated, so broken autoformatting doesn't need to be fixed. Second, as a consequence of these fixes, should Lightbot have been removing these concealed "year in" links from infoboxes? Again, despite the fact that such edits are within the scope of the bot's approval, I would argue no - concealed links such as these within infoboxes are explicitly envisioned within the MOS.

Third - although I didn't go down this path previously since I was hoping to keep the discussion yesterday focused to the first two concerns - should Lightbot have been removing concealed "year in" links outside of infoboxes? I would actually argue yes, that's a good task for the bot. However, while those concealed links aren't the best way to provide access to that information, removing them without adding a direct link to a "See also" section as has been suggested here and elsewhere is equally unhelpful - an inelegant link to good information is better than no link at all, at least IMHO.

How could all of these concerns be settled both within the scope of the bot's approval and consistent with the MOS? Infoboxes using a date that includes a concealed link could be fixed by removing the markup around the day/month and leaving the markup and piped link at the year in place, so as not to give an impression that the linking is for the purposes of autoformatting. This particular solution was proposed by at least a couple of editors yesterday. Concealed links elsewhere in articles piping to a "year in" article could be removed and replaced with a "See also" link, as was suggested by others. In fact, the former of these is exactly what I did yesterday on the 59 radio station articles where I chose to fix Lightbot's edits, and I'll probably go back over the next few days and do the latter on those articles and others where Lightbot's edits were not to the infobox.

I don't pretend to understand bot coding, and I have enormous respect for those users who build and run them as a result. I believe that Lightbot could handle this solution, but Lightmouse would have to confirm whether or not that's the case. If Lightmouse chooses to add this code to the bot and continue to include the radio station articles in its sweeps as a result, I think it would be a net benefit to the project and I'd back their efforts if questioned. On the other hand, if Lightmouse prefers to remain disengaged on the issue, we'll try to address these situations manually as time and workload permits. Mlaffs (talk) 15:13, 12 November 2008 (UTC)

Mlaffs, I agree with you that removal of autoformatting also removes any errors it contains. Many editors think/thought 'dates must be linked' and simply decided that a 'year-in' link is better than a link to an unfocussed year article. You can see that in many cases, the 'year-in' links are merely another variant of autoformatting or solitary years. If the autoformatting or solitary years weren't there, the 'year-in' links wouldn't be either in many cases. We all know of articles full of autoformatting and solitary years. You can see many articles that are similarly full of 'year-in' links and that is consistent with the idea that no particular thought has gone into them. In some cases, it has been a search/replace action where autoformatted dates have been turned into broken autoformatted dates containing a 'year-in' link. I hope that explains why I am inclined to think that broken autoformatted dates should be reinstated rather than reinforcing a 'year-in' link that should not have been part of autoformatted date in the first place. On top of that, you can add my cynicism about click-through rates. I don't believe people click on solitary years and a link that looks like a solitary year will be treated like one.

Now, on the technical side, it would be very easy for you to use my script to delink the dates. It is not configured specifically for the 'year in radio' links, but you would find it much easier than manual editing. All you have to do is add the line:

importScript('User:Lightmouse/monobook.js/script.js');

to User:Mlaffs/monobook.js and clear your cache using the instructions on that page. Then when you are in edit mode, you will see some new 'buttons' below 'What links here' at the left of your page. You can use these buttons to delink the dates and then you can put the 'year in radio' link back in. Even though I don't think it is a good idea, I am offering that as a way for you to get what you want. I hope that helps. Lightmouse (talk) 18:29, 12 November 2008 (UTC)

Chronological articles: exception

In the spirit of compromise, I tentatively support the recent inclusion of the exception concerning the linking of chronological items:

"Chronological items—such as days, years, decades and centuries—should generally not be linked unless they are demonstrably likely to deepen readers' understanding of the topic. Articles about other chronological items or related topics are an exception to this guideline."

However, can someone tell me what "related topics" means? How are these defined? Tony (talk) 14:07, 11 November 2008 (UTC)

In the spirit of Potter Stewart: we'll know them when we see them. This is a guideline, not an Inland Revenue code. Septentrionalis PMAnderson 19:20, 11 November 2008 (UTC)

If no one's gonna tell me what "related topics" means, I'm gonna take out the reference to it. I'll give it a couple of days. Feedback welcome. We do not want fuzzy guidelines, since that's an invitation for disputes at article talk pages. Tony (talk) 08:49, 12 November 2008 (UTC)

You are speaking for yourself, Tony. There is no we here. Septentrionalis PMAnderson 22:33, 12 November 2008 (UTC)

Sometimes it makes sense to leave some latitude for interpretation - I think it's risky to just put "articles about chronological items" because there might be other similar exceptions. It's not a huge deal, there's always IAR. Dcoetzee 22:45, 12 November 2008 (UTC)

Indeed. Firstly, its a mere guideline. Secondly, the first sentence contains the word 'generally' and that implies flexibility. Thirdly, there is IAR (as User:Dcoetzee says). I don't see a need for the second sentence, but if it were reworded to say:

• Articles that describe calendar entities themselves (e.g. [[Sunday]], [[August]]) can contain cross-links to other calendar entities.

I think that would be an improvement. But I think that is covered anyway by the first sentence. Lightmouse (talk) 00:23, 13 November 2008 (UTC)

I think the current reasoning it pretty solid. I think "chronological items or related topics" provides a defined but broad framework to work with. I think we can agree that "related topics" is intended to preserve date linking in articles that may not fall under the "chronological" category but still may demand date links for similar reasons as the chronological articles do. A rewrite would only be needed if this part of MOS was being "gamed" or used in a way not intended by its original meaning. Do others agree? Sillyfolkboy (talk) 01:58, 24 November 2008 (UTC)

I propose:

For lists, such as lists of works in a biography article, where unpiped links would visually overwhelm short lines of text, piped links should be used, and readers should be informed at the top of the list that year links are to specific topical year articles (such as 1990 in television). See T.S. Eliot#Bibliography.

The reasoning would seem to be obvious. That T.S. Eliot list is an excellent example where 1930 in poetry and similar year-in links make the list much more difficult to read, and, frankly, annoying. I think the discussion so far has considered links within prose, where this problem doesn't exist. Reconsideration (talk) 23:44, 25 November 2008 (UTC)

I rather like Lighmouse's wording (perhaps calendar items). Do people support Reconsideration's proposal? In any case, perhaps we should wait until MOSNUM talk works itself out on these matters. Tony (talk) 11:11, 26 November 2008 (UTC)

I want to make sure this is not used as a backdoor for re-introducing date links. Language here should be specific and restrictive. "… other chronological items or related topics …" does not meet that requirement.--Goodmorningworld (talk) 12:55, 26 November 2008 (UTC)

Here is the revised proposal with 'calendar entities' changed to 'calendar items':

• Articles that describe calendar items themselves (e.g. [[Sunday]], [[August]]) can contain cross-links to other calendar entities.

What do you think of that? Lightmouse (talk) 09:38, 27 November 2008 (UTC)

Thanks, Lightmouse—it's good, except "can" is ambiguous. "May"? Grammar fix towards the end. And more importantly, are we going to bite the bullet and insert a year or decade in the examples? What about this: "Articles on calendar items themselves (e.g. [[Sunday]], [[August]], [[1830s]], [[1954]]) may contain crosslinks to articles on other calendar items."Tony (talk) 12:46, 27 November 2008 (UTC)

Looks good to me. Lightmouse (talk) 12:55, 27 November 2008 (UTC)

I am of the mind that all dates "demonstrably likely to deepen readers' understanding." Dates are always relevant. It bothers me that date links are being removed. Kingturtle (talk) 16:30, 31 December 2008 (UTC)

Yes, but date links do come at a cost, which must be balanced against their utility. When I ask people for examples of useful links, usually we get nowhere. Do you have some? Tony (talk) 16:36, 31 December 2008 (UTC)

Dates are relevant, but the information in the date articles is not. Date links lead to articles that are full of unfocused trivia, which is usually irrelevant to the article the date was linked from. Except for articles about these chronological items, I have never seen a "useful" date link. Dabomb87 (talk) 18:50, 31 December 2008 (UTC)

"demonstrably likely to deepen readers' understanding of the topic" is a conceit, and tremendously patronizing. How about we stop treating our readers like children and allow them to browse as they like? — Hex (❝?!❞) 17:35, 3 January 2009 (UTC)

I don't understand what you mean. Readers are free to browse as they like. What makes you think they aren't? Lightmouse (talk) 18:21, 3 January 2009 (UTC)

Dabomb87, Information in the date articles is not relevant? There are WikiProjects dedicated to keeping date articles to high standards. You should read the long history of threads there. Much work and effort is put into keeping those articles free of the "unfocused trivia" you fear. Furthermore, although you have never seen a useful date link, I have seen date links as educational. History is contextual, and one of the contexts is time. The Peterhouse, Cambridge was founded in 1284. It is interesting contextually and culturally that the same year Wales and England were formally incorporated into the entity of England and Wales (done through the Statute of Rhuddlan). It would be useful for the readers of Peterhouse, Cambridge and Statute of Rhuddlan to be able to click on 1284. In fact, these items are of such note that the people at Wikipedia:WikiProject Years include each of these articles in the article 1284. If the events are so important to 1284, should their articles have 1284 linked to 1284? I think so. Kingturtle (talk) 15:30, 4 January 2009 (UTC)

True. But there have been links to years in places such as "However, there is no formula for general quintic equations over the rationals in terms of radicals; this is known as the Abel–Ruffini theorem, first published in 1824, ..." Now, it is totally irrelevant to the modern understanding of the topic that the theorem being discussed was published in 1824 rather than, say, 1624 or 1874, so a link would be rather pointless. And I daresay that most links to years were of this kind.

Also, whereas some articles (e.g. 1345) do provide context to someone clicking such a link from an article about something that happened in that year, others (e.g. 2008) are pure listcruft. (But the solution to this problem would be moving 2008 to List of 2008 events, or deleting it altogether, and writing a new 2008 article from scratch similar to the 1345 one.) -- Army1987 – Deeds, not words. 17:32, 5 January 2009 (UTC)

Are you sure the article 1284 is relevant to the article Peterhouse, Cambridge? I found 23 historical events listed there. One repeated the event of Peterhouse's foundation; 21 others were interesting in their own right, but totally irrelevant to the foundation of Peterhouse (like King Charles of Naples being captured in a naval battle); and then there's the Statute of Rhuddlan. I had a good look at that article but couldn't find anything that had a bearing on Peterhouse. I think it's simple coincidence that they occurred in the same year. If somebody could show that the incorporation of Wales had a bearing on the founding of Peterhouse, then I'd suggest that ought to be in the Peterhouse, Cambridge article. As it is, I'm completely unconvinced that sufficient relevancy exists to justify a link to the article 1284 from Peterhouse, Cambridge. --RexxS (talk) 23:36, 6 January 2009 (UTC)

A list of notable events is not a useful link. If there are a couple events from that year that are relevant, they could probably be mentioned in the article itself. It may be a useful article, and is accessible through the search box. Now, if year and month-day articles could be written and organized to the extent of [[1345]—which actually does provide context—I would reconsider my stance. Also, Tony mentioned a while back that year articles before a certain year should be conflated into decade articles. I would support that idea also, which I believe would make those articles more useful and easier to provide context in. Kingturtle, understand that I am not attacking the year articles and demanding that they be deleted. I am just noting that the way they are organized now, they are not much use as articles to be linked to. Dabomb87 (talk) 03:17, 7 January 2009 (UTC)

Year links are much better-served when they are linked to an article about a year in that topic: 2008 in poetry, 2005 in sports, etc. Dabomb87 (talk) 03:21, 7 January 2009 (UTC)

The other thing that bothers me about this your assumption that you know exactly what readers want and what they are looking for. You're removing a way to explore Wikipedia. It might not be your cup of tea, but there are people out there who enjoy date associations. Kingturtle (talk) 12:56, 8 January 2009 (UTC)

Links to years in poetry/literature in lists of books written

Many biography articles of poets have had links in the "Books" or "Works" sections (essentially a list of works near the bottom of the article) that link to individual "____ year in literature" or "____ year in poetry" articles. For example, see Thomas McCarthy (poet)#Books. This seems useful to readers, but it isn't reflected in style. Shouldn't it be? Reconsideration (talk) 12:59, 21 November 2008 (UTC)

I'm reposting this part of a comment I left on Lightmouse's talk page: Another problem is that the full phrase "1930 in literature" can look a bit bulky on a list. See T.S. Eliot#Bibliography. I think it would make the list a bit harder to read, therefore annoying to readers. Reconsideration (talk) 18:35, 21 November 2008 (UTC)

Tables

How does "Table entries are an exception to this; each row of a table should be able to stand on its own." - this even make sense? If I have a table with the same person's name list in each row, why would I link to it every single time?  BIGNOLE  (Contact me) 13:32, 30 November 2008 (UTC)

It doesn't make sense to me either. Lightmouse (talk) 17:15, 30 November 2008 (UTC)

People may scroll down to only the entry they want, or even sort the table if it's sortable. --NE2 17:36, 30 November 2008 (UTC)

Overall, I do not support multiple links on the same phrase in a table, however, the counter-arguments I considered are: 1) In a large table, it might be tedious to search for the first instance of an item -- particularly if it isn't linked either! 2) As a reader, I wouldn't be quite comfortable going to a different place in the table, and making the assumption that the Wiki link applies equally to the other entry, 3) It's confusing to have two items linked in a cell, and to have a third, apparently similar one, not linked, 4) Subjectively, personally, a patchwork of linked and unlinked words in a column looks ugly. A table with "USA" or "UK" linked every time (or not at all) looks better. That all being said, I still prefer the minimum amount of table linking. Alpha Ralpha Boulevard (talk) 19:36, 30 November 2008 (UTC)

NE2, and to Alpha, what about when a table of TV episodes for 1 season (approximately 20 rows) has the same director linked over and over again, when there is almost no way to look at the single table and not see the first and last episode in the same frame (the only real way would be if you had your resolution set very high). I mean, when the first four rows have the same name linked over and over again, at what point does "exclusion" become "over done"? I could understand linking each "first" instance, in each individual table, but not every single instance in the same table that only holds 20 odd episode rows.  BIGNOLE  (Contact me) 23:02, 30 November 2008 (UTC)

I'm gonna put my two pence on in this one. Another reason the episode list in question does this is because individual episodes have link anchors. As such they can be linked to for reference and a user shouldn't be forced to scroll up to get the link to the director, writer etc. that they desire. (e.g. If I was linking The Strip episode, a user should need to scroll up find another link to go to the Allan Heinberg article). Rambo's Revenge (talk) 23:17, 30 November 2008 (UTC)

They'd have to scroll up to the episode that is listed 3 rows above? That's so strenuous that we need to link every single instance?  BIGNOLE  (Contact me) 23:20, 30 November 2008 (UTC)

Repeat linking is worse than scrolling/arrowing/paging. If I am interested in a table, I scroll/arrow/page to put the table on a single screen. If it is too big for one screen, I am already using scroll/arrow/page to read it. Lightmouse (talk) 09:50, 1 December 2008 (UTC)

• Those who support our wikilinking system please note: blueing out every entry in a list makes it look as though this is a global formatting choice; many—even most—readers are unlikely to hit any of the links in the column under these circumstances. There are several altervatives that are likely to garner more hits. Shall we discuss them here? Tony (talk) 11:00, 1 December 2008 (UTC)

I believe we should. I can respect, and accept that linking once per table (if the same information is present in the other tables) could help the reader not have to search for a link to that particular page, but linking a half dozen times in a 20 row table is a little excessive. Scrolling is not strenuous.  BIGNOLE  (Contact me) 12:34, 1 December 2008 (UTC)

True, and yes, scattering a few links through a long column can look untidy. What do you think of this, for example? I'm OK with it—at least readers are alerted to the first occurrence of the items when they're linked, and are more likely to hit them.

When all year-in-X items in a column are linked, one alternative is to link only the top one (visually not too untidy) as a gateway into the whole set via the navbox at the top of that linked-to article; another is to insert an explicit note above the table; another is to include one or two of the links in the "See also" section. Tony (talk) 13:10, 1 December 2008 (UTC)

In regards to normal wikitables, I believe that only the first occurence of the item should be linked. However, many reviewers think that sortable tables—including Tony's example—should have their items linked every time because when a column is sorted, you don't know which item will come first in a column. This seems to be a precedent set by our Featured lists. Dabomb87 (talk) 13:35, 1 December 2008 (UTC)

I don't think this is something that can be (or should be) codified to cover all possible examples. I think this is an area where editor discretion should be allowed. You cannot compare a sortable table of 100 items with a small table of 20 items embedded in an article. They are wholly different things which editors and readers treat differently. Tony's example looks unprofessional to me, it looks ramshackle. In that case, in my opinion it should all linked or none; its current format does not look good to me. I agree with Dabomb regarding the sortable tables because the first linked item will change. I think any attempt at rigidly codifying this would be bureacracy gone too far. Regards. Woody (talk) 13:51, 1 December 2008 (UTC)

We should not create links for aesthetic reasons. Lightmouse (talk) 14:03, 1 December 2008 (UTC)

I agree. I can see where some would have issue when you're dealing with sortable tables, but list of episode article don't have sortable tables - at least not generally. Why should we link all instances when you can clearly see just about all of them in a single frame?  BIGNOLE  (Contact me) 16:28, 1 December 2008 (UTC)

Linking nationalities

While I agree with most of this guideline, I do not agree that nationalities in the lead of niographical articles should be delinked. This is a critical part of an individual's identity and is worth linking. Recently a script seems to have been used that delinks nationalities, and I feel this is going too far. -- Necrothesp (talk) 14:16, 21 December 2008 (UTC)

Just how does linking "UK" or "Australia" or "Indian" or "French" or "United States" help readers who consult the English-language Wikipedia? Is the object to provide a magic blue carpet for discretionary browsing? I can't imagine how the article on the United States, which is understandably voluminous and general in its angle, would increase the understanding of the reader of an article on an American band. How? Do you have examples? I believe, also, that it's possible to type the name of the country into the search box to the left, above the "Go" button; but why a reader would want to interrupt their reading of an article on the band by going to such a huge-picture article is unfathomable. We could provide links to "human" and to "Earth", too, in case there's a need for the big stuff. Tony (talk) 14:31, 21 December 2008 (UTC)

RFC started

I have started an RFC for a centralized discussion of the issue of linking units in articles. I arbitrarily chose the talk page of WP:MOSNUM, but I am also leaving notices on as many relevent talk pages as possible to attract centralized attention to this. See Wikipedia talk:Manual of Style (dates and numbers)#Centralized discussion for linking of units of measurement. Please carry on all further discussion at that location. Thanks. --Jayron32.talk.contribs 21:21, 31 December 2008 (UTC)

Linking miles and chains

I'm working on some articles that give distances in Miles and Chains. Under this guideline, miles should not be linked as it is a common unit of measurement, though I'm inclined to link to chain (length) as it is not in common usage. The problem is that with the {{convert}} template, though able to deal with miles and chains, only seems to be able to link both or neither. So which is the better option - linking both miles and chains, or leaving both unlinked? I've raised this at Template talk:Convert, but had no reply as yet. —  Tivedshambo  _(t/c) 23:15, 1 January 2009 (UTC)

Personally I think the best option is modifying the template so that you can choose whether to link or not (or simply having two different templates). Dcoetzee 05:28, 2 January 2009 (UTC)

The template does allow you to select whether to link or not, but does not discriminate which. So for example: {{convert|2|mi|24|chain}} gives 2 miles 24 chains (3.7 km), but {{convert|2|mi|24|chain|lk=in}} gives 2 miles 24 chains (3.7 km). —  Tivedshambo  _(t/c) 09:31, 2 January 2009 (UTC)

What about binning the convert template? You can even run it for the computation, then key in the results manually with full control. Tony (talk) 12:30, 2 January 2009 (UTC)

Shakira is an American singer

Hint: she is from the Republic of Colombia. Jennifer Lopez is an American singer too. Someone removed the wikilink from the Jennifer Lopez article, and as a result I can't tell what the author meant by "American". This is unacceptable.

Having worked on the perpetuity of information in a famous National Library, I know that semantic is more important than text, and the latter evolves much faster than one would expect. Wikipedia articles successfully identify concepts, and a wikilink is a firm way to refer to a concept. A word refers to something today, and to something else a few years later. As an example, there is a discussion at the Spanish wikipedia on whether Estados Unidos refers to the United States of Mexico or to the United States of America, and I would not be surprised if the term "United States" became more and more ambiguous in the next ten years, even in the English language. Given the size of Wikipedia, it is time to write clearly.

Putting every expressed concept in a blue link might have drawbacks, so here I propose a new kind of "light wikilinks". They could be implemented in various ways, for instance a web user may see a light dotted underline that shows the meant concept in a tooltip when clicked. What do you think about it ? Thanks ! Nicolas1981 (talk) 15:03, 4 January 2009 (UTC)

Merge proposal

It has been agreed above (see #Idea for merge and later discussion/poll) that this page, and WP:BUILD, are to be merged into WP:MOSLINK. Is anyone doing anything about implementing this?--Kotniski (talk) 12:11, 8 January 2009 (UTC)

Something should be done; it appears to me, even though different proposals are jumbled together above, that merging of all three pages into Wikipedia:Manual of Style (links) is supported. Now we need to get somebody to handle the tedious details. Gene Nygaard (talk) 13:09, 8 January 2009 (UTC)

Oops, only two pages. My problem was the overuse of shortcuts by the people putting in the merge proposals, ant not realizing that Wikipedia:Only make links that are relevant to the context has sortcuts at WP:CONTEXT as well as WP:OVERLINK and who knows what else. The article's actual name should have been used, not a shortcut, in the merge suggestion. And the fact that the discussion was on the wrong talk page; if you are discussing moving two different articles into one article, the discussion would best be done on the talk page of the article proposed as the target of the move, not on one of the pages to be moved. Gene Nygaard (talk) 13:20, 8 January 2009 (UTC)

True, but anyway, the discussion links seem always to have been pointing to this talk page. Are you saying, then, that WP:BUILD should not be merged? Seems a bit pointless to leave it separate, since it's basically a single paragraph explaining why we make links in the first place.--Kotniski (talk) 13:51, 8 January 2009 (UTC)

I've now hacked out a merged version of WP:MOSLINK, WP:OVERLINK (aka WP:CONTEXT) and WP:BUILD, as we agreed was desirable. It's at WP:Manual of Style (links)/merged. Please comment on that talk page. I know it still needs brushing up, but let me know if I've left out anything major. If there are no objections, I'm planning on substituting it for the current version of MOSLINK, and redirecting this page to there (or rather to the overlinking section of there).--Kotniski (talk) 19:59, 9 January 2009 (UTC)

MERGE NOW DONE as described above.--Kotniski (talk) 13:30, 11 January 2009 (UTC)

• The merge is half-baked an premature. I don't know if this is yet another trenchline being drawn on the date-linking issue, but there are just a few toooo many changes going on for my liking. Now's not the time to such through radical mergers where things could get lost or misinterpreted. I see the potential for yet more conflict here and we need to see the outcome of the Locke's Arbitration case before we can come to any firm conclusion on how to merge it. I am undoing it for now. Ohconfucius (talk) 02:05, 12 January 2009 (UTC)

• Instead of being disruptive and reverting (and partially reverting at that; it took Kotniski an hour to merge these pages) over one sentence and two words, please instead join discussion at WT:MOSLINK. —Locke Cole • t • c 02:14, 12 January 2009 (UTC)

• Look who's accusing someone else of disruption - While I respect he's spent considerable effort, that in itself is irrelevant how many hours were spent - if it's premature, spending 200 hours on consummating the merger doesn't change squat. Ohconfucius (talk) 02:48, 12 January 2009 (UTC)

"Premature" seems a bit unfair - there was widespread support for the principle of the merger months ago, yet no-one could be bothered to actually write the merged version. Then when I actually wrote the merged version, there was unanimous support for it. Then suddenly when it went live, a few people started vehemently opposing it - effectively on the grounds that the version that went live differed from the agreed version in one word, "demonstrably". I'm quite happy for this word to be put back if anything hinges on it, but whatever the result over that, it's quite unreasonable to say that the merger was made without consensus just because of that one-word difference. Since one of the reasons why people supported the merger was to reduce the number of separate guideline pages, restoring this page goes quite against the spirit of what the community decided. Can we agree, in the cold light of morning (it's morning here, at any rate), that the right decision is to make this page a redirect again?--Kotniski (talk) 06:55, 12 January 2009 (UTC)

Go ahead. You will surely hear from me if I find anything. Ohconfucius (talk) 07:04, 12 January 2009 (UTC)

OK, I have. If you think there's anything from this page that hasn't been adequately reflected in the new version, I suggest bringing it up at WT:MOSLINK.--Kotniski (talk) 07:19, 12 January 2009 (UTC)

• I do believe it was premature; however, as I stated at MOSLINK, I'm OK with it now. I'm sorry there was a kerfuffle, and I do appreciate the skill and hard work you put in, Kotniski. Tony (talk) 07:33, 12 January 2009 (UTC)

Cleaning up

These archives and the discussions on this talk page should probably be moved to their new home (or, at the least, this page should be manually archived and then redirected to WT:MOSLINK, then the archive box there should be updated with all the archives from here). Thoughts? —Locke Cole • t • c 07:12, 12 January 2009 (UTC)

• I would archive this page as it is now, copy over the archive box to WT:MOSLINK as archives from WP:CONTEXT, then redirect this namespace. Ohconfucius (talk) 07:18, 12 January 2009 (UTC)
  • Good idea, I agree :-) Nicolas1981 (talk) 11:29, 12 January 2009 (UTC)