Wikipedia talk:Only make links that are relevant to the context/Archive 2

This is an archive of past discussions on Wikipedia:Only make links that are relevant to the context. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

From Village pump

The article Richard Neustadt has more linked text than plain text, and to me looks like a Las Vegas billboard. Too many dates are wikified. It looks very ugly! Does anyone else agree? -- Viajero 14:33, 3 Nov 2003 (UTC)

Maybe you should create stubs for all the red links, then it will be uniformely blue. :-) andy 15:37, 3 Nov 2003 (UTC)

Oh no! Given the subject, all that red and blue on a white background looks quite patriotic :) -- Finlay McWalter 15:56, 3 Nov 2003 (UTC)

I always disliked all the years in Wikipedia being linked. Tempshill 23:06, 3 Nov 2003 (UTC)

Wikipedia talk:Make only links relevant to the context is not policy, and it is rude to revert other people's linking. RickK 01:56, 4 Nov 2003 (UTC)

Can you please clarify: does this refer only to links authors put in their articles or also to links others also add to yours? I mean, there are a lot of people who go through articles here doing nothing but adding links. -- Viajero 14:27, 4 Nov 2003 (UTC)

If the Richard Neustadt article said something more about just what he accomplished with all those years in all those jobs, the link density would be a lot lower. He was presumably a historian in the Navy - did he work with Samuel Eliot Morison perhaps? What kind of advice did he give to all those presidents? This article is a good start, but it's completely colorless (figuratively speaking :-) ). Stan 07:28, 4 Nov 2003 (UTC)

I have become gradually converted to more extensive linking. I now link all dates, all proper nouns, and all abstract nouns. I don't link common nouns unless there is some particular reason to do so. Adam

I wish someone would explain to me the rationale behind linking to dates. Personally, I have never clicked on a day, a year, or a decade in the course of reading an article. Is there something wrong with me? ;-) Theoretically, I can imagine that there might be some use for seeing what links to a given date, but if you click on What links here for 1969 for example, you get an undifferentiated list of five hundred article titles. What possible use can this have? One can also search Wikipedia for 1969. What is the difference?

That being said, I find the guidelines suggested here reasonably sensible: just link birth, death, and epoch-defining events. Don't put in links for subsidiary information that doesn't exist, like book titles or geographical dependencies. But when one refers other users to this page, one hears (like above) that the page is "controversial" and "is not policy". Heh?

Even more so than bold and italic, wikilink is a form of emphasis; the way the software is written, a link is not hidden until you pass the mouse over it but rather it jumps right out the page, and I am entirely unconvinced by people who argue that you can simply ignore them. Sorry, you can't; they distract the eye. Hence I believe Wikilinks should be used with discretion, and only for articles related that bear some kind of high-level relation, and certainly not for common English nouns like water or gun. In visual terms, Richard Neustadt looks an embarassingly amateuristic, a travesty of good taste. Are we slaves to the software and stick links everywhere because we can, or do we use our intelligence and pick and choose carefully? -- Viajero 14:27, 4 Nov 2003 (UTC)

I agree with Viajero's rant. When I write, I avoid linking phrases unless I myself see a connection. If it's an article about one country's relation to another, well, then that's the connection! Iraq was invaded by a coalition led by the United States. (If you click on the "What links here" command at the United States page, you'll see every article containing a [[ ]]-style link to the US.

Also, if I'm using jargon as in scientific articles, I think it helps the reader to be able to click on unfamiliar terms like "global warming hypothesis" (a nice, short explanation that they can read in a minute or two) when reading about the politics and economics of the Kyoto Protocol. It would NOT help if Fuzheado deleted the GW hypothesis page, because then the link from the medium sized Kyoto Protocal article would go to the super-large, poorly-organized global warming article. --Uncle Ed 14:45, 4 Nov 2003 (UTC)

While links are primarily aids to the reader, they are also aids to maintenance; for instance, backlinks often find orphaned or misconnected articles (JFK was elected in 1960, not 196), and we use them to find refs that need disambiguation. Also, you can't assume readers know as much as you do, and you can't assume that searching works usefully; while I know about the fine points of Captain in a naval context, I want uncertain readers to look at the article rather than make assumptions, and the blue lighting is a hint. Lots of free links in articles like Richard Neustadt is also a sign that we're still missing necessary articles. BTW, Welfleet is a red link because it's misspelled and mislocated - it should be Wellfleet, Massachusetts (Neustadt had homes on each side of the Atlantic) - yep, links are a good way to detect misspellings... :-) Stan 00:32, 5 Nov 2003 (UTC)

I have to agree with Viajero. This article looks a mess as it stands and most of the links told me nothing I needed to know in order to increase my understanding of the article. And it's far from being the only article I've seen that's in this state. Readability must be the prime aim, surely? And all that coloured and underlined text certainly does not aid readability; quite the opposite. Bmills 11:08, 5 Nov 2003 (UTC)

Stan, In your new article Airmail, I notice you have red links for rocket mail, missile mail, space mail. These are "missing necessary articles" ?!?!?! As for pigeon mail, balloon mail, dirigible mail, and zeppelin mail, why not simply expand them in Airmail? -- Viajero 12:51, 5 Nov 2003 (UTC)

And Astrophilately is clearly explained by the sentence in which it occurs. Bmills 13:26, 5 Nov 2003 (UTC)

Heh, this may be hard to believe, but each of those subjects is a recognized collecting specialty, with multiple books in the philatelic literature. One of my rules of thumb is that any subject with a scholarly book dedicated to it will easily justify a 1-2 page article of its own. For instance, the astrophilately article, which I expect will be the smallest of the bunch, still needs to say something about which flights carried what, postmarks on the various space stations, process, the political fights, etc, which would be too long and complicated for the generic airmail article. I don't understand the idea of not making links to topics just because one doesn't know anything about them; one of the things I enjoy about WP is discovering that an unlikely-looking link in an article leads into a whole field of whose existence I was previously unaware. Stan 14:28, 5 Nov 2003 (UTC)

Stan, clearly your Airmail article was not a good example of overlinking. You added those links with deliberation, and it sounds like you will probably write the corresponding articles in due time. That's fine. I enjoy learning new things too; now I know what missile mail is! (And hopefully will learn even more in a short while!)

For me, the issue remains the unreflective linking of all dates and proper nouns, and even common nouns and verbs in an article, irrespective of whether there is any high-level conceptual relation (there are some examples mentioned in the preceding discussion; I can think of more). I realize you consider the latter a "subjective" judgement, but it is no more subjective than the process of deciding what go into an article. The problem is, taken to its logical extreme, linking everything to everything results in articles where there is more linked text than plain text, like in Richard Neustadt, which I think is (typo)graphic embarrassment; it is anti-reader. It is linking simply for linking sake, without logical basis, and it makes this enterprise look amateurish.

At this point, we can one of two things:

We can attempt to rationalize linking and expand upon the guidelines listed on this meta page, and try to promote a more conservative approach to linking. Or
We say, ok, link everything to everything. But then let's turn off the colored links on the article pages and use CSS mouseovers to highlight links only when the mouse cursor passes over them. Plain black-on-white text again! Maybe generate a box (like the TOC) at the bottom of the article listing the articles the text is linked to. Yes, I realize displayed links are helpful when composing an article, so we display them in the Preview mode.

Such a scheme would be marginally less writer-friendly but so much more reader-friendly. I expect it would even increase the quality of articles too, as writers would be less hypnotized by the colored links. -- Viajero 19:32, 6 Nov 2003 (UTC)

I would say that there seem to be two types of links - there are links indicating a relation to another article on a similar subject, or on which an article is dependant; and then there are also links to other articles which are incidentally referred to (but which a reader may occasionally want to refer to). I would suggest that the most user-friendly way to deal with this would be the implementation of different markup for important and incidental links... maybe important links are blue/red and underlined, while incidental links are only highlighted on mouseover. While I realise that this would probably take effort to implement, it would make pages look considerably cleaner to have only important links highlighted. Maybe (assuming infinite time available to the programmers), a user could have a cookie-retained choice to remove incidental links entirely, or to have all (or no) links highlighted. 84.92.118.165 15:17, 12 August 2005 (UTC)

Add my vote to the side of conservative links that pay attention to the context. Excessive linking is just plain hard to read. The arguments for mass-linking do not convince me. If I want to find every reference to a topic (like a date), I can use the search function at the top of the page. We are trying to create an encyclopedia that is attractive and easy to use. Too many links make it harder - just as too few links. Authors and editors have a responsibility to make judgements. Rossami 16:02, 6 Nov 2003 (UTC)

Searching really doesn't work that well; just try finding all the references to 142 AD by searching, and beware, Google searching does not reflect the current state of the database accurately. Stan 17:26, 6 Nov 2003 (UTC)

I just tried this. 142 displays two linked pages; searching Google with 142 AD returns eight hits, of which half spurious, with "142 AD" OR "AD 142" returns just one. Hence, I am not convinced that Google is necessarily worse. As for reflecting the current state of the database, Google seems to spider WP daily, meaning only the most very recent changes don't show up. -- Viajero 18:42, 6 Nov 2003 (UTC)

Try "142" without the AD; articles discussing events of 142 don't add "AD" everywhere. There are lots of other examples where searching doesn't work well, that's just a really obvious one. Also, I don't think Google spiders the whole thing daily, it regularly shows me material I edited out weeks ago. Stan 19:51, 6 Nov 2003 (UTC)

Just tried it. Test results below:

142 has 22 entries in "What Links Here". 12 are other date-recap pages and two are these pages discussing the example. By my count, that leaves eight relevant pages. Google search finds 193 entries, most of which have nothing to do with the year 142. I guess I can see your point for references to years because you get so many false positives especially for low years AD.
A quick test on a random date found the following: (I picked February 29 so the count would be managable.) What links here = 80, many of which are other date-recap pages. Google search "February 29" = 39 hits but they all had high relevance.
Testing a non-date: Henry Ford. What links here = 63. Google search = 209 but relevance drops off at about 50.

My conclusion: The facts support Stan. My alternative of google-searching is not a panacea. But I still think that excessive links make the article harder to read and are not worth the effort. Rossami 21:33, 6 Nov 2003 (UTC)

All date links should be Wikified. I want to be able to have instant reference to the historical context of any event deemed worthy of recording the date of. Otherwise, why are you bothering to put the date that something occurred? It doesn't make sense NOT to Wikify all dates.--64.254.131.70 14:41, 27 Apr 2005 (UTC)

As a side-note, I am aggressively deleting redundant links (two or more links to the same place within a single article). Even granting the most liberal of the arguments above, they serve no purpose. They just clutter up the page and make future maintenance much harder. Rossami 16:02, 6 Nov 2003 (UTC)

That's the usual rule - only link a first appearance. Stan 17:26, 6 Nov 2003 (UTC)

I disagree. If an article is five screens long, I don't want to have to search backwards through the text to find the first instance of a term if I find myself interested in it later in the reading. First occurence within a subsection is a more reasonable standard.--64.254.131.70 14:41, 27 Apr 2005 (UTC)

more from village pump

A recent edit of Statue of Liberty has a section that now reads:

French sculptor Frederic Auguste Bartholdi was commissioned to design a sculpture with the year 1876 in mind for completion, to commemorate the centennial of the American Declaration of Independence. It was agreed upon that in a joint effort the American people were to build the pedestal, and the French people were responsible for the Statue and its assembly here in the United States. However, lack of funds was a problem on both sides of the Atlantic Ocean. In France, public fees, ...

The edit added links to French, American, French, Atlantic Ocean, and France.

I see a lot of this, and I don't understand it. What purpose is served by linking the word "French" twice, rather than just on its first appearance in the paragraph? And, what purpose is served by linking to these terms in the first place? There seems to be a frenzy in Wikipedia to wikilink anything at all for which an article exists. Isn't it preferable to link only those that are clearly relevant to the article from which they're linked? Dpbsmith 04:24, 1 Feb 2004 (UTC)

I agree. I often remove links where they bear absolutely no relevance to the article. Particularly dates, where the date has no real significance. For example in the article on Homer Simpson the sentence

Various episodes have given his date of birth as May 12 1956, May 10 1955, and even 17-23-1956 [sic].

contained 4 links to May 12, 1956, May 10, and 1955 before I removed them! HappyDog 04:45, 1 Feb 2004 (UTC)

But this is a hard call. The original idea of hypertext was that every word would be a link. It seems clear that this is not the way to go. But one of the really wonderful things about wikipedia is the robust links between pages. Let's not get too strict about removing links.... RayKiddy 05:29, 1 Feb 2004 (UTC)

See the debate: make only links relevant to the context vs. build the web. I prefer the former, since it makes it easier to read, and when only a few words are linked, it draws your attention to the more important words. --Minesweeper 05:58, 1 Feb 2004 (UTC)

At Wikipedia: Manual of style#Free links we find

...The use of so-called "free links" to other topics, for example, George W. Bush, is encouraged. Use the links for all words and terms that appear in your article for which it could be worthwhile to read the linked article. However, don't overdo it. Do not link every occurrence of a word; simply linking the first time the word appears will usually be enough.

which I think is about right. It's not always an easy call, but for instance linking to the number three from triangle is I think good (and I must do it), while linking to the number six from The six o'clock news would be quite wrong, linking to the number three from The three bears is IMO a line call, it depends on whether there is something important about the number three that makes it particularly apt for this story.

Helpful links are those which either a reader is going to be looking for, or isn't going to be looking for but will recognise as helpful and say "Aha!" when they see them. But deciding this isn't always easy.

Linking to the same article twice, or to articles of no likely relevance, just clutters the page and makes it harder to see the useful links.

I probably tend to err towards overlinking as the lesser of two evils. Andrewa 06:12, 1 Feb 2004 (UTC)

We should link all dates, places, names, etc. The more links the better, as long as it doesn't hinder the reading experience. It's WWW, not paper. Optim 13:39, 1 Feb 2004 (UTC)

Well, this is precisely the point: excessive links do hinder the eye, as you demonstrated above. Tell me: what exactly is the point of linking every date: how often have you followed a date link? At what is the point of linking every occurence of New York City??? -- Viajero 16:48, 1 Feb 2004 (UTC)

Some ideas:

It seems to me that there may be different ways of working here. These may reflect our Myers-Briggs types or something similar. So, some of us may see the clutter as objectionable, and others as not. Some may see trivial links as helpful, and some not. And, we should expect our readers to be similarly varied.

Is there any reason that Wikipedia can't implement the original idea of hypertext, that every word is a link without requiring any special syntax on the part of the author? I think we would still want ways of piping and highlighting explicit links, and I don't know how it would affect "what links here", whether we'd make that highlighted and/or piped links only, or whether we'd want to have some parameters so you could choose how it worked.

And, I don't know how or whether it would impact our database design.

What I'm trying to do here is have my cake and eat it too, to have the best of both worlds by both maximising links and minimising clutter. Anyway, food for thought? Andrewa 02:58, 2 Feb 2004 (UTC)

There has already been quite a bit of discussion on why we can't link every word. Some technological and some based on our assumptions about the ideal reader experience. You might be interested to read the discussion above and the prior discussion that has been moved to the archive page. Rossami 17:22, 2 Feb 2004 (UTC)

Just for the record, I did read the above discussion before posting. I guess many don't and I accept that your comment is meant to be helpful.

I've now read this archive too. Thank you! There is so much archived discussion it's impossible to read it all, often I can't even find my own previous contributions when I try! This archive does deal with the reader experience, but it seems to me that despite this many still assume that every reader is going to have the same experience (and generally, will think the same as they do, despite the overwhelming evidence that we are a rather varied lot).

I can't see where the technical problem is discussed, and I'm very interested, could you be more specific? Andrewa 18:42, 3 Feb 2004 (UTC)

Wading through the old conversation threads is tough. Even knowing what I was looking for, I had to read through 3 times to find it. I was referring to the difficulties with what was called "auto-linking" above and in the archive. Basically that there is no way to link every word without missing all the links that depend on multiple words. (Others said it better than I am doing.) Rossami 00:01, 4 Feb 2004 (UTC)

I support having a choice: let the [[]] code give non-underlined links, and add a new code for underlined links, to be used when the writer wants to draw attention to a related topic/definition/name/etc. Exploding Boy 11:00, Feb 2, 2004 (UTC)

Worth a try. This does seem to offer the best of both worlds. I'm fascinated at the suggestion above that we can't make every word a non-highlighted link for technical reasons. I'm guessing these may be capacity/performance considerations, or it may just be that the proposal was misunderstood. Andrewa 18:42, 3 Feb 2004 (UTC)

I propose a system of formatting the links so that there are different level views of them. Perhaps three or four levels ranging from level one which is an obvious underlined link to level 3 or 4 which would be completly invisible unless a mouse over event occured on the link. The middle level could be blue (or whatever colour is logical) high lighted text but not underlined. This way people who prefer lots of linking could delegate the less important links to a lower priority which would facilitate ease of reading and allow a more rich tapestry of links for those who prefer more linking.--Mikeroodeus 04:03, 22 Feb 2004 (UTC)

This sounds like a very good system to me. In addition, logged on users could have something in their preferences to remove all links below a certain level. Combined with user stylesheets, this would allow everyone to see the wiki in whatever manner best suits them. 84.92.118.165 13:43, 16 August 2005 (UTC)

Overwikification

I apologize if I'm raising this issue in the wrong forum or if it's already been discussed. But I've noticed several examples of what I'd call overwikification. Check out these articles: List of Ambassadors and High Commissioners to Canada, Survivor: Pulau Tiga, Tom Berenger, Twelfth United States Congress. There's over two hundred red links to empty articles on these four pages alone. And looking at the titles of these empties, it's likely nobody is ever going to fill them in with any article, let alone one worth reading.

Maybe we need to discuss some informal standard an empty article needs to rise to before someone creates a red link to it. MK 22:30, 6 Mar 2004 (UTC)

I can't really see the harm. It's possible that they may just inspire an interested party to write a series of articles for them - it's certainly happened before. Ambivalenthysteria 22:53, 6 Mar 2004 (UTC)

I'm with MK. I find it prevalent and an annoying irritation. I suspect the practice of wikifiying almost everything regardless of whether the linked article exists was more useful when Wikipedia was getting started. I don't believe a word should be linked unless you either a) know that an article for the linked term exists or b) feel strongly that the article should exist and have a reasonable expectation that someday it will exist. Dpbsmith 22:59, 6 Mar 2004 (UTC)

Well, I can't speak for everybody, but I "strongly feel" that an article for every member of the U.S. House (past and present) should exist, along with an article on every movie in IMDB and every band at Ultimate Band List. I've got no problem with these articles. Meelar 23:02, 6 Mar 2004 (UTC)

I agree that in narrative text, overwikification is very annoying. But I'd draw a distinction between narrative text and lists: in lists of films, books, credits, etc, like those pointed to by MK, it's less annoying to have them all wikified. And many of them will come to have articles eventually: several of those in Tom Berenger's list, for example have three or more links to them already. - Nunh-huh 23:07, 6 Mar 2004 (UTC)

What Nunh-huh said. On lists like 12th Congress, all of those names eventually (sometime in the next decade) should get articles of their own. In actual articles overwikification is a huge problem, especially with dates. But in lists is cool...

What? you want us to re-wikify them all after they exist, which means once the links are available, we have to search them all in wiki and wikify them? I guess that's more annoying! Besides, i have to check out all the links in my article to see if they exist before i submit it. which one is more annoying? i strongly believe that a red link gives Wikipedians a desire to creat it. What we need to worry about, is the vandalism, IMO. PS: I DON'T like repeated wikified links in one article... --Yacht 12:17, Mar 7, 2004 (UTC)

So people are claiming that there actually will someday be an article on Philémon Yunji Yang, the current Cameroonian ambassador to Canada? A solid biography of Mr Yang (I'm assuming Phil's a mister) and a brief but thorough history of his diplomatic and political career? Hopefully it'll appear before someone else has suceeded him as Ambassador because an article on the former Cameroonian ambassador to Canada seems pointless.

If we're going to have a standard that any subject potentially deserves an article because someone may be interested, why do we discourage vanity pages and high school articles?

My personal suggestion is that when you're wikifying an article, ask yourself these questions before creating a red link: Should this article exist? Would there be information in this theoretical article that goes beyond what's in this existing article? Is it ever likely to be written? Would someone be likely to be looking for an article with this subject?MK 16:26, 7 Mar 2004 (UTC)

Very amusing. Phil's been taken care of. Which is one down; one hundred and thirty three to go. Which will complete one page of the four I mentioned. Which were only four random examples of the hundreds of similar pages that surely exist. Starting to get an idea of the size of the job?MK 15:04, 8 Mar 2004 (UTC)

Yes, MK. The job sounds massive--almost as though someone was trying to build the world's largest free encyclopedia using only volunteer editors. :-) Seriously, MK, it doesn't sound any more daunting than, well, doing what we've been doing for the last several years. Red links mean we add more articles. And frankly, there is a difference between a vanity page for "the owner of Mountlake Terrace's largest 7-Eleven" and an article describing, in brief, the diplomatic actions of Cameroon's representative in Canada. Are you opposed to an article on the United States' ambassador to the United Kingdom? If so, I wonder who does qualify for articles. If not, then why should we look down on ambassadors from poor countries/Third World countries/countries most Wikipedians have not traveled to? I think there is a bias here. After all, I'm sure I could find a page full of red links to people MK thinks should have articles, and I can't imagine you want those pages dewikified, else you'll never know the articles need writing. The problem you are asking to have solved is not a problem, in my opinion. Jwrosenzweig 17:49, 8 Mar 2004 (UTC)

I guess there's little point in arguing against the consensus. Personally I think that Anglo-American relations are more objectively important than Cameroonian-Canadian relations and I think this is a reflection of reality not bias. If everything is equally relevant, than why say that an ambassador is more important than a store owner?

Let me say again, I have no objection to people writing these pages. The problem is people not writing these pages. But pages are still be creating, literally by the thousands, that nobody will ever write. Do we really want to reach a point where the majority of this project is empty boxes? In my opinion this would make a mockery of what we're doing here. MK 07:23, 9 Mar 2004 (UTC)

To me this seems to be 2 slightly different arguments going on here, firstly what deserves an article, and secondly that red links are bad. The first argument I am going to steer well away from. The second one however, seems easy, (philosophically at least). If there are a reasonable amount of people who find the missing links annoying we should add a user config option to not display incomplete links, (yes I know this isn't the place to suggest software mods but I was just trying to help solve a disagreement, if others agree I could add the request on sourceforge). In my suggestion there would be 3 choices for the display of bad links, red, question mark or plain text. Steven jones 02:50, 13 Mar 2004 (UTC)

Self links

How about adding the following sentence to the policy:

Do not leave any links from the article to itself, including through redirects, since this annoys readers considerably.

Any objections? Gadykozma 09:07, 12 Sep 2004 (UTC)

That's already the policy (though it is not spelled out on this page - I can't remember which page it is on). Feel free to add it if you think it would be helpful here. Rossami 13:30, 13 Sep 2004 (UTC)

It's kind of obvious, and people don't do it intentionally anyway. Gadykozma 13:43, 13 Sep 2004 (UTC)

Fabricated stories v. reality

There are many articles that deal with imaginary subjects: SF, TV, movies, urban legends ... . I will try to over-link these articles. This possibly helps people to go back to the real world. It's usually next to impossible for a SF movie to link to another one, unless filmmakers wanted to do so. But all fictional subjects link to the shared human knowledge base. If we over-link, some very important articles will be more oftenly visited. -- Toytoy 03:06, Sep 22, 2004 (UTC)

Trying to subvert the author's intent with excessive linkage is a losing stategy: when everything is linked to none of the links will ever be followed. On the other hand you can, for example, go around adding paragraphs discussing the social/political/economic/philosophic premises underlying a science-fictional scenario, and judiciously link to those subjects. -- Doom 19:16, Oct 7, 2004 (UTC)

Text deleted from Wikipedia:Link editing and Wikipedia:Links to minor subjects

These pages (which have not been discussed for quite a while) seem to discuss issues similar to those discussed here. I have therefore made them redirects to this page.

Text deleted from Wikipedia:Link editing

Wikipedia has extensive guidelines for the writing of articles. However, it has few for the choice and editing of links. As a result, some articles are under-linked, others over-linked, there are articles created earlier which lack links to those created later, and external links often appear by surprise in the middle of articles. This is a draft policy for editing links, not articles.

One guide to areas of coherent discourse that should heavily inter-link is the Main Page. One expects psychology or mathematics or history articles to link densely to other articles on the same topic. If one sees a lot of links to psychology in a history article, or to history in a mathematics article, however, it may be time to figure out which are really essential or neutral. One mistake to avoid is censoring out-links to other topic areas - this is this is an encyclopedia, not a textbook so one expects philosophy of mathematics and philosophy of science and philosophy of physics to be well integrated with philosophy in general, even if mathematicians, physicists or other scientists might prefer it not be.

Another guide to areas that are coherent is the Wikipedia:WikiProject list. These are closely related article clusters on things like voting systems, places in London, or all the world's ecoregions. As choice of a voting system depends on its comparison to others, a place in London can only be reached or understood in relation to others around it, and ecoregions likewise can only be understood as unique in relation to others on their borders or elsewhere in the world, one expects dense linking and detailed comparison. However, in some cases, there may be special articles just on comparisons or categories, in which case, those overviews are the place to discuss and compare.

A third guide would be the Markov chain of links that users actually follow - showing what they perceive as coherent subjects or topic areas, and which are the most common pathways from one article to another. The most heavily travelled link paths can be shortened by including clarifying texts and links to the articles most commonly reached in two or three links while browsing - thus making the Wikipedia a far more convenient place to get quick briefings.

Sadly, this data is not now being made available to contributors or editors, and this should probably be a priority for some future software release. There is probably no other way to become user-driven with regard to the way that the encyclopedia puts paths in front of its users. A forest or garden path is beaten flat and thus the most commonly taken route s visible. Wikipedia by contrast is more of a jungle that quickly overgrows.

Opinion: Please remember that every link makes the text slightly more difficult to read, and may encourage some readers to divert from the topic. Fewer links will invest greater importance in the remaining links. In general, I think that Wikipedia articles are overlinked, in particular with general links that can be of no immediate use to the reader in understanding the topic. Whoever started linking broad chronological items on Wikipedia, e.g., 17th century, 1990s, 2001, deserves to be horse-whipped; it's now a scourge. Tony 23:57, 1 September 2005 (UTC)

Text deleted from Wikipedia:Links to minor subjects

This page is not official at this time.

Wikipedia contains many articles on subjects that are relatively unimportant, and this is fine, because Wikipedia is not paper, and the articles can all fit.

However, what irritates some Wikipedians more than such articles themselves are links to the articles that make the minor subject seem more important than it is. When creating an article about minor subjects, a good policy is to avoid the temptation to link to it from pages where it is not particularly relevant or important. Instead, limit yourself to creating links from pages where the subject really is highly relevant.