Wikipedia talk:Manual of Style/Linking/Archive 12

This is an archive of past discussions. 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 5

←

Archive 10

Repeated links

'Repeated links' now says "where the links are in a table or in a list, as each row should stand on its own" seemingly justifying an indefinite number of repeats arranged in lines down the page. Can we make a common sense amendment to this sentence? I propose:

'"where the links are in a table, the table should stand on its own, relinking words that may have already appeared in body text."

Any comments? --Klein zach 01:59, 19 April 2011 (UTC)

Good suggestion. Tables tend to be grossly overlinked, because this guideline seemingly allows it. --Ohconfucius ^¡digame! 02:04, 19 April 2011 (UTC)

I support this. I had the same issue above, as a user was readding links to every row even when you can see all of the rows within a standard resolution screen. BIGNOLE (Contact me) 04:11, 19 April 2011 (UTC)

I support anything that stops (or discourages) a link being repeated in nearby rows in a table. The trouble is that I'm not getting that sense from the suggested wording. "...relinking words that may have already appeared in the body text" doesn't seem to achieve what we are all thinking here; or does it, and I'm not understanding it? Doesn't that wording still give a green light to anyone wanting to repeat the same link in every row? GFHandel. 04:33, 19 April 2011 (UTC)

I think there is a critical difference between a table standing on its own as opposed to each row of a table standing on its own. I am unsure as to what "relinking words that may have already appeared in body text" means, but would guess that is an exemption to link words that have already been linked to in the preceding text. --Ohconfucius ^¡digame! 04:40, 19 April 2011 (UTC)

Sorry, my wording was less than ideal. I did mean "an exemption to link words that have already been linked to in the preceding text". Perhaps I should quote the whole text with (my suggested) part in bold:

"In general, link only the first occurrence of an item. There are exceptions to this guideline, including these:

where the later occurrence is a long way from the first.
where the first link is in an infobox, navbox or similar meta-content.
where the links are in a table or in a list, as each table or list should stand on its own with its own independent set of links.

How about that? --Klein zach 06:54, 19 April 2011 (UTC)

IIRC, the reasoning for the current wording, giving the exception as line-by-line, not the table as a whole, is that in sortable tables one cannot ensure which line is going to appear first, and so each line being fully self-contained ensures the reader doesn't have to hunt through the entire table to find the relevant link. For short tables, it's not that big a deal, but some tables are quite long, and the links actually serve the reader. I think we should keep the current wording; it exists for a reason. oknazevad (talk) 18:13, 19 April 2011 (UTC)

what about incliuding mention that the exception is for sortable tables, and not non-sortable tables? BIGNOLE (Contact me) 18:53, 19 April 2011 (UTC)

"...it exists for a reason" seems like a reason why nothing would ever get debated/changed, so instead: have a look at this table. The Venue column is sortable, however it was long ago decided not to repeat the links to Queen's Theatre, King's Theatre, etc. in every row. (and there has never been a localized problem with that decision). Tables don't retain a user-preference of sort order, so the table is always initially in a sort order that makes the decision to link the first occurrence sensible. The idea being that when the user clicks a different sort order (e.g. Libretto in the example table), then all bets are off, and the user shouldn't be too shocked that the links are no longer in the original order—and it is expected that someone using that table would be knowledgeable enough to restore the original sort order (with the HWV column) if required. Please remember that there is no perfect solution for repeated links in tables, but I feel that linking only the first occurrence (even in a sortable table) is generally more desirable than the dilution of more valuable links that happens when every occurrence is linked. GFHandel. 23:09, 19 April 2011 (UTC)

The table that prompted me to start this is here: List of operas by Auber. As you will see Opéra-Comique and Eugène Scribe are each linked 30 or 40 times in this table. This also illustrates GFHandel's point. I don't think the reader of these (relatively complex) tables is so naive that he or she doesn't realize that links are also moved by resorting. Most readers will not re-sort. Those that do, will be doing so for a purpose. --Klein zach 00:38, 20 April 2011 (UTC)

My 2c. While I object to repeated links in text, I don't have a problem with repeated links in tables, or, indeed, plain lists. Period. Wwwhatsup (talk) 03:29, 20 April 2011 (UTC)

Done I've made the change as we have a reasonable consensus for change. Please say if the phrasing needs further tweaking. Best. --Klein zach 07:19, 21 April 2011 (UTC)

I've just added rfc tag, as I don't think this was debated widely enough. --Rob Sinden (talk) 09:15, 7 June 2011 (UTC)

I have added a fourth exception - where two references map onto the same article via a redirect - see my changes to Zara Phillips. Martinvl (talk) 09:08, 14 June 2011 (UTC)

I cannot parse what you added; your example above also seems inconsistent with what you did at Zara Phillips. I'm reverting for now pending further clarification. --Ohconfucius ^¡digame! 10:30, 14 June 2011 (UTC)

The current text for three consecutive paragraphs in the lede for Zara Phillips now contains:

... her niece Savannah Phillips in 2010, she is ....

...[paragraph not relevant]...

She has an older brother, Peter Phillips, ...

The original link to Savannah Phillips had been redirected to her father, Peter Phillips. This morning somebody removed the link from Savannah's name. I reinstated the link even though both end up at the same page on grounds that the reader would not automatically connect Savannah Phillips to Peter Phillips. Does this clarify things? Martinvl (talk) 12:09, 14 June 2011 (UTC)

I'm not stupid, but I still don't get it. How will other people understand this addition? At the very least, could you write out here something a bit easier to understand, as a draft for your addition? Tony (talk) 13:12, 14 June 2011 (UTC)

How about the following text to describe a permitted repeated link:

"If two articles that are apparently unrelated have been merged into a single article using WP:REDIRECT (for example the biography of a young child has been merged into the biography of its parent) and both articles are referenced from a third article, then both articles should be Wikilinked in that article, especially if the context does not indicate a connection."

Martinvl (talk) 14:15, 14 June 2011 (UTC)

Needs clarification

At Wikipedia:Manual of Style (linking)#What generally should not be linked it says in part "Avoid linking the names of major geographic features and locations,". What is meant by major in this case? Is it used to indicate importance or size? CambridgeBayWeather (talk) 23:40, 30 May 2011 (UTC)

If the only choices are importance or size, importance is clearly better because the Earth's mantle has much more size than London. But we don't have more specific rules on that issue. Art LaPella (talk) 04:06, 31 May 2011 (UTC)

I would take it to mean the terms "mountain" or "ocean" or "doctor" are not linked. A normal English reader should understand the term. But I wouldn't have excluded Mount Everest, England, Islam, nor Arabic, even though all are well known. I would think we would want them linked the first time in an article like everything else. Student7 (talk) 02:02, 2 June 2011 (UTC)

I think "Avoid linking the names of major geographic features and locations, religions, languages, and common professions" pretty clearly includes at least England, Islam and Arabic if it includes anything, so does that mean you want to change the guideline? Art LaPella (talk) 03:06, 2 June 2011 (UTC)

I think I would favor that change, yes. I think the guideline went a bit far. "Geographic features" and "common professions" makes a lot of sense. I think specifically named places, religions, and languages are not down to that level IMO. Even IF considered "common" by many of us. There are some people for who this is not common. Others who would appreciate a link to the topic. Student7 (talk) 14:17, 3 June 2011 (UTC)

It is well-established on en.WP that the names of commonly known countries and cities are not linked, particularly English-speaking ones. There are almost always more important links to preserve in the vicinity without the dilutionary effect of such overlinking. We have had this discussion many times, most recently at WT:MOS last week; please let's not revisit it. Tony (talk) 16:41, 3 June 2011 (UTC)

@Student7, I never really considered it referring to things like mountain, ocean or doctor. @Tony1, I'm not looking to change the guideline but for clarity. If the word major is used for size then editors would not link Victoria Island (Canada) but they would link Cyprus. If on the other hand major means importance or commonly known the reverse of the previous sentence would be true. The whole sentence should be cleared up by rewording it. One possible suggestion is "Avoid linking the names of common or well known geographic features and locations, religions, languages, and professions". CambridgeBayWeather (talk) 15:22, 4 June 2011 (UTC)

That wording is according to my interpretation of the guideline, and would have no problem with it. --Ohconfucius ^¡digame! 16:51, 4 June 2011 (UTC)

Yeah, seems ok. Could you make it "well-known"? Tony (talk) 03:16, 5 June 2011 (UTC)

You mean as in "Avoid linking the names of well known geographic features and locations, religions, languages, and professions"? CambridgeBayWeather (talk) 10:07, 5 June 2011 (UTC)

He means the hyphen: well-known not well known, see WP:HYPHEN. (I would have changed that as a typo, anyway.) Art LaPella (talk) 16:17, 5 June 2011 (UTC)

I assume that this solves one problem that satisfies everyone.

As for the other part, Proper Names that are commonly known to US residents and sometimes Europeans, are not always recognizable to someone whose first language is not English or who does not have a European background. These are not uneducated readers; they just haven't taken American Geography or watch American television. This is why I prefer to link place-names when first encountered. I am not pushing for a change here since I have really had no problem with that, except with myself when I first noticed rather recently, a proscription not to link stuff in the lead, when avoidable. (I have found it hard to avoid, to tell the truth). Student7 (talk) 18:20, 8 June 2011 (UTC)

What Student7 say. Cardiff might be well known in the UK but I think a sizeable fraction of Texas couldn't tell where it is off the top of their head, and vice versa for Austin. On the other hand, the names a few really major cities such as London or Paris are actually quite unlikely to be unfamiliar to a reader, wherever they are from. Maybe “well known geographic features ... and professions” should be replaces by “geographic features ... and professions which are well known worldwide”? ― A. di M._plé^dréachtaí 22:44, 30 July 2011 (UTC)

RFC: restructuring of the Manual of Style

Editors may be interested in this RFC, along with the discussion of its implementation:

Should all subsidiary pages of the Manual of Style be made subpages of WP:MOS?

It's big; and it promises huge improvements. Great if everyone can be involved. Noetica^Tea? 00:43, 25 June 2011 (UTC)

Overlinking of magazines or publishers

An IP-hopping editor has made a habit of linking The New York Times and Time (magazine) (not [[Time (magazine)}|]]), and professional organizations, such as the American Geophysical Union, when it appears in a citation template in a reference. Other editors find it a distracting sea of blue. Any comments? I, personally, think the links unnecessary, but not distracting. Any comments whether something needs to be done, and how to phrase it in WP:OVERLINK? — Arthur Rubin (talk) 15:13, 7 July 2011 (UTC)

If there is a specific element within the source document that is cited/quoted directly (such as a quote from an editorial, as opposed to a third party spokesman), the journal's name would appear in the running text of the article; it will most likely be linked at its first appearance. In most other cases, sources are fungible (ie it wouldn't make much difference if we substituted another source) and do not merit linking. Names of publishers are rarely necessary in everyday news sources; I would expect the reader to know about 'publishers' such as BBC, CBC, and for readers to look them up separately only if they were curious. Many other 'works' themselves are often better-known than their publishers. --Ohconfucius ^¡digame! 15:20, 7 July 2011 (UTC)

Good question. I have linked lesser-known publications so that interested editors may check out the source to determine reliability. For a large article, I can appreciate that overlinking happens for the better-known sources. I do not expect that everyone reading these articles has heard of sources that many of the rest of us are familiar with.

I am more annoyed with someone sticking in a source that "sounds like" a reputable organization but merely has a well-chosen name for what is essentially and unedited blog (and no article in Wikipedia). Student7 (talk) 18:35, 9 July 2011 (UTC)

A similar case in January 2010 inspired the lengthy thread Wikipedia talk:Citing sources/Archive 27#Links within citation. There was no consensus on whether these links are a good thing. Adrian J. Hunter^{(talk•contribs)} 07:35, 29 July 2011 (UTC)

Avoid linking plain English words

This is ambiguous because it doesn't specify 'link to' or 'link from'. A link from a common word such as "nun" may be linking to an uncommon word such as "Dorothy Kazel". QuentinUK (talk) 13:43, 28 July 2011 (UTC)

Ah, that would be piping from a common word to what one hopes is a high-value link. This is called "Easter egg" linking, and we try to minimise it; occasionally, it's inescapable. Tony (talk) 14:30, 28 July 2011 (UTC)

Sometimes I signal that by including the grammatical article in the link caption: The equation describing the ... is named after him, where it's clear I'm not linking to equation because in that case I'd keep the The out of the link. But don't do that unless it's obvious from the context where the link goes. (@Tony: Can you give an example of when “it's inescapable”? When all else fails you can recur to parentheses, as in a nun (Dorothy Kazel).) ― A. di M._plé^dréachtaí 22:14, 30 July 2011 (UTC)

Linking to an image page

The article has a section on "Links to Wikipedia's categories" that explains prefixing the link with a colon so that the Cat will appear as a link and not get interpreted as a command to add a Cat to the bottom of the page. I noticed that something similar happens with links to images. If one places a wl to an image, it gets interpreted as a command to display the image. On the other hand, if one wishes to include a link to the image – say, as part of a Talk page discussion – then the colon does that. I think the article should make this clear (unless it does and I missed it). Either rename the section "Links to Wikipedia's categories and images" or create a new section "Links to Wikipedia's images" with instructions like in the "categories" section.—Biosketch (talk) 08:27, 31 July 2011 (UTC)

nbsp in wikilinks

Hi

I have seen a recent trend to include &nbsp in wikilinks - e.g. [[The link to number 7]]

Is a Wikilink inherently protected from breaking on a line, in other words is an nbsp necessary?

These introductions are causing a problem when editing or viewing diffs as it prevents correct linking and results in a "Bad title" error page. Chaosdruid (talk) 18:26, 2 August 2011 (UTC)

This is explained at MOS:NUM#Non-breaking spaces and the footnote it points to. It isn't just a recent trend; I've been doing it since the April 2010 discussion that footnote links to. A wikilink is not inherently protected from breaking on a line. You can experiment with this example of [[World War II]]: World War II by un-maximizing your window (that is, click the square by the red X in the upper right corner), and adjust the window size until "World" shows on one line and "War II" shows on another line. I just repeated that experiment to make sure. I'm unaware of any problems, but of course if you're the first to tell me about something happening everywhere, then I got the wrong advice and we need to stop, and remove the nbsps from the links we have.

So what "problem when editing"? When editing you see the nbsp, which causes a problem if you don't know what it means, but that problem isn't restricted to links. I suppose it might also cause a problem among people who think it won't work in links, even though it does, but the worst they could do is remove it. "When viewing diffs, as it prevents correct linking"? I don't get it; links don't work when viewing diffs anyway. In normal reading mode, the links do work; click my World War II link above. "a 'Bad title' error page"? What is that? How do you get it? Can you show us an example? Art LaPella (talk) 21:42, 2 August 2011 (UTC)

It's not clear to me what user Chaosdruid means. The wikisoftware now seems to convert all spaces, nonbreaking or not, into underscores in the link. That should work, and I have not seen any problems. --Robert.Allen (talk) 22:34, 2 August 2011 (UTC)

The problem is that when editing, or viewing diffs, if I click on one of the links with nbsp in it, I get the error page instead of the target article - Links do work when viewing diffs, at least they do for me (you do know about holding ctrl then clicking to open them in a new tab?).

Example: click edit for this section, then hold ctrl and click on your World War II example. Doing that opens a special page "Bad title".

"How to resize your browser 101" was a little beneath my knowledge lol :¬) Chaosdruid (talk) 00:42, 3 August 2011 (UTC)

"click edit for this section, then hold ctrl and click on your World War II example" is a little above my knowledge. I have never heard of such a procedure, and when I tried it nothing happened, neither when I clicked World War II nor when I held ctrl and clicked this example. In edit mode, of course, it looks like [[World War II]] and [[example]]. Is that what you want me to click while holding down the "ctrl" key? Firefox 5.0, Windows XP Professional, Service Pack 3. Anyway, it looks as if others will need to decide if this is a significant enough problem to warrant removing nbsps from wikilinks. Art LaPella (talk) 04:11, 3 August 2011 (UTC)

Having a nbsp inside a wikilink is a great inconvenience in edit mode - it never displays properly. When I select a wikilink in edit mode and mouseover, a menu pops up taking me to the relevant article; a nbsp screws that up. For that reason, I have always eschewed that method, even in my dates script (except for images). Thanks to the above, I find that the nbsp inside wikilinks works as it is meant to in display mode. I still intend to avoid using the nbsp inside links. --Ohconfucius ^¡digame! 04:50, 3 August 2011 (UTC)

Links should be tested using Preview. --Robert.Allen (talk) 05:56, 3 August 2011 (UTC)

I'm sorry. Links should be tested by whatever means is most convenient/effective. The most important constraint for most people is time, so whatever we do is usually driven by that principle. The edit screen is difficult to navigate and is sometimes difficult not to preview, and use of Popups (with its mouseover feature for wikilinks) is a great tool to make life easier. Being forced to use preview for a minor check because someone has inserted nbsp is a real productivity sump. --Ohconfucius ^¡digame! 06:26, 3 August 2011 (UTC)

(ec)

@Art - edit mode and click this - if you do not hold ctrl while clicking it will try and open it in this tab.

If it doesn't do anything, it may be that you do not have some of the extras enabled in your Wiki preferences that I have - though I think it is possibly something else. I also am running FF5, though with Vista 64 bit (which should not really affect it).

The main gripe for me is when checking that links point to the correct place, that the main articles contain the material summarised in the one I am editing/comparing changes via diffs. If this was to become a widespread thing I really think that Wikilinks should have some sort of non-breaking protection added as that would prevent any problems. At present nbsps only work in the preview window or in reading mode.

@Robert.Allen - we are talking about the links in edit mode and as diffs, why would we have to preview every time we want to check a link or diff, or why have to go to the actual document when we have the facility to hover over them? For example, looking at the diffs for this page [1], when I hover over Arts addition of "[[World War&nbsp ;II]]" I do not get a preview of the start of the World War II article, I get "/* cache key: enwiki:resourceloader:filter:minify-css:5:f2a9127573a22335c2a9102b208c73e7 */" Chaosdruid (talk) 06:31, 3 August 2011 (UTC)

I clicked "edit" and found [[World War II|click this]]. That made me choose between clicking "World War II" and "click this". So I tried it both ways. I tried both ways without the ctrl key. I tried both ways with the ctrl key. I tried both ways with the other ctrl key. Nothing. Oh well. I haven't installed "Popups" as Ohconfucius has, so maybe there's something else I didn't install. Art LaPella (talk) 06:48, 3 August 2011 (UTC)

The link user Chaosdruid provided was for a History version comparison. In any case, when one scrolls down to the link it is certainly possible to open it in either a new window or a new tab without errors. If these users have special software installed to customize their edit windows which produce popups, then this certainly applies to a only few editors. If that's the case, the problem is probably with this customization software, and not relevant to most editors. --Robert.Allen (talk) 08:40, 3 August 2011 (UTC)

I don't use popups, but sometimes I highlight links and use the middle mouse button to call up a Google search on the highlighted text. Any nbsp needs to be edited out of the search box and a new search done. That's no big deal to me, but if it bothers many people, a way to avoid it is, instead of
[[World War II]], use a pipe:
[[World War II|World War II]].
--Stfg (talk) 09:49, 3 August 2011 (UTC)

@Art - what happens if you do the same process with this one? - click edit and click on this link World War II (with and without holding the ctrl key)

@Robert - how exactly is it "certainly possible", in other words what process did you have to go through to make it possible, as that implies that it does not do it on its own without some sort of intervention by the editor? Also I cannot see how you have determined your numbers on how many editors use pop-ups or not, can you please tell me how you did that?

@All - perhaps it would be simple to gain consensus for a recommendation that the nbsp is only used in piped links, and as part of the pipe, not the actual link? After all, surely we are here to make progress, not limit us to what works when we disable everything. Chaosdruid (talk) 18:58, 3 August 2011 (UTC)

Popups is an EXTREMELY widely used tool, so it should not be dismissed as a 'minority'. --Ohconfucius ^¡digame! 01:40, 4 August 2011 (UTC)

Still nothing. Assuming a left click. Hovering over the link in the green rectangle on the history page doesn't do anything either. Art LaPella (talk) 00:23, 4 August 2011 (UTC)

@Chaosdruid: I use the Safari browser. To open a link I generally use one two methods: hold down the command key and click the link to open it in a new tab in the same window OR hold down the control key and click to reveal a popup contextual menu from which one can choose to open the link in a new tab or a new window. Both of these methods work fine with these links that contain nobreak spaces without piping.
@Ohconfucius: after checking your link I must agree, "Navigation popups" are used by a lot of editors, and if they are a minority, they are probably a very important minority.
@All: it still seems possible that the problem can be fixed in the Navigation popup Javascript. I'm not sure how long it has been since support for nobreak spaces in wikilinks was added to the wikisoftware, but I only noticed it and started taking advantage of it recently. I suspect it may be easier to get the Navigation popup Javascript modified than to get all editors to pipe these links. (If Chaosdruid is not using the Navigation popup Javascript, then I can't imagine where his problem lies.) --Robert.Allen (talk) 02:04, 4 August 2011 (UTC)

@Art - I guess then, that you have not ticked the option to "open links in a new tab" (in your "My preferences" page "Gadgets" under the Browsing section), or that your "Editing" page does not have some of those other functions enabled (under "Advanced options"). Apologies for asking you to try that so many ways and times, it must be a little frustrating :¬)

I assumed (I know, ass out of u and me *blushes*) that everyone with Firefox had this functionality - see [2] and [3]

@Robert - This post was originally made about issue which did not include popups (see original post at the very top of this section), this is when simply editing a page and clicking, or looking at diffs and clicking. Pop-ups are an entirely different thing, I was just giving the result pop-ups gives when the link is hovered over. Chaosdruid (talk) 02:11, 4 August 2011 (UTC)

Correct. Despite 39,377 edits since 2004, clicking "Gadgets" shows (if a techie can actually believe this) nothing but EMPTY BOXES. Art LaPella (talk) 05:01, 4 August 2011 (UTC)

I only see an option for opening external links, not wikilinks. I think I'm still in the dark on this. Sorry. --Robert.Allen (talk) 07:33, 4 August 2011 (UTC)

Since popups is being used as a way of checking links while editing, didn't they ought to behave in the same way as links in ordinary windows? Otherwise, the check seems unreliable. So perhaps this should be considered a bug in popups? --Stfg (talk) 08:39, 4 August 2011 (UTC)

Popups while editing? I do not get pop-ups in the edit window while editing, I use left-clicking the link and holding ctrl to open in a new tab. (the same as when looking at diffs) Chaosdruid (talk) 13:30, 4 August 2011 (UTC)

Sorry if I used the wrong term, but that functionality is not available until you enable something. I thought that something was being referred to as "popups" in this thread. Whatever it is that has made it possible you to do that in the edit window and in diffs, if you want it to give you a reliable check of a link, then it needs to behave in the same was as the software behaves when you click a blue link in the read window. Otherwise the check is unreliable. That's all I meant. --Stfg (talk) 15:14, 4 August 2011 (UTC)

No problem, I mentioned it earlier, but it might have been lost in all those @s: "I assumed (I know, ass out of u and me *blushes*) that everyone with Firefox had this functionality [ctrl + left-click] - see [4] and [5]"

It is a basic default functionality of Firefox, and I believe of IE. If Art cannot get it to work, then maybe something is different in Art's FF or Java setups (xe might have turned it off), rather than it being a Wiki software thing. Chaosdruid (talk) 17:57, 4 August 2011 (UTC)

@Chaosdruid: Are you using or have you tried wikEd by any chance? Apparently it shows links in the Edit window. (I just tried wikEd, and I got the error you are talking about when I tried to open a wikilink in the Edit window which contained an nbsp.) --Robert.Allen (talk) 18:08, 4 August 2011 (UTC)

I imagine that the standard version of the edit window also shows links :¬) It does not allow them to be opened whilst editing - how would one open links from there? preview, scroll, find it, ctrl+click (shudders) all those extra movements just to open a link seem a little wasteful to me. That is one of the reasons I use WikiEd (used by around 2,000 editors) and mainly because the standard editor is not that good at anything apart from editing.

That would not explain the diffs view links not working though, nor the pop-ups problem. Chaosdruid (talk) 19:22, 4 August 2011 (UTC)

It seems to me that all these add-on editing aids just need to be updated so they keep up with the improvements in the wikisoftware. I think you will find it difficult to roll back the change, since it was probably implemented in response to requests from other editors. It was interesting to learn about the add-ons however. Thanks! --Robert.Allen (talk) 19:02, 5 August 2011 (UTC)

I tend to use {{nowraplinks}} --Trevj (talk) 05:10, 17 September 2011 (UTC)

Multiple links in a table?

For a table of Fraternity Conventions, if out of say 50 conventions, 3 were held in Chicago, would all three Chicago entries be linked? This is a sortable table.Naraht (talk) 11:03, 13 August 2011 (UTC)

Normally yes, the link should be repeated (particularly as a sortable table), but I do ask to consider if you need to link the very well-known geographic city of Chicago in a non-geographic article. --MASEM (t) 11:50, 13 August 2011 (UTC)

So you would support removing the links to cities in both fraternity convention tables and fraternity chapter lists that includes cities?Naraht (talk) 13:47, 14 August 2011 (UTC)

Big cities. Art LaPella (talk) 17:25, 14 August 2011 (UTC)

So if they've had two conventions in Chicago, Illinois and two in Shreveport, Louisiana out of the 80 conventions they've had, should it be 0, 1 or 2 links for Chicago and 0, 1 or 2 links for Shreveport? Naraht (talk) 22:20, 14 August 2011 (UTC)

WP:OVERLINK doesn't define "likely to be already known by almost all readers", but I would say 0 for Chicago and 2 for Shreveport. Art LaPella (talk) 23:10, 14 August 2011 (UTC)

That just looks really wierd in a table, to have some cities linked and some not. Both links are of equal importance to *that* article, aren't they?Naraht (talk) 00:30, 15 August 2011 (UTC)

Not if the reader recognizes Chicago but not Shreveport. If you want to change the guideline, I generally leave that to others. Art LaPella (talk) 01:07, 15 August 2011 (UTC)

Well, this of course flags up the regularly raised - and never-answered - point of how you can ever define what is or is not "well known" or what, for example, is a "big city" in the context of a global encyclopedia like WP; or where any consensus on the practical details of these points has ever been agreed (which is a separate point from any suggestions of changing the guidelines). Even if that were possible and had been done, it ignores the fact that links are about enabling navigation to detailed encyclopedia pages - and should not necessarily be excluded simply because some, or even most, people might know that Chicago is "a city in the US, somewhere in the middle north".

Equally, in terms of tables and lists as opposed to main article text, surely we have to be a bit more liberal in how we apply what is, after all, only a guideline anyway - as Naraht says, and many others have observed in the past, it looks odd to have some terms linked and not others, as well as, per the first point, being somewhat arbitary. N-HH talk/edits 13:07, 15 August 2011 (UTC)

Its understandable that not every possible WP reader will understand what "Chicago" is. But more than a slim majority would - likely closer to 80-90%. We can probably say the same for most world capitals (at least, those in NA, Europe, Asia, and Oceania) and 10M+ population cities like NYC, LA, Houston, etc. And if all we are doing is linking the name of the city in a non-geographic article to help people ID what that city is, then its absolutely not necessary to link those cities that are likely to be well known to WP readers. On the other hand, if this was an article about, say, lakes or some other geographic feature, I could see the city being linked to provide that geographic reference. This is not that type of case, however. --MASEM (t) 14:29, 15 August 2011 (UTC)

Well, actually, no. Let's not throw out years of more intelligent, selective, useful wikilinking because someone has turned up after six months' absence to start lobbying for maximal linking. Tony (talk) 14:18, 15 August 2011 (UTC)

Tony. Where am I lobbying for maximal linking? Someone came here asking a question, and I gave my take on it. Why do you need to fall back into cheap sarcasm and strawmanning (after six months)? And do you have anything constructive to say? N-HH talk/edits 14:48, 15 August 2011 (UTC)

Masem: thanks, but I still maintain that it is hard to see where we draw the line, or how we can even try to work out where. Plus, as noted, I think there's a difference when we're talking about lists and tables; and, it's not about ID-ing as such, it's about navigability as well - and of course, in a way, the more well known a place is, the more likely people are to want to go to the page (I'm pretty sure pageview stats bear this out, eg that the US page is one of the most popular here). N-HH talk/edits 14:48, 15 August 2011 (UTC)

In this specific case I find it hard to believe the city names are there for navigation as opposed to simply providing a wikilink to an unfamiliar proper name. I'm not saying the location column of the table is unnecessary, but it does offer much context compared to the rest of the details given. Therefore, it is unlikely that someone is going to land on that article and use the location links because they want to learn more about that location; eg they aren't very germane to the subject matter. But, as they are proper names, we should be linking the uncommon ones for those to learn what the city is. This is obviously not true for every table, but in this case, I would argue that the links are only there for uncommon proper name aid, nothing more. And hence why Chicago shouldn't be linked city it is a reasonably well-known city. --MASEM (t) 15:18, 15 August 2011 (UTC)

I'd have to wonder how many people look at most lists and tables anyway tbh, let alone navigate from them. But, leaving that to one side, and pushing the Chicago point a bit further - I know where Chicago is; I know what state it's in; I've been there twice, which surely puts me in a <1% minority of English-speaking people. But even then I don't know how many people live there, when it was founded, how many colleges there are in the city, where its convention centres are etc. Nor (until just now) have I looked at the WP page on Chicago, but might well choose to out of curiosity if I came across a link to it on another page. Sure, some of those considerations probably fall short of giving it a "relevant/germane" qualification in the case being cited here, but not all do, and it just seems it's a little inflexible and limited - on several levels - to simply say "it's well known, so don't link, even in a table" and leave it at that.

Anyway, I think we may agree to disagree on this. But I do think it's important that people who ask a question about this kind of thing know there are differing views and interpretations of the style guideline. Also, I can't help but noticing (when I just went to look for it) that since I was last here, that bit of the guideline that said "Think before removing a link—it may be useful to other readers" has gone AWOL. Was it agreed somewhere to remove that? N-HH talk/edits 16:17, 15 August 2011 (UTC)

Answering my own question .. no, it was not agreed - it got slipped away with this edit towards the end of last year. Some of changes unilaterally made by that edit were later put right, by the two edits on this diff, but that deletion seems to have escaped restoration - by omission rather than because there was any discussion/agreement about it specifically. (Here's the general discussion. I, for one, missed it back then). N-HH talk/edits 16:27, 15 August 2011 (UTC)

Underlinking - places, languages and geographic features

I've noticed that wp:overlink is being enforced over zealously of late. I've even found myself being interested (however tangentially) in something mentioned in an article but being forced to copy and paste the text into the search bar because no link was provided. Either that or one is forced to scroll up or down several screenfulls to find the link.

While common word and expressions should only be linked when relevant to the topic, I don't think this rule should apply to proper nouns. If a place, language or geographic feature are relevant enough to be mentioned in an article they should be linked. We should avoid making judgements over what the reader is likely to be familiar or what might interest them. — Blue-Haired Lawyer ^t 13:41, 25 August 2011 (UTC)

"United States" is the most commonly used example of something that shouldn't usually be linked. The María Guinand article, for instance, briefly mentions "the Oregon Bach Festival held in Eugene, Oregon, USA". Simply mentioning the USA in that context is unlikely to make someone want to read all about it. Linking "USA" is more likely to distract them from useful links than to be useful in itself. Readers came to that article to read about María Guinand, so they are unlikely to want to read all about the US instead. Not 100.000% certain, just 99.999% unlikely. Of course we should make judgments of what might interest the reader; you would surely object if I copied the New York telephone book onto your user page. If you think "United States" should always be linked (and not just in places like the California article) then why is there any word you don't think should be linked? Art LaPella (talk) 17:10, 25 August 2011 (UTC)

To BHL - I agree with you, up to a point. As wp:overlink itself says, I think common geographic etc terms (however exactly those are to be defined anyway) do need to be "relevant" to be worth linking, rather than mentioned simply in passing. However, most delinking edits, done by script or otherwise, go way beyond even this, often leaving obvious links bare and random inconsistencies as to what is left linked and what is not (eg "XX is a city on the border between France and Belgium"). Also, on the grounds of navigability, I think the balance should be to retain a link rather than strike it; plus, as you say, it is impossible to make a sensible judgment about which reader knows what when they come here, and what links they might want or even "ought" to be using when they do get here. Anyway, I won't repeat in any details all the arguments I've made ad nauseam about this in the last six months since I first noticed the scripts and other edits running through pages on my watchlist (see archives). Be prepared as well for other familiar faces here to come in from the other side, insisting that this is all correct and proper, and that they are "improving the encyclopedia" (implying you are not, or do not wish to), as they usually do when the latest outsider pops in here querying edits of this sort ..

To Art, the US page is one of the most-visited pages here. People want to go to that page. Why make it harder? Yes, it is probably overlinked, as a whole, but I think it's hard to say we should pretty much never link it, especially in geographic articles about places in the US. And I really don't see the comparison between providing a link (which simply makes a word on a shared page turn blue and provides easier navigation) and dumping 1000s of pages of text onto someone's user page. Apologies, but that's a very odd point to make. Also, BHL I think is pretty clear about the broad principles under which they think things should be linked - and not linked. Please let's not go down the road of suggesting that everyone who simply queries, however casually, the extent of some of the current delinking of places, languages etc is somehow demanding links to "the" and "suitcase" in every article in which they appear.N-HH talk/edits 17:29, 25 August 2011 (UTC)

The US page is one of the most-visited, probably because we have so many irrelevant links to it, not because people reading about María Guinand suddenly want to read all about the US. If we linked every occurrence of "the", that article would be among the most-visited. "geographic articles" – sometimes, which is why I mentioned California. Depending on how you define it, only about 1% of United States links are from "articles about places in the US" (using a random sample of "What links here"), so let's not miss the forest for the trees. Dumping 1000s of pages is relevant to the actual quote I was rebutting: "We should avoid making judgements over what the reader is likely to be familiar or what might interest them", and a similar point applies to overlinking. I didn't suggest Blue-Haired Lawyer wants to link "the", which he has already denied – I asked why not, considering that he always wants to link the US. For other editors, the oft-damned sea of blue argument is a very relevant counter-argument whenever they advocate linking anything the reader might possibly want to read, not what he more likely wants to read. Art LaPella (talk) 20:35, 25 August 2011 (UTC)

I hadn't the common American habit of giving the address of a place after it's name to mind when I wrote my proposal. Mentioning the USA in the sentence "the Oregon Bach Festival held in Eugene, Oregon, USA" seems a bit superfluous to me, it's not like there's an Oregon anywhere else?! (I just checked, there are but they're all in the US too.) But I don't really think the USA is mentioned, no more that the definite article is mentioned, even though the words USA and the appear in the example. Eugene is mentioned and appropriately linked to, even though it's by no means obvious that readers of an article about María Guinand would be interested in a small town in Oregon. The problem is that on my reading of the policy it shouldn't be linked to, as the town is not "particularly relevant" to the article on María Guinand.

But I'll give my own example now. None of the countries mentioned in the lead of the euro article are linked to but shouldn't they be? They all use the euro as their currency, isn't that enough relevance? — Blue-Haired Lawyer ^t 23:19, 25 August 2011 (UTC)

I think I would have linked those countries. Wikipedia has plenty of wiki-warriors who enjoy such issues. ("Oregon, USA" is arguably an example of Wikipedia's overreaction against "US-centrism".) What gets me relatively excited is to click United States, click "What links here", and look at typical examples in that list, with United States links that are very unlikely to be used. The existing WP:OVERLINK guideline already acknowledges "particularly relevant" exceptions, and I would call Euro such an exception. My reading of the OVERLINK guideline differs from yours on Eugene, because Eugene isn't "likely to be already known by almost all readers" including Europeans. Art LaPella (talk) 23:56, 25 August 2011 (UTC)

Linking a whole row of euro countries in the first sentence: aside from the wash of blue at the opening and the assumption that a reader will want to divert immediately to another article rather than to read about the euro, linking to whole country articles is to assume the reader doesn't know:
(1) where those countries are (fair enough, but a narrow-based question that should be answered in a continental map), and
(2) that they want to read a full article about every aspect of the country they click on.
This is why whole-country article-links are often below the line of utility.
The "eurozone" and "European Union" are linked in the first sentence, which both give a description of the countries in the super-state in more focused context. But people just don't read articles to divert to link-targets five seconds after starting. And if they don't know where or what those countries are, would they spend a long time clicking on each link in succession? Would they click just on "Germany"? Or would they actually read the subject article on the subject of the euro first, given that they've gone there explicitly in the first place to read about it.
Much more focused links in the context of the euro are provided in the first table—to the original currency of each country. And the map below, strangely, names the countries that aren't in the zone, not those that are. That could be fixed, rather than resorting to the multiple link diversions to huge, mostly irrelevant articles that almost no readers will follow through. (BTW, either WP's map or the map at Ref 2 at the end of the sentence is wrong in one respect: Slovakia—is it in or out?) Remember that WP articles are not "reliable sources"; the ref presumably is, and should stand in preference to unfocused internal links). Then we have a map of the Eurozone, to show the geography of the whole zone (if not that of each individual country).
If links to individual country articles were provided, they should be to either or both the economy section of the country article or to a daughter article on the economy of that country (like Austria#Economy, which has a section on currency, or even better, Economy_of_Austria). Second-best would be to deceptively pipe these to the country-names alone right at the top, just where readers are getting into the topic of the euro, and are least likely to divert, especially when the pipe will indicate an over-broad target (better that WP had not damaged wikilinking in the past by soaking it in deceptive links). First choice would be to link the "Economy of" articles explicitly in a "See also" section. That would be a plainer, better positioned, more sophisticated use of wikilinking.
Now, I'm not prepared to engage in an extended debate yet again about this matter, which has been discussed before; so you won't hear much more from me. In summary, we need to uphold the smart use of wikilinking, and not to rely on generalised carpet linking in place of fixing maps and presenting readers with explicit and focused links that are not deceptively piped—which has only one advantage: it's quick and easy for editors to do). Many editors are willing to actively optimise the use of wikilinking, and very few object. Tony (talk) 01:47, 26 August 2011 (UTC)

The reason to link to the countries in the Euro article would be to study their economies, so "Economy of Austria" etc. would be ideal if it were the Main Page featured article. Since it isn't, AWB could process around 50 United States "What links here" articles in the time it takes to research an "Economy of Austria" solution. Editor time is finite. Art LaPella (talk) 13:50, 26 August 2011 (UTC)

Wouldn't it be more appropriate to link to the equivalent "Economies of X" articles instead of the country itself in such a case? Also, arguably, when a group of highly-related topic are present, such as what I'd consider "Economies of (European Country)" articles, this would justify a upfront navigation box to allow editors to jump between them quickly (in addition to a end-of-page nav box). --MASEM (t) 13:59, 26 August 2011 (UTC)

Yup, if it can fit in at the top without trouble. Otherwise, the "See also"; and it's sometimes possible to put in parentheses a short note prominently in the lead, like "(articles on the economies of each country are linked in "See also"). You might have a neater wording, or the navbox might work. Either will be much more useful for readers, Masem. Tony (talk) 14:24, 26 August 2011 (UTC)

Overlinking vs. overpolicing

Wow, I just read this page after probably not looking at it for a few years, and I can't believe some of the suggestions given here. In particular:

"Avoid linking the names of major geographic features and locations, nations, languages, religions, and common professions."

Really?? So we shouldn't link to the Nile or Germany, or English anymore, just because they're well known things? I'm sorry, that's just plain stupid. This is an encyclopedia. We should link to topics we have articles on that the reader might want to find out more about. That should be the overriding concern. (Of course, if this guideline was just discouraging links to, say, the actual words river or nation, or language, that would be different.)

"Too many links can make the lead hard to read."

Again, really? All that blue is really that distracting? How long have people been using the web now? Yes, overlinking is a real phenomenon, but articles are not disambiguation pages, and should not be pared down to "only the most relevant links". As has been pointed out above, underlinking is a real problem, since a reader who wants more info about an unlinked term must then search for it separately (perhaps even having to first open up another window/tab so they can easily continue reading the article — I have to do this kind of thing all the time when I copyedit/wikify articles).

"Overlinking should be avoided, because it makes it difficult for the reader to identify and follow links that are likely to be of value."

The source cited in a footnote to this sentence is talking about linking between different web sites in general, not links between articles in an encyclopedia. It's almost not even relevant, if you look at the examples the author is talking about. But to address the guideline itself: In contrast to the "having to search" problem I mentioned above, this "too hard to read" issue seems to me to be a mostly made-up problem. I can say with some confidence that I, personally, have never read an article with lots of links in it and thought, "Wow, this is hard to read." I have thought, "Wow, this article is overlinked," but only because it linked to things like "regular English words" or lots of "insignificant" dates.

This reminds me of the old inclusionist/exclusionist schism, which (ISTM) was largely resolved in favor of inclusionism in most areas. (If you want to know what I'm referring to, you'll have to go search for it yourself. ;)

So, anyway.... to narrow this down to a specific case, here's what actually brought me here in the first place. Recently I wikified the lead section of the University of Houston article from this (I've taken out the links to references):

The University of Houston is a public research university located in Houston, Texas. Founded in 1927, it is the flagship institution of the University of Houston System. The third-largest university in Texas, UH has over 38,750 students on its 667-acre campus in southeast Houston. It was known as University of Houston–University Park from 1983 to 1991. The Carnegie Foundation classifies UH as a top-tier research university. The institution ranks among the Top 50 American Research Universities, and is in the Top 300 Academic Ranking of World Universities.
The university offers over 300 degree programs through its 12 academic colleges on campus—including programs leading to professional degrees in law, optometry, and pharmacy. The institution conducts nearly $120 million annually in research, and it operates more than 40 research centers and institutes on campus. Interdisciplinary research includes superconductivity, space commercialization and exploration, biomedical sciences and engineering, energy and natural resources, and artificial intelligence. Awarding more than 7,200 degrees annually, UH's alumni base exceeds 250,000. The economic impact of the university contributes over $3.1 billion annually to the Houston economy while generating about 24,000 jobs.
The University of Houston hosts a variety of theatrical performances, concerts, lectures, and events. It has over 400 student organizations and 16 intercollegiate sports teams. Annual UH's events and traditions include The Cat's Back, Homecoming, and Frontier Fiesta. The university's varsity athletic teams—known as the Houston Cougars—are members of Conference USA and compete in the NCAA's Division I in all sports. The football team regularly makes bowl game appearances, and the men's basketball team has made 19 appearances in the NCAA Division I Tournament—including five Final Four appearances. The men's golf team has won 16 national championships—the second-most of any NCAA golf program.

To this:

The University of Houston (abbreviated UH or UofH) is a public research university located in Houston, Texas. Founded in 1927, it is the flagship institution of the University of Houston System. The third-largest university in Texas, UH has over 38,750 students on its 667-acre campus in southeast Houston, as of fall 2010. It was known as University of Houston–University Park from 1983 to 1991. The Carnegie Foundation classifies UH as a top-tier research university. The institution ranks among the Top 50 American Research Universities, and is in the Top 300 Academic Ranking of World Universities.
The university offers over 300 degree programs through its 12 academic colleges on campus—including programs leading to professional degrees in law, optometry, and pharmacy. The institution conducts nearly $120 million annually in research, and it operates more than 40 research centers and institutes on campus. Interdisciplinary research includes superconductivity, space commercialization and exploration, biomedical sciences and engineering, energy and natural resources, and artificial intelligence. Awarding more than 7,200 degrees annually, UH's alumni base exceeds 250,000. The economic impact of the university contributes over $3.1 billion annually to the Houston economy while generating about 24,000 jobs.
The University of Houston hosts a variety of theatrical performances, concerts, lectures, and events. It has over 400 student organizations and 16 intercollegiate sports teams. Annual UH events and traditions include The Cat's Back, Homecoming, and Frontier Fiesta. The university's varsity athletic teams, known as the Houston Cougars, are members of Conference USA and compete in the NCAA's Division I in all sports. The football team regularly makes bowl game appearances, and the men's basketball team has made 19 appearances in the NCAA Division I Tournament—including five Final Four appearances. The men's golf team has won 16 national championships—the second-most of any NCAA golf program.

(I'm not just linking to the diff because my insertion of newlines before the "ref" close-tags to improve "skimmability" of the source makes the diff almost useless to see the actual changes.) Another editor reverted all of my changes, citing this page as his justification when challenged. Is the latter version really overlinked? Really? (Be sure to "mouse-over" the links to see what articles they link to.) I mean, OK, maybe you could quibble about, say, golf or law... but to not link research university or Houston in the opening sentence? That just seems crazy to me. And to not link terms like superconductivity? Seriously?

My point is this: if editors are citing this guideline to limit links to this extent, things have gone too far. We need to clearly state that the overriding concern is to link to terms that people might "reasonably" want to know more about, not to limit links to only topics directly related to the topic of the given article. - dcljr (talk) 09:51, 31 August 2011 (UTC)

That delinking looks extreme to me. We don't usually link to words like "Germany" (with exceptions like the Denmark article) because, as you put it, "We should link to topics we have articles on that the reader might want to find out more about." If an article mentions Germany only because (for instance) a meteorite happened to land in Hesse and people might not know where that is, then someone reading about meteorites is unlikely to suddenly want to read a Germany article that doesn't even mention meteorites. But "superconductivity"? Many readers would have no idea what that is, and would want to read the article or at least the introduction. Art LaPella (talk) 21:28, 31 August 2011 (UTC)

'[[Public university#United States|public]]' and '[[research university]]' are bunched/chained, as are 'Houston' and 'Texas'. The bunching can be reduced by separating the words, or by removing the less important link. Ordinarily, I wouldn't link Houston, but as the subject is based there, it is arguably germane. Here, I would say unlink 'public' and 'Texas'. 'Law', 'pharmacy', and golf are common enough so they would not miss being unlinked, as would '[[college football|football]] and [[college basketball|basketball]] – which college doesn't have a football and basketball team? --Ohconfucius ^¡digame! 01:22, 1 September 2011 (UTC)

I agree with Art LaPella that the unlinking went too far. Some items are obviously undesirable as links ("Texas", which is linked to prominently in "Houston", and bunches; and "engineering" and some of the other topic/profession items is starting to treat wikilinking like a dictionary (Pillar: "Wikipedia is not a dictionary"). "Public" is a deceptive link that hardly any reader would ever click on, and it bunches with "research university", itself a pretty self-explanatory term. Common country-names: please no. Tony (talk) 01:46, 1 September 2011 (UTC)

I guess I have a very different idea of what kind of links are desirable than the people interested in maintaining this policy. For example, I see no problem at all with linking both Houston and Texas. Why not? Some people might want to look at the Houston article, some at Texas. Why prevent one group from getting to their destination in just one click? I don't buy the whole notion that "more links confuse people" — in fact, I don't even think that "bunched" links are much of a problem: after all, don't most browsers allow the reader to see (/hear?) the target of the link before they follow it? (And BTW, it seems somewhat disingenuous to claim that the links to Houston and Texas are "bunched": they're separated by a comma in a standard way that "everyone" knows is referring to geographical/political entities. This can't possibly be considered confusing in any way.) The real justification for the public link (the reason I did it, anyway) is that the terms "public" and "private" can be used in opposite senses in different countries, so the link can clarify the sense for those who are concerned about that. So I would just warn people to keep in mind that just because they might not see the reason someone would want to follow a link doesn't necessarily mean there isn't a good reason (that's why I warned people to actually look at where my links went — not that I chose the best targets in all cases, but some links might seem more relevant once you see where they go). I completely disagree with the idea that major countries, languages, etc., should not be linked when only mentioned "in passing". This seems to me to be a horrible, horrible, horrible idea. Did I mention I think it's a horrible idea? In fact, I would go so far as to say that it's probably the worst guideline I think I've ever read on Wikipedia (among those that had some kind of consensus, I mean). No, really. Think about how you actually use Wikipedia: Do you only follow links that are significantly related to the article's subject? I know I don't. I go wherever looks interesting, and that depends on my mood/goal at the time. Even if I'm looking for more specific information than a link will provide, sometimes I think, "Maybe something relevant will turn up in that article's 'See also' links, or even in the categories." Why prevent that kind of "browsing" for no good reason? (I haven't seen a good reason presented yet, anyway.) - dcljr (talk) 22:53, 1 September 2011 (UTC)

Note: Apparently what I said above about public/private is not true about universities—which I guess I would have discovered if I'd actually read the article I linked to! [g] My confusion, ironically, serves to prove my point: that linking to an explanation of the term might be a good idea. - dcljr (talk) 02:08, 2 September 2011 (UTC)

no irony. The usage 'Public school" is but an English quirk. It's not usual to expect that terminology would carry the opposite meaning. --Ohconfucius ^¡digame! 02:30, 2 September 2011 (UTC)

Does your mood lead you to reread the Germany article, when mentioned only in passing in a completely unrelated article, more often than it would lead you to read about almost any other word? Art LaPella (talk) 03:16, 2 September 2011 (UTC)

Just to reiterate for your benefit, Dcljr, that not everyone here takes such a rigid view on linking and delinking. Indeed, I've always asserted ever since I first came across this problem that most disinterested editors and readers would probably err on the side of retaining more links (a glance at the succession of threads opened by people previously unknown to this talk page querying what's going on suggests that is correct). In addition, it is worth noting what the guideline actually says in more detail - eg that common terms are fine to link if "relevant". Also, until it was quietly edited away with a slightly misleading edit summary, the page used to say "Think before removing a link—it may be useful to other readers". The guideline itself is not as strict in what it says as you fear and as some would like to believe. In addition, it is, after all, only a guideline written by a small sub-group of editors - not a policy.

As for the Houston points being discussed here, I'm sorry, but removing links to "Texas" and even, arguably, to "US" is a clear breach of the guidelines. It's hard to argue that Texas is well known enough to be excluded, and, even if it were, it still falls within the "relevant" exception. Equally, removing the link to "Houston" from the University of Houston page lead, is pretty unwarranted (removing "flagship" I can more than happily live with). I can't see in what way that is helpful or indeed in line with the guideline here. Dcljr's bid to retain them seems wholly correct, by any standard.

And finally (@Art .. and others). Please, please, please stop with the "link every word" digs. I don't know how many more times I can ask this, or why you might think it's still appropriate or relevant to do it. N-HH talk/edits 08:58, 3 September 2011 (UTC)

"Every word" digs are relevant as long as arguments are presented for linking that could be applied to every word, even though the "every word" conclusion of such arguments is explicitly rejected. Repeated pained objections from several people frankly look like evasion of that point. In this case, why would your mood be more likely to make you click Germany than any other word? Art LaPella (talk) 18:04, 3 September 2011 (UTC)

Because Germany, even when linked in passing, is (almost always) referring to a geographical entity that is itself worthy of an encyclopedia article (and not only worthy, its inclusion is positively de rigueur). Words like more or often, or than, OTOH, refer to ideas that are themselves not typically subjects of encyclopedia articles — and if an encyclopedia were to have articles about them for some reason (e.g., some kind of "encyclopedic dictionary"), they would be talking about the words as words in the English language, not as signifiers of other encyclopedic topics. BTW, this might be a good place to explain why I linked the word "flagship", even though it is a "regular English word": My thinking is that it is a somewhat obscure term that is likely to not be understood by a significant precentage of readers (e.g., younger ones, and especially non-English natives), so I decided to link it to assist those who needed an explanation of the term (exactly the same reason I linked de rigueur just above). I was not slavishly following a rule that "Thou Shalt Not link regular English words", nor did I avoid linking it because "then I'd also have to link words like 'the'" (an objection that brings to mind Emerson's famous "hobgoblin" quote). No, instead I used my own judgment, based on years of experience reading and editing Wikipedia, as to what links readers might benefit from. Clearly I have a "more expansive" view than others on this point, but not quite as expansive as your ridiculous straw-man characterization. - dcljr (talk) 10:03, 7 September 2011 (UTC)

You clearly have not read - or at least understood, or thought through the conclusions of - anything I have said then. I do not know what more I can do to help you with that. N-HH talk/edits 19:00, 4 September 2011 (UTC)

Just to clarify, to avoid any confusion, since you seem a little confused (this is more polite than accusing you and others of wilful misrepresentation) - my personal view is that in a page that opens, for example, "The University of Houston is a university in Houston, Texas, USA, that offers a variety of courses and admits students from various disciplines ...", "is" should not be linked; "a" should not be linked; "university" should possibly be linked; "in" should not be linked; "Houston" should definitely be linked; "Texas" and "US" should possibly be linked; "that" should not be linked; "offers" should not be linked; "a" should not be linked; "variety" should not be linked; "of" should not be linked; "courses" should probably not be linked; "and" should not be linked; "admits" should not be linked etc. Quite why I should have to point this out, Lord only knows, but there you are. There are words that pretty obviously no one is ever going to demand are linked, words where there will probably be broad agreement that they should be and words that are more open to debate one way or the other. If you and others could stick to discussing the last of those three - which probably amount to 0.1% of all the words that appear in any given article - that would be marvellous. Thank you. N-HH talk/edits 19:32, 4 September 2011 (UTC)

Then apparently you can't help me understand why "USA" or "Germany" should be linked, but not the unmentionable words. Oh well. Art LaPella (talk) 20:49, 4 September 2011 (UTC)

I can - I hope - at least get you to focus on that question, and on the key issues of relevance and related information and context, rather than make daft "jokes" about things that no one is asking for, not least because pages don't even exist for a lot of those words to link to. You and Tony and others have used this for a long time as an excuse to derail any serious discussion and to taint those of us who might disagree with you as being in favour of a "sea of blue". Re-read the above thread - you bring up an absurd and irrelevant point; I ask you not to drag that irrelevant point out again; you then insist it is a relevant point; I explain in detail why it is not; and you then accuse me of not being willing to debate the substantive point because I have focused, temporarily, on the irrelevant point that you dragged up. Sorry, that's just stupidity or trolling, take your pick. As for "unmentionable words" I have absolutely no idea what you are talking about. Or Germany for that matter, in the example under discussion. N-HH talk/edits 21:22, 4 September 2011 (UTC)

Well no, that isn't what happened. But if you don't want me to mention the Voldemort words again, here's how: "I think 'USA' should possibly be linked because [fill in a reason here, a reason that wouldn't apply to every word in the dictionary as well]". Not a reason that convinces me; I'm no style expert by the standards of this page. Just a reason. Art LaPella (talk) 21:59, 4 September 2011 (UTC)

I wasn't under the impression that Wikipedia must necessarily develop out of rigorous adherance to previously defined rules. I thought we were just trying to make something that would be useful as an encyclopedia. To answer your request, see one possible reason in my other comment above (with the same timestamp as this comment). If that doesn't do it for you, then I don't know what to tell you. - dcljr (talk) 10:03, 7 September 2011 (UTC)

If Dcljr and N-HH are the same person, then thank you for answering the question. Of course if the consensus is to link "Germany" just because it's a typical encyclopedia article, then we might as well reverse the rule, and recommend a link to such countries from every possible article, because the same reason will apply. I suppose we could get the same effect by listing the most frequently read articles on the left side of every page, relevant or irrelevant, just in case someone wants to click them. Art LaPella (talk) 00:15, 8 September 2011 (UTC)

Of course we are not (I'll assume that was a joke of sorts rather than a serious suggestion of sockpuppetry). Although I do agree with their comment higher up on the page that of course no one is ever going to link every word, not least because most verbs, prepositions etc are not encyclopedic topics anyway. Anyway, I ducked out here because it seemed a bit pointless for me to say much more, and I can also only handle one debate/argument at a time here. As for Germany/US or whatever, I am on record many times as saying I do not think countries and other such terms should always be linked. However, I do think, for example, that all countries should be linked, at least somewhere on a page (even if only the infobox), when we are on geographic or place articles (eg "Bremen is a town in Germany"; "The US is a north American country neighbouring Canada and Mexico" etc). It's all about context, and whether they are mentioned in passing or are more specifically "relevant" - as the guidelines say - to the main topic/sub-topic at hand. That's quite a clear and rational criteria and it should be obvious to see why that might be one reasonable option, and where it might - and might not - lead. I'm not sure why you're pushing so hard for me to explain something I've already explained, or for me to explain at length why it will not have alleged consequences that it very clearly will not.

Someone came here and asked for opinions on a specific point - I gave mine, as I have previously on the broader questions. I'm broadly sympathetic to what they were saying, and knew they'd get a bit of heat for suggesting it. It may be they want more linking overall than I might prefer, just as many regulars seem to want far less linking than seems reasonable and sensible to me - but guess what, that's fine. There's far too much rigidity in (de)linking and an assumption that there's a provable "right" way to link and unlink. And that isn't coming from me and any alleged bid of mine to "link every thing, every country, always"; if anything it's coming from those script-based delinkings that remove, for example, every link to France, World War II etc from thousands of articles, regardless of context. N-HH talk/edits 19:16, 8 September 2011 (UTC)

I'll just limit myself to denying an accusation of sockpuppetry (I did wonder if you were the same person, but I wouldn't have said that if I had realized it amounted to accusing you of a Wikicrime.) And this section has me on record as considering a specific delinking to be "extreme". Art LaPella (talk) 20:53, 8 September 2011 (UTC)

[back to left margin, since this is not just a response to the previous comment]

Just a few final (??) comments from me in this thread:

To those who felt that the "delinking" of the University of Houston lead section shown above went too far (which, by my reading of the comments, is actually everyone who has participated in this thread), please note that two subsequent attempts to add fewer and fewer additional links to that article's lead were both unsuccessful, being reverted away by the same editor as before. I can't help but see this as an inappropriate attempt by User:RJN to eschew collaboration and "own" the article in question. Perhaps someone more "sympathetic" to RJN's viewpoint can convince him/her to at least link to Houston somewhere in the University of Houston article! Currently there are none that I can see. This is clearly an unacceptable situation.
BTW, for comparison sake I checked the articles for 10 "arbitrarily chosen" U.S. universities named after states or cities (just the first 10 that came to mind), and in 9 of those cases, the state or city was linked in the first sentence. The only case where this wasn't true, University of Texas, had been edited recently by RJN to remove those links! This says to me that RJN is, at the very least, swimming against the tide of what the majority of WP editors want linked in a lead section, at least in this kind of article.
Art: While I did find your comment about N-HH and I being the same person a bit disconcerting, I realized you only said that because I was replying to a question you posed to him/her. I chose to reply because N-HH had not yet done so, and he/she seemed to be making the same general argument here as I was. Also, you had already used a word that I had used ("mood") in a reply to one of N-HH's comments, so I figured it was sort of "open season" as far as replying to things, regardless of whom they were directed towards.

OK, go on about your lives... - dcljr (talk) 18:18, 20 September 2011 (UTC)

(long break)

For the record, because subsequent comments left on RJN's talk page about the University of Houston (UH) article have gone nowhere, I've raised the issue on the article's talk page. Now, I'm hoping that RJN and other UH-interested parties can talk about this, so I would humbly request that discussants here not go over there and rehash the same arguments they've given here. I'd recommend holding off—perhaps watchlist or bookmark that page—and only join in after some actual discussion has started over there. Thanks. - dcljr (talk) 22:22, 18 October 2011 (UTC)

Moved

My summary seemed to get lost, perhaps it was too long, but the raionale for the move of this page can be found at Wikipedia talk:Manual of Style/Archive 123#RFC: Should all subsidiary pages of the Manual of Style be made subpages of WP:MOS?. — Martin (MSGJ · talk) 11:47, 1 September 2011 (UTC)

That discussion says that renaming everything makes information easier to find, but doesn't explain how. So I hope the rest of you know something I don't. Art LaPella (talk) 03:59, 2 September 2011 (UTC)

Yes, all subsidiary pages of the MoS (involving parentheses in their titles) should be moved to subpages (with a solidus or slash). Tony (talk) 04:39, 2 September 2011 (UTC)

Current consensus on links in quotations?

I see this archive entry which seems to say "It depends." In a BLP the subject is quoted as using the expression "floating signifier" which used to link to empty signifier which explained what it meant, which most people wouldn't know. (I didn't til looked it up myself.) An editor removed it writing: "as a general rule, in my opinion, we should avoid piped links in direct quotations." But it seems this would be an exception to the not very explicit rule. If I'm wrong, do tell, before I revert it. (Note other editors on page aren't too friendly towards subject and prefer he seem as obscure as possible, including in sentence where this is used, so need NPOV opinion.) Thanks. CarolMooreDC (talk) 11:50, 1 September 2011 (UTC)

Sometimes links within quotations seem pretty safe, but sometimes they don't. If there's doubt, you might consider footnoting it and putting the link in the footnote with a brief explanation. Tony (talk) 04:41, 2 September 2011 (UTC)

That had occurred to me. I assume on the word itself? CarolMooreDC (talk) 12:28, 2 September 2011 (UTC)

Propose restoration of removed material

I would like to "formally propose" restoring the following bullet point that was removed with no explanation (none specific to the removal, I mean) back on 12 December 2010:

Think before removing a link—it may be useful to other readers.

This is essentially what I've been saying here for the past week.... - dcljr (talk) 10:39, 7 September 2011 (UTC)

Support. It's hardly controversial, surely. In fact, it should go back in now really without any further discussion here, as there was no consensus - or even any discussion - about removing it. Those who wish to see it gone can then make that proposal. N-HH talk/edits 19:19, 8 September 2011 (UTC)
Oppose—we could add "think before doing X" in hundreds of places in style guides. This is just clutter, and style guides need to be as succinct as possible. Equally, we could say "Think before linking an item"—yet more clutter. Tony (talk) 09:29, 9 September 2011 (UTC)
Oppose – per Tony. We should be careful everywhere, not particularly here. This is just noise. Dicklyon (talk) 23:08, 9 September 2011 (UTC)
Oppose – "Think before you link" seems like a more useful reminder to have in the guideline, judging from the rather systematic overlinking that exists across Wkipedia. --Ohconfucius ^¡digame! 00:27, 10 September 2011 (UTC)

I agree that calling on people to think before doing either would be sensible advice, as well as that, equally, to have both would be rather stating the obvious and somewhat redundant. The point here is that the guidelines very specifically (by design or accident) used to ask people to be more cautious before delinking, and that injunction was removed without any discussion. A few editors delink certain terms en masse by script; nobody links or relinks in the same way. Hence, the issue is slightly different - whether one believes WP is as a whole over- or underlinked, the focus here is as much on the action as the aims behind it. N-HH talk/edits 13:14, 10 September 2011 (UTC)

Propose another addition

Under What generally should be linked, I believe we need another bullet point — something like:

articles that provide necessary context for the current article, especially in the lead section.

Now, I guess this might be "controversial" given the above discussion, but the way I see it, this is entirely consistent with WP:LEAD#Contextual links (and is apparently necessary to remind people of that other guideline). - dcljr (talk) 10:57, 7 September 2011 (UTC)

Weak support: broadly agree, although it's possibly too vague, such that anyone can take anything from it (or maybe that's a good thing in a guideline?). But having said that, if it's in wp:lead, it's odd that it's not here. N-HH talk/edits 19:26, 8 September 2011 (UTC)
Unsure why the lead should be treated differently from other parts of an article. It's case by case: often a more specific link is possible under the lead, avoiding the so-called Easter-egg links that are piped with general items the readers are very unlikely to click on. Tony (talk) 09:27, 9 September 2011 (UTC)

Because the lead is where context is established. - dcljr (talk) 18:41, 20 September 2011 (UTC)

Nah: the lede is among the most densely linked real estate. Context for an article, where judged necessary, is nearly always provided within the background section – the first section after the lead. Merely planting a link in the lede to provide that context is just plain lazy and is of no service to the reader. --Ohconfucius ^¡digame! 09:50, 9 September 2011 (UTC)

Indeed there's no need to go out of your way to add text just to plant a link. The suggestion, though, is that the main contextual information should link to the articles that cover it in more detail. This information normally occurs in the lead naturally, especially the first sentence. We might be intending something different by the word "context". As I am using the word, the topic's definition sets the context for the rest of the article, by placing the topic in the broader context where it's meaningful and important. For example:

The Court of Chancery was a court of equity in England and Wales that followed a set of loose rules to avoid the slow pace of change and possible harshness (or "inequity") of the common law.

The links lead to information about the terms needed to explicitly delimit the topic. These express the most basic facts about the topic: what kind of thing it is, and how it most significantly differs from others of that kind. The rest of the article provides details that fill in the context set by the definition, with links to contextual information relevant to various subtopics. The links in the definition lead to contextual information for the article as a whole. —Ben Kovitz (talk) 20:31, 15 October 2011 (UTC)

Nah2 – ledes with overlinking as an alternative to actual context are annoying. Don't encourage it. Dicklyon (talk) 23:09, 9 September 2011 (UTC)

Well, we arguably have clashing guidelines if we don't do something with this guideline (or with that one). I agree that leads are often overlinked - to the extent that that matters really - but I'm not quite sure we should be pushing "valid" links further down in article simply to keep the lead clearer of blue or to avoid something being a bit "annoying" to some readers/editors. That just seems to reduce functionality and navigability, especially to those people who never get much further than the lead, to no particular benefit. N-HH talk/edits 13:20, 10 September 2011 (UTC)

Ohconfucius & Dicklyon: I don't understand your objections. The context is presumably already there, we would just be encouraging the linking of it to relevant articles. For example, instead of saying, "Acid rain is a rain or any other form of precipitation that is unusually acidic" (which is the unlinked version of the actual opening text of acid rain), it should be (assuming the same [awkward] wording): "Acid rain is a rain or any other form of precipitation that is unusually acidic". That is, if you're presenting right at the outset the defining characteristics of something (what it is, where it is, etc.), you should provide links to relevant articles so that people can look up more information about those things. (And again, before anyone says it: this is not an argument to "link everything you can in the first sentence"!) I mean, in this particular example (which I found through Special:Random), without the links the first sentence is not helpful at all. "Acid rain is acidic rain." Oh, yeah, thanks. But with the links there, it makes some kind of sense to start this way (not saying it couldn't be phrased better, mind you and note that the sentence does actually continue beyond what I've quoted). Regardless, I'm not arguing about wording, just the linking. - dcljr (talk) 18:41, 20 September 2011 (UTC)

Hmm. Okay, maybe I should provide a better example now that I have a little time to do so. Here's a few different examples (also found through "Random article") of what I'm talking about. The particular "contextual links" I'm alluding to are underlined. The rest of the links are not important for this particular thread. (I've linked the bold terms because those are the articles these examples came from. As in the earlier thread, I've removed things like <ref>erences, pronunciations, etc. And sometimes I've made minor grammatical changes to the article text so as not to distract from the main issue being discussed.)

Vallipuram was an ancient capital of the Northern Kingdoms of Sri Lanka. Point Pedro is the nearest town. Vallipuram is a part of Thunnalai.
(actually, I would have linked "Northern Kingdoms of Sri Lanka", but I wasn't sure what the best target would be)
XtratuM is an open source hypervisor specially designed for embedded real-time systems available for x86, PowerPC, recently for LEON2 (SPARC v8) processors.
(I guess "open source" is debatable; "open source software" would probably be a better target to provide the necessary context)
Matthew 5:19 is the nineteenth verse of the fifth chapter of the Gospel of Matthew in the New Testament, and is part of the Sermon on the Mount.
In economics, Benefitive Treasury Measure (BTM) is an economic indicator that attempts to correlate a relationship between immigration and government tax receipts or government spending; put more simply, it is cost-benefit analysis of immigration on a macroeconomic scale.
(targets of linking and inclusion/exclusion of "government" in linked text is debatable, but beside the point: something should be linked to explain these terms)

I hope that my point is clear: in every lead section, there are terms that are so germane to understanding "what this thing is" that they really should be linked to appropriate articles to provide the proper context for what's being discussed — whether "most people" will follow the links or just a few. It just seems to me to be a matter a good style to provide such links (not to mention [again] that this is actually what the majority of WP articles currently do). Which terms are covered by this guideline and what the targets of the links should be is entirely context dependent, and I see no good reason to try to create a list of what kinds of things should always or never be linked (or linked to). (Examples, yes: to elucidate the guideline, and these obviously can be debated here. But not checklists.) And I'm out of time again.... - dcljr (talk) 21:01, 20 September 2011 (UTC)

So... no more comments about this? Even though the two "negative" comments above seem to have been based on misunderstandings? (Ohconfucius: The opening sentences of an article pretty much always try to establish context. Dicklyon: I'm not talking about links as substitutes to good prose. And while I'm at it, Tony: Like I said, the lead is where the context is first introduced; apart from that, I'm not actually trying to treat the lead any differently. Those who would push the linking farther down the page are the ones who would be treating the lead differently from the rest of the article.) So, by my (admittedly biased) reading, this leaves a "weak support", another whole guideline that this is consistent with, and no other particularly relevant objections (other than the allusion to link density). Does anyone have anything new to say in response to my above attempt at clarification? Or, even better, a possible rewording of the original bullet point? - dcljr (talk) 10:15, 15 October 2011 (UTC)

"Immigration"? It's far too general an article to be worth linking in this opening context; it's a dictionary term. The first pillar of WP says "Wikipedia is not a dictionary". "Immigration" is even too general/common to bother linking in the See also section. "Economic indicator" ... well, maybe, but anyone who goes to Benefitive treasury measure is going to understand enough about the world to know what "government spending" is. Again, the lead is the worst place to blue-carpet the text—you want the reader to keep going with the article, don't you? Not to click away from it straight away an an aimless chase. These things can be linked below in the article text, once, or in the See also section. Otherwise, it's a disservice to our linking system.

Tony (talk)

03:38, 16 October 2011 (UTC)

Regarding "immigration": the problem with your analysis, ISTM, is that immigration is one of the two (or three, depending on how you count) things being correlated by this economic indicator. It is therefore one of the fundamental parts of the definition of this thing. Without that word (or something equivalent to it), you can't define what this thing is. Given that fact, I'm saying that it's a prime candidate for linking simply for that reason. If you have a more specific article than Immigration to link the term to, that probably just means that the actual wording of the sentence could be improved. And, BTW, the word "immigration" is in the dictionary; the contents of the "Immigration" article is not. Please don't confuse the two. Regarding "government spending": the link to that article is not merely to a dictionary definition either. So it doesn't matter if most people will "know what 'government spending' is"; it only matters whether they will want to know more. Anyway, these examples were primarily meant to illustrate the concept of "contextual links" as I understand the term, not lead to a debate about which specific terms should be linked to which specific articles in these specific examples. As for your question, "you want the reader to keep going with the article, don't you?, my answer is, "Of course not!" I want them to do whatever they want to do with respect to the article. They can keep reading if they want; they can follow a link out of the article and never come back; they can even close their browser and walk away. I don't care. But for those who want to keep reading I want to provide them well written, interesting, informative text (not try to prevent their escape from the article by providing "no way out"). For those who want to follow a link to more information about something, I want to provide them that link. Since I cannot predict with certainty what kind of information they might be looking for—unlike you, I don't have such a highly developed sense of what links people will and won't need to follow, and why—I can only go by rough heuristics like "relevance", "relatedness", "context", "unusualness", and the like, with a dose of "reasonableness" thrown in for good measure. And finally: "Not to click away from it straight away an an aimless chase." What kind of idiots do you take our readers for? - dcljr (talk) 09:53, 17 October 2011 (UTC)

The problem is that Tony and the other delinkers think that the readers are complete idiots who can't handle the choice of following a link or not. -- cheers, Michael C. Price ^talk 10:02, 17 October 2011 (UTC)

Support and link to WP:CONTEXTLINK: The most salient facts about the topic are in the lead. They are usually the most appropriate for a link. Links mainly lead to context for the topic, and the lead lays down the most fundamental elements of that context. Also, the lead appears first and usually starts with a definition. A reader who needs the most basic information about the topic may well find that definition confusing; linking from the definition to its main elements is especially valuable. —Ben Kovitz (talk) 15:40, 15 October 2011 (UTC)

" The most salient facts about the topic are in the lead. They are usually the most appropriate for a link."—And the lead is usually the least likely place a reader will want to give up the article and divert somewhere else, almost always to a more general topic. An argument could be made that it's better to link to the first occurrence of an item in the body of an article, not in the lead, but I won't push that. Or in the "See also" section, if there are no other occurrences. The lead should normally have links, but only carefully selected ones, or the boat will be capsized by the blue carpet effect and by bunching, which is discouraged by the guideline (rightly so). $Tony (talk)$ 03:31, 16 October 2011 (UTC)

Tony, I share your loathing of "blue carpets", where meaningful emphasis is lost because too much is emphasized, or even reversed because the few non-linked words stand out against the blue. I don't think that linking to the most salient elements of context produces anything like that sort of mess, though. It does occasionally produce two consecutive links, but I think that's a flaw to weigh against the flaw of failing to link when it's called for. I especially share your sentiment that links should be "carefully selected", whether in the lead or elsewhere. Tonight I will gather my thoughts and try to write something that explains the whole of why I think contextual links are important, and why I think they serve rather than harm readability, at least when thoughtfully selected. Arguing little points one at a time might be creating an illusion of disagreement. In an edit summary, you mentioned your concern about "disservice to the linking system". That also sounds like one of my leading concerns.

On another subject, are you sure that the lead is the least likely place for a reader to click away? Speaking for myself, when I am just starting to browse, I often find myself clicking links in leads, quickly running through related articles, sometimes to get the lay of the land, sometimes because I don't know quite what I'm looking for until I've found it, often because I don't understand the definition and I need to back up to read something more elementary. I treat the leads as mini-versions of the article and quickly read and navigate those and then settle down and read for detail if necessary. I don't have any data about how common this is, but I can't believe it's unusual. In any event, I'm wary of sacrificing the link structure's faithfulness to the material in favor of how we think readers ought to or even in fact usually do use Wikipedia. —Ben Kovitz (talk) 06:23, 16 October 2011 (UTC)

Manual of Style/Links

Manual of style pages have the following titles:

Abbreviations Accessibility Anime- and manga-related articles Article message boxes Biographies British Isles-related articles Canada-related articles Capital letters Captions Chemistry Comics Command-line examples Computing Cue sports Dates and numbers Disambiguation pages Embedded lists Film France & French-related Glossaries Hawaii-related articles Icons Images/Draft India-related articles Infoboxes Iran-related articles Ireland-related articles Islam-related articles Japan-related articles Kosovo-related articles Latter Day Saints Layout Lead section Legal Linking Lists Lists of works Malaysia-related articles Mathematics Medicine-related articles Military history Music Music samples Novels Philippine-related articles Philosophy Poland-related articles Pronunciation Proper names proposal Record charts Road junction lists Self-references to avoid Singapore-related articles Snooker Spelling Stand-alone lists Stringed instrument tunings Tables Television Text formatting Thailand-related articles/Draft Titles Trademarks Trivia sections Visual arts Words to watch Writing about fiction Writing about fiction/Draft revision

It seems to me that this page would be more consistent with those if it were "Manual of Style/Links". Thoughts? Lightmouse (talk) 23:34, 20 September 2011 (UTC)

linking countries in tables

I'm probably treading over old ground with this question, but I think it needs clarifying. As far as I'm concerned the current guidelines are ambiguous when it comes to whether you are allowed to link countries in tables. On the one hand we have a guideline that states that countries that should not be linked, yet another that states that links in tables should be allowed to stand on their own. I would like clarification as to whether, countries should be linked in tables or not, and if possible a guideline that specifically states whether they are allowed or not, to end the issue once and for all. NapHit (talk) 18:51, 7 October 2011 (UTC)

"Allowed [or not]" is probably slightly strong terminology, since we are talking about a guideline. Also, the guideline does not say countries "should not be linked" - it only says "major" nations (whatever they are, as ever) should not be linked "unless particularly relevant to the topic of the article", ie when they are simply mentioned in passing. So, in fact, "major" nations should be linked if related to the topic and "minor" nations should be more liberally. Anyway, the comment that tables should stand on their own is, as far as I can see, more an exception to the guidelines on repeated links, which are usually deprecated. The point is: don't avoid linking a term in a table simply because it has already been linked in text. In principle, beyond that, the general rules on linking apply the same to tables as they do to text.

In reality tables are often much more heavily linked - correctly in my view - since the "aesthetic" and "distraction" arguments against overlinking (which are contested anyway) do not really apply outside of narrative prose and because tables are seen by some as having a navigation function. I always find it odd when you see tables that are part linked and part unlinked (eg where "Belgium" is linked but "France" not). It seems wholly arbitrary, needlessly reduces utility and navigability for the the broad range of readers, and in fact is somewhat uglier than having most terms in the table blue and hyperlinked. Anyway, as ever, there are no compulsory rules here, and others will disagree with at least the second half of the above. Does that help? N-HH talk/edits 17:21, 8 October 2011 (UTC)

Yes and no. Yes for the fact that what you say confirms my view that countries should be linked in tables. No because as you say other will disagree, but I suppose that is part and parcel of the project. Hopefully there will be a few more views on the matter, but I agree with what your saying. Personally, I would just like a guideline stating the above, but I'm sure that would be hard to achieve. NapHit (talk) 23:17, 8 October 2011 (UTC)

Counties, US states, ethnic groups etc

And beyond any issues about nations/countries ("well known" or "major", or otherwise), see above ad nauseam, I've noticed that we now seem to be reaching a point where even English counties, ethnic groups (by seemingly random selection) etc are being delinked. I note on that editor's talk page that several others have chipped in on these points from a sceptical viewpoint. I appreciate that the county point is based on the "avoid linking of consecutive terms" principle rather than the "well known" argument, but it surely is going too far to argue that on a page about a town/city we should not have an early and accessible link to the county it is part of or found in. Sure, bunched-up linking may be best avoided where possible, but it is not outlawed. It also seems a little odd to suggest that readers are incapable of working out when the individual items in a string of words might be linking to different pages, given that it becomes obvious as soon as the mouse places the cursor over it. Equally, there is absolutely no rational or guideline-based justification for delinking Italian-American while retaining Swedish-American, or similar edits. N-HH talk/edits 18:13, 9 October 2011 (UTC)

No links in the lead?

And then we have this edit. Not surprisingly (in light of my comments above), I've reverted it and invited the editor to talk about it here. - dcljr (talk) 09:05, 14 October 2011 (UTC)

I agree - again, perhaps unsurprisingly - that that would be a step way too far. I think most people would too, although you never know. It seems to have been done in good faith, by an editor unaware of the wrangling that goes on in this obscure backwater; but as a general point, people really need to get consensus before making changes to this or related pages, even if certain changes or "tweaks" seem obviously beneficial to them. N-HH talk/edits 12:09, 14 October 2011 (UTC)

The suggestion was: "Try writing the lead without any links at all; then link below where the lead is expanded." I actually think writing first without links is good advice, though it applies equally to any prose text on Wikipedia. You're in a better position to tell which facts have the most salient connections to other topics after you've got some facts down. It would be nice to mention this somewhere in the page, though not in an especially prominent spot. If the suggestion is really to leave the lead bare of links and link only from the body, I think that's a very bad idea. To start with, it violates the lead's ability to stand on its own as a concise version of the article (WP:MOSINTRO). —Ben Kovitz (talk) 20:50, 15 October 2011 (UTC)

removed overlinking in image captions edit

I have removed an edit made by Ssilvers [6] that promoted repeating links into image captions:

1. The edit was made deceptively ("ce" as summary).

2. Change was made without discussion. (and previous version was stable for at least 2 years, I checked).

3. Image captions are not a type of metacontent like a navbox or table.

4. Image captions are one of the WORST places to put formatted text. That is some of the most high gain text in an article.

5. It really doesn't make sense to REPEAT a link into a caption (of course if you need it, fine, but not REPEAT it). I mean if I have an article on the Battle of Waterloo and Napolean was linked at the beginning of the article, do I really need to wikilink the term in a painting of him?)

6. It really doesn't make sense if you have repeated images of a subject (let's say there are two paintings of Napolean, would I wikilink in both image captions? Or is there some parallel and separate train of linking just for the captions?)

96.238.184.111 (talk) 18:44, 14 October 2011 (UTC)

I would observe that infoboxes and image captions seem to live in a cocoon of their own by some sort of historical consensus. Linking practices within these are usually abominable, but to open up a battle front where there is litle hope of winning is just too stressful. --Ohconfucius ^¡digame! 02:28, 15 October 2011 (UTC)
I don't think the second half of Ssilvers' edit encourages wikilinks in captions. It encourages wikilinks when the only previous link was in a caption. For instance, in the Waterloo example, suppose the reader wants to know what else Napoleon did before Waterloo, so he is looking for a link to the Napoleon article. Do we expect the reader to find that link in a caption at the top of the Waterloo article? Or as Ssilvers would prefer, do we also link the first mention of Napoleon in the main text, regardless of captions? Art LaPella (talk) 06:23, 15 October 2011 (UTC)

I also read the Ssilvers edit as pointing out exceptions to the guideline to link only from the first occurrence of a term. Image captions are perceived as a separate thread from the main text, so a link from a caption shouldn't count as preceding a link from the same term in the text. This and the other suggestion (in a long article, the body is perceived as a separate thread from the lead) are already common practice, so I'd favor restoring the edit. (It was indeed not a mere copyedit, though.) On the idea of never linking from captions, that's a whole separate topic. —Ben Kovitz (talk) 21:00, 15 October 2011 (UTC)

Salient relations

I'm surprised that the article doesn't already include more about "salient relations between topics". I understand this to be the main principle explaining why wikilinks in the text are important and for distinguishing between good, over-, and under-linking. The article currently contains a tidbit about the "knowledge tree", but not much else. Here's an attempt to explain the missing concept. Apologies for the length. If I had more time, I'd write something shorter.

The link structure is itself part of WP's summary of all recorded knowledge. The links tell which topics have some salient relation to any given topic. Salient here means standing out from all other topics' relations to the target topic due to playing a distinctive or important role—standing out relative to the great majority of relations that do not stand out.

Nearly any statement of facts about one topic must be expressed in terms of other topics. For example, from an article about a gardener:

Robinson began his garden work at an early age, as a garden boy for the Marquess of Waterford at Curggaghmore. From there, he went to the estate of an Irish baronet in Ballykilcannan, Sir Hunt Johnson-Walsh, and was put in charge of a large number of greenhouses at the age of 21. According to one account, as the result of a bitter quarrel, one cold winter night in 1861 he let the fires go out, killing many valuable plants.

WP has articles about many topics mentioned here: garden, work, age, the Marquess of Waterford, estate, Ireland, baronet, greenhouse, testimony, dispute, winter, and more. Almost none of these play a distinctive or important role in describing Robinson. In fact, Robinson is notable because he revolted against highly artificial gardens such as require plants grown in greenhouses. That's why this story was included in the source.

So, the topics with a salient connection from Robinson are simply garden, greenhouse, and Sir Hunt Johnson-Walsh. Robinson plays no particularly distinctive or important role in relation to the others. For example, this story is not a special or outstanding fact about boys or boyhood in general. But Robinson did play a distinctive role in relation to gardening, greenhouses, and the life of Sir Hunt Johnson-Walsh. (Marquess of Waterford is actually a title, which many people have held, and the fact given is rather vague, so it only has borderline salience in relation to Robinson.)

To make those relations explicit, the Robinson article should contain links to those three topics, and none of the others.

Abstractly: If you need topic B to state a fact about topic A, then B is somehow relevant to A. But normally the relation is salient only if A is somehow distinctive or outstanding in its relation to B. To put that another way, link from A to B only if the existence of the link means something to B. If that seems weird, notice that it's normally true automatically if B is essential for saying something important about A. So, you can get the same effect just by linking from the most important terms needed to say the most salient facts about A. That's the easier way to think about it.

There can be other reasons for linking, of course. This is simply the main one. And of course, linking just from the first occurrence of a term is enough to do the job. There's no need to spoil readability by turning half the page blue.

Generally, we do not link simply in order to make it easier to navigate to other topics. If we wanted to make it as easy as possible to get to other topics, we would just make MediaWiki link every word or phrase to the article about it, without highlighting some links in a distinctive color. Just click anything and you're on it. Our actual policy, though, gives links a visual salience that represents their logical salience. The differing ease of navigation, where only the saliently related topics are easy to navigate to and the rest are not, might be the greatest benefit that the link structure provides readers. The easiest, most natural way to browse Wikipedia is to click along the links in the text. A reader traveling along the links effortlessly finds a path shaped by a combination of the relations that exist within the source materials and the reader's particular focus.

Nor do we link simply because we think most readers don't know what a word means (WP:DICT). A reader can consult a dictionary or type an unfamiliar word into the search bar. A link leads to further information; a definition is part of that, but an article contains much more information than a definition (including further links from the linked-to article). The presence or absence of a link derives from the relations within the source material, not from an attempt to second-guess the reader. That is why the links are so valuable—to readers. Readers can use the information contained in the links any way they like. Our job as editors is simply to make the links provide that information.

Some less-common ways of using links: Links that express salient relations between topics make the "What links here" link in the nav bar provide meaningful information. (It sure could be improved, though.) Robinson might not be mentioned in Sir Hunt Johnson-Walsh, but the fact that he links there is significant; maybe he should be mentioned there. Data-mining software can crawl and analyze the link structure. If wikilinks are pictured as arrows between topics, natural "families" of topics should appear as clusters in the picture, something like maps of scientific fields derived from citations in journal articles.

Overlinking damages the link structure because it corrupts the information it contains. It's actually slightly worse than indiscriminately adding facts to an article, because too many gratuitous links spoil the kinds of things you can do when you look at "What links here". For example, Ward Cunningham doesn't show up for a long time on Special:WhatLinksHere/Wiki. He's drowned out by topics like Mick Fleetwood, who links to Wiki because an external link about him goes to a wiki, and someone gratuitously pointed that out and linked to Wiki even though there is nothing important to say about Mick Fleetwood in connection with wikis. Maybe the editor thought "wiki" would be an unfamiliar term to many readers, and tried to be helpful.

Even with the current high level of overlinking, though, "What links here" searches often turn up interesting facts about less-central topics. For example, every notable virologist shows up automatically at Special:WhatLinksHere/Virology. There are lots of kinds of searches you could do, most of which probably haven't even been thought of yet, all because the link structure faithfully reflects important relations within the source material. What links here is only a convenient illustration. The main benefits are the way logically salient relations are made visually salient and help shape the path of a reader's browsing.

Hopefully that persuades someone to share my opposition to most guidelines that make the kind of topic the criterion for linking or not linking to it (such as "link to jargon terms, don't link to common terms"), rather than the kind of relation that one topic has to another. This is also why I think the relation of "critical to defining the topic" nearly always does merit a link. Salient relations (standing out) can only be held by a small minority of topics, by definition. So, I think linking only along salient relations means that relatively few words will be linked from, and when there is a link, it will mean something.

Would anyone object if I add something about salient relations to the article, in a prominent place?

—Ben Kovitz (talk) 10:25, 16 October 2011 (UTC)

Of course they will. [wink] I agree with much of what you're saying here, but I'm not prepared to go quite as far as you in the area of "selectivity". One small thing: you say, "Nor do we link simply because we think most readers don't know what a word means (WP:DICT)", but isn't that exactly what Wiktionary is there for? WP:DICT stresses that we shouldn't have dictionary definitions for articles, not that we shouldn't link from WP articles to WT entries (indeed, it even says in the lead, "the two often link to each other"). If I'm mistaken, please point to the part that says that. Oh, and to answer your final question more directly: You probably should post it here before you try it out in the article. - dcljr (talk) 10:14, 17 October 2011 (UTC)

I don't disagree with everything Ben says, and I do agree with Dcljr's post. Ben, I think you're taking one dimension of the wikilinking system—the navigational—to an extreme, without accounting for its other dimensions. To raise one point that's been done to death on this talk page, wikilinking needs to be rather selective if it's to aid the readers. We could link every single word via a system that didn't colour the linked items (like my online Encarta dictionary—works well for that), but that would remove our significant service to readers in isolating what, in our knowledge of the topic and the catogory, we judge is the most useful, the most focused on the topic. Relatedness per se is only one criterion.

Tony (talk)

10:27, 17 October 2011 (UTC)

Tony, it sounds like we agree: the thoughtful selectivity is what makes the links especially valuable. And, I wouldn't say that navigation should crowd out all other concerns. There can be other reasons for linking, too. I think we all know that our topic—everything—is filled with so many irregularities that no rule nor single principle can cover them all. —Ben Kovitz (talk) 10:59, 17 October 2011 (UTC)

Yes, Tony, please reread Ben's post and yours. Is that really what you wanted to say? You didn't get Ben and me confused, did you? - dcljr (talk) 06:52, 18 October 2011 (UTC)

Dcljr, can you tell me what "selectivity" you think goes too far? It's likely that I was unclear; the above was "typing out loud" to find out what's been bothering me about the delinking campaign. One clarification I can make right now: the above is only about links between articles in Wikipedia, not links to Wiktionary. Regarding WP:DICT, I meant that we don't link (within WP) in order to define a word we think the reader doesn't know, we link in order to lead to further information about a related topic. Doing the latter often happens to accomplish the former, of course. —Ben Kovitz (talk) 10:50, 17 October 2011 (UTC)

Ah, I see that I completely misread your meaning in the line I quoted. Mea culpa. Although... that being said, I wouldn't necessarily rule out linking a somewhat obscure term for those who need to find out simply what it means (as I have done with the link above). I mean, obviously this will often be an indication that the article text needs improving (if you find yourself linking for the purpose of dictdefs), so no one has to take up arms in opposition to what I've just said. I'm not pushing it as a guideline... Sorry if I seemed somewhat dismissive with my short response; I didn't have a lot of time to devote to it, plus I figured others would comment, anyway, so...

Let's see... the statements that I would take issue with in various ways (and to varying degrees) include:

"To make those relations explicit, the Robinson article should contain links to those three topics, and none of the others." Without getting bogged down in which specific terms you recommended be linked and which you didn't, the criterion that lead you to those choices is only one of many that could be used, so it's not necessarily true that only those terms should be linked. This is more or less what I was alluding to with my "selectivity" remark. (Interestingly, this sounds very much like Tony's comment, except the way I read it, he got it exactly backwards and ended up restating your points as if he were disagreeing with you. [?!] Talk about "violent agreement".....)
"link A to B only if the existence of the link means something to B" This is interesting advice, but I think too restrictive. Change "only if" to "if" and I'll agree with you. IOW, I suspect that this criterion works better to encourage good links than it would for discouraging (only) bad ones.
"There can be other reasons for linking, of course. This is simply the main one." Again, those "other reasons" are not necessarily of lesser importance. It depends...
"The differing ease of navigation, where only the saliently related topics are easy to navigate to and the rest are not, might be the greatest benefit that the link structure provides readers." Hmm. I don't entirely buy this because... (see next item)
"A reader traveling along the links effortlessly finds a path shaped by a combination of the relations that exist within the source materials and the reader's particular focus." Not if you haven't provided a link that they would have wanted to follow. In that case, it's not as effortless, and might even be downright frustrating for them.
"too many gratuitous links spoil the kinds of things you can do when you look at 'What links here'". Frankly, unless some revolutionary advance in usability is made (selecting namespaces is a good start, but not sufficient), WhatLinksHere is going to remain almost useless (to human readers) for all but the least linked-to articles. This is decidedly not an argument to reduce the number of links in our articles. It's a technological problem that requires a technological solution.

- dcljr (talk) 06:50, 18 October 2011 (UTC)

About not providing a link that the reader wanted to follow: As terrible as this sounds, I think making certain browsing paths frustrating is actually a good thing. The ease of access to closely related articles (through links, especially in the lead) provides a set of "paths of least resistance" that the user explores while quickly browsing. The links reflect the most important close relationships within the source material, so you tend to view a natural cluster of topics as you browse. Your exact browsing path, however, follows only a small subset of the links, reflecting your purpose for browsing that day, your interests, the exact background knowledge that you have and that you lack. This happens automatically, without us trying to anticipate any of these things about the reader.

This is really pretty amazing, if you think about it. I regularly bring myself up to speed on new subjects by taking little tours suggested by the links (though not consciously). For example, a biologist friend told me that a certain kind of gene is relevant to something I'm doing. The first couple articles about genes I looked at were way out of my depth, but I skimmed and browsed a bit, and in maybe half an hour, I found myself getting a basic overview of a certain mechanism in developmental biology, which I did not even know I was looking for but which was the reason my friend thought the gene was relevant. A little tour of genetics, starting in some highly technical place I'd never heard of before, customized to my one-of-a-kind reason for being interested! The point I was trying to make above is that such tours reflect both the logical connections within the subject and each reader's unique interest and background.

On the other hand, if a reader wants to do something "against the grain of the links," like visit every Greek-derived word in the article, that should feel like going against the grain if Greek-derived words play no special importance in the subject matter. But, if, for some weird reason, most of the Greek-derived words do have links, that should reflect something peculiar, interesting, and real about the subject, like way early botanists drew heavily on Greek. (Or is this not a good example of what you meant? If not, can you tell a better example of a reader finding a path frustrating?)

My main point above is that we as editors shouldn't try to anticipate any of this. Our job is to write prose that's faithful to each topic and to make links that are faithful to the logical relations with other topics. That's what makes the magic happen.

On the other hand, pushing a single factor to the exclusion of all others is unrealistic and silly, and I wouldn't want that. Sometimes, sure, link to a peculiar word not so much because there's a salient relation, but because there's no real doubt that most readers need the word explained. That injects a flaw into the link structure, but sometimes usability has to trump "salient relations" or whatever we want to call it. My long "typing out loud" above was an attempt to understand something important but not much covered about links, not an attempt to cover every factor to consider (well beyond my expertise, anyway). I guess I didn't make that clear.

About the difference between "links to" and "links from": "if" rather than "only if" is fine with me. My point there is that it's OK if A links to B but B doesn't say anything about A or shed any obvious light on A. The fact that B is needed to state a particularly important fact about A means that A is outstanding in some way in relation to B. That's the right kind of relevance, and it should be included in the "path of least resistance". (That sort of link made a big difference on my "customized, peculiarly focused tour of genetics". I had to jump "up" in the topics a lot.)

About the current lameness of "Wikipedia:What links here": This is not nearly as important as "paths of least resistance" and visual salience reflecting logical salience, but there are possibilities for automated use of the link structure that no one has thought up yet. Actually, if you google, you'll find a bunch of creative uses of the link structure. Obvious ideas that occur to me now are simply systematic ways of paring down the list of "Wikipedia:What links here" in custom queries: leave out user pages, include only links from people, include only pages that also link to X, etc. It should be easy to make a query to find all people with some important connection, say, to Greenhouse, or all mathematics articles with some connection to Enrico Fermi (not a mathematician per se, but I've heard he had amazing skill at mentally restructing a calculation to quickly get an order-of-magnitude estimate, and I'm curious about that). So you say, "Show me all articles that link to both Mathematics and Enrico Fermi."

The main thing is: don't try to anticipate any of this! Don't skew the link structure to reflect your own theories about what's more and what's less useful. Users decide what's useful; editors decide what's salient. Don't make rules based on the kind of topic or word being linked to; phrase the linking guidelines to focus on relationships between topics. Modulo the inevitable exceptions and competing concerns, of course.

—Ben Kovitz (talk) 11:51, 18 October 2011 (UTC)

That is a take on salience and overlinking that I can largely agree with. If only that's how most articles are linked. The problem is that en.WP is chock-full of links that one or another editor think some others may think useful, thereby using it as 'navigational' or to provide a dictionary definition. I have catalogued a substantial number of words that are seemingly linked as a matter of course, often without any thought. It is the practice which is detrimental. Such links can be removed advantageously in >90% of cases; for some of the more common terms, that percentage rises to 99.9%. For every hundred links to a common term, there are perhaps 5 that could be advantageously rendered salient to the subject. Therein lies the challenge. --Ohconfucius ^¡digame! 07:16, 18 October 2011 (UTC)

Ben, should wikilinking shoulder the entire navigational burden? WP has navboxes and categories too. And a "See also" section at the bottom of each article. $Tony (talk)$ 12:07, 18 October 2011 (UTC)

Of course not! That would be crazy. Is there any disagreement about this? —Ben Kovitz (talk) 12:16, 18 October 2011 (UTC)

Well actually, pace Ohconfucius, I think the bigger problem is that a small group of editors have taken it upon themselves to remove a large number of links - often by automated scripts that remove links to certain terms wherever they appear, regardless of the context on that page - because they assert, definitively, that they are "not useful", or even, somehow, "detrimental"; as well as on the basis of the now comprehensively rebutted "dictionary" issue. Personally, I see a lot of articles that I think are overlinked, but I think the technique for dealing with that primarily needs to be through manual copyediting; the principles under which that is done should be as per the existing guidelines, which stress relevance and navigation as being relevant criteria for linking, and which very specifically do *not* mandate removal of so-called "dictionary" links or the blanket removal of so-called "common" or "well known" terms; and especially not the arbitrary and selective removal of totally random terms as often happens on the basis of some tortuous reasoning.

As for the "salient" option, I can kind of see the value of that. But equally, I think we're slightly dancing on semantics to some degree when we start comparing salient, germane and relevant etc (the latter of which is already accounted for in the guideline). I think the example of dog offers a good example, which I think I have used before - the guidelines as written, common sense, the expectation of most readers and editors surely, and any additional/alternative requirement of germaneness or salience would all allow, or even require, a link to the "dog" article in the lead of the Doberman article. However, that part of an article that recounted the occasion when something happened to them as a famous pianist took their dog for a walk would not link to dog. The problem for me is that some seem to be arguing against the first point.

Finally, as to navigation, I agree we shouldn't be looking to simply navigate everywhere at every opportunity, but I can't see the problem with offering a considerable amount of flexibility and options to readers (falling some way short of "link everything!", which is never going to happen anyway, even if anyone really wanted that, not least because pages don't exist for most words, WP not being, er a dictionary). Those readers are all different of course, coming here with different levels of interest and knowledge and for different reasons, and with different online reading styles and habits etc etc. I don't get how some editors can assert the right to decide what people - supposedly as an undifferentiated whole - are looking for or, worse, what they ought to be looking for in terms of links to other articles. I neither accept that as a principle nor do I see how you could assess how that should actually work in practice and in the context of individual articles. Focusing more on the relevance (or even salience) requirement - which, as noted, currently exists in the guideline - offers a clear, fairly objective standard as to how to link between topics. Individual editors making their own judgments about what is "common/well known" or "useful to the reader" - whether this is done in an extremely reductive fashion, as some do currently, or in a more maximalist and generous way - very definitely does not. N-HH talk/edits 14:23, 18 October 2011 (UTC)

I didn't read all of that, but IMO the main consideration in deciding whether a paragraph of Article A should link to Article B should be:

the conditional probability that someone will want to read (not necessarily straight away, remember about tabbed browsing) the lead of Article B (or, for section links, the linked section) given that they have just read said paragraph of Article A.

Other considerations should be:

the principle of least astonishment (hence, do not write Glasgow, Leeds, Liverpool, Manchester and Sheffield unless it's obvious why some are linked and some aren't);
a roughly consistent link density (though if by following the first two principles you'd still end up with a paragraph with more than 3 [less than 0.3] links per hundred characters, odds are that it is too terse [too vacuous] and you'd better explain the stuff more explicitly [cut fluff out]).

All the rest can only be rules of thumb about how to best satisfy these desiderata. ― A. di M. 20:47, 27 October 2011 (UTC)

My post or the whole thread? Anyway, apologies for the length of what I said at least, but the conversation here rarely moves forward, so one finds oneself constantly having to repeat - and explain in ever greater detail - what should be rather simple and readily acknowledged first principles, and rebuttals of claims about what the guidelines say, whether I and others want to "link everything", where consensus may or may not be and why wp:dictionary has nothing immediately to do with linking issues etc etc.

Anyway, I mostly agree with what you say, but, on the first point, I'm always wary when we start trying to assess too analytically how people might want - or even ought, as some put it - to navigate between pages. As I said in the last few sentences of my essay, I just don't think it's possible to do that in any satisfactory way with such a varied readership; nor can it lead to consistent principles for linking, since every editor will have their own views on what others supposedly might want - the criteria need a strong objective focus as well. There's a risk of each case falling down to "well someone might be interested in that other article, hence we'll link" vs "well who would really, how many of them? Let's delink". As, of course, happens currently. N-HH talk/edits 15:02, 28 October 2011 (UTC)

I think going by probability that a reader will want to click a link is a very bad idea. Perhaps it should be discussed in a separate section. The links provide information when they correspond to important logical connections between topics. Our job is to provide information to readers, not reflect what we think readers are doing with that information. This is explained in more detail above. (Sorry for the slow reply.) —Ben Kovitz (talk) 14:56, 18 May 2012 (UTC)

Germane –> "relevant and appropriate"

There's been a slight tussle by two editors over whether this long-standing term germane—approved at a massive RfC more than two years ago and inserted officially by Ryan Postlethwaite, an ArbCom clerk—should be changed. I'm uneasy that it hasn't been put to this forum, let alone gained significant support. But the editor pushing this does have a point, that germane does mean relevant and appropriate in the dictionary, and is a less-well-known word. It's neither here not there in terms of the meaning, but might be plainer English. If other editors here don't object to the change, perhaps it might be acceptable. At the moment, it's at WP:BRD status, as far I can see. $Tony (talk)$ 07:49, 20 October 2011 (UTC)

I see some problems with the edits made:

The previous version with the explanation ("relevant and appropriate") in parentheses made it clear (in my opinion) that the explanation also applied to subsequent uses of "germane". This is lost if "germane" is replaced by "relevant and appropriate" in one instance.
The current version omits the explanation completely. This also removes the emphatic "and"; so the implication "not only relevant but also appropriate" is lost.

So, on balance, I would prefer restoration of the original, including the parenthetical explanation that clarifies the precise meaning of germane that is intended (which is, apparently, more than the Wiktionary definition "related to the topic being discussed or considered"). --Boson (talk) 10:05, 20 October 2011 (UTC)

I thought my edit was well-thought and explained. I was a tad miffed that it was undone without so much as a word of explanation, and even more miffed at the condescending tone in which it was suggested I "look up" a word I obviously understand. I saw no point to using the word "germane" when an explanation of the word was given immediately after it. I was entirely unaware of the "massive RfC", and if editors feel "germane (revelant and appropriate)" is ideal, I certainly have no objection. I do object to "germane and topical", and don't understand the grounds on which that change was made. Joefromrandb (talk) 17:34, 20 October 2011 (UTC)

Is there any objection to changing from "germane and topical" to the original "germane (revelant and appropriate)"? Joefromrandb (talk) 11:36, 22 October 2011 (UTC)

Fine by me. I just objected to removal of the word 'germane'. --Ohconfucius ^¡digame! 12:03, 22 October 2011 (UTC)

Joe, I'm confused: you removed and reinstated a number of times. Please don't remove "topical" unless there's consensus. I've reverted to the long-standing.

Tony (talk)

13:58, 22 October 2011 (UTC)

Reverting to the long-standing version was what I meant to do. Sorry. Joefromrandb (talk) 15:15, 22 October 2011 (UTC)

Linking United States

WP:OVERLINK provides: "Avoid linking the names of major geographic features and locations, languages, religions, and common professions." I interpret this to mean we should not be linking to United States (a major geographic location) in the body of articles. I made a series of edits today delinking the term in a number of articles by an editor (User:Jweiss11) who regularly links the location. An example of my delinking is found at this diff. The editor in question has objected to my delinking the "United States." I will hold off on delinking for now, but would appreciate input from the experts here. Should the "United States" be delinked in examples such as the diff above? Cbl62 (talk) 06:55, 27 October 2011 (UTC)

It should definitely be unlinked in those contexts; and in rare cases where a link might be appropriate, it should generally be to a section or a daughter article.

Tony (talk)

07:14, 27 October 2011 (UTC)

It might be helpful to readers if there was information on where the subject originates from. Such would allow a focussed link without resorting to an article where the relationship is not of the first order of separation. --Ohconfucius ^¡digame! 07:28, 27 October 2011 (UTC)

I'm curious why then today's feature article, Rudolph Cartier, sports a link to Austria in the lead. Jweiss11 (talk) 19:26, 27 October 2011 (UTC)

Not sure whether that link is so vital, but at least, unlike in the Joe Reilly case, none of the other articles linked by that paragraph is Austria-related but more specific. (And I don't think screenwriter should be linked when Nigel Kneale is, or George Orwell when Nineteen Eighty-Four is, so I'm gonna remove those.) ― A. di M. 20:14, 27 October 2011 (UTC)

On Joe Reilly, the link to Boston College obviates the need for United States to be linked? Jweiss11 (talk) 22:59, 27 October 2011 (UTC)

I'm not sure if I would link Austria either, but I would surely link Suriname for instance; not a major geographic location. Austria has 8 million people and 80,000 km², so readers might be hazy about where it is. The United States has 312 million people, 10 million km², and speaks the language of English Wikipedia. It's very unlikely that someone reading about Joe Reilly would suddenly want to read a general article about the United States that doesn't mention Mr. Reilly at all. Art LaPella (talk) 00:40, 28 October 2011 (UTC)

Indeed. Let's not be lulled into thinking that readers routinely click on links: they don't, yet editors seem to plaster them all over the place in the hope that they will. Judicious linking that avoids those that overlap in topic (Orwell and 1984, for example) should not be doubled up unless there's a compelling reason. Usually, the more specific link should be retained.

Tony (talk)

02:30, 28 October 2011 (UTC)

FWIW, Guyana speaks the language of the English Wikipedia, too. ― A. di M. 09:38, 28 October 2011 (UTC)

I'm old enough to think of Guyana and Suriname as new names for British and Dutch Guiana, but I don't get the point. Art LaPella (talk) 23:21, 28 October 2011 (UTC)

It surely depends on the context of the article and text in question. Let's not forget that the actual, more complete - and qualified - wording of that part of the guideline is "Unless they are particularly relevant to the topic of the article ... avoid linking the names of major geographic features and locations" (my emphasis). However, there no clear definition of what constitutes a "major" geographic location - arguably the US if anything falls within that for most people, but where is the line - France? Belgium? Nigeria? Iran? I can't see how doing maths about population numbers is really going to help make that sort of judgment; nor is the question anyway about whether people might know or not know "where" a specific country is. The links to country articles do not simply lead to geolocation co-ordinates or a one-line dictionary definition, eg "country in Europe" or whatever. Regardless of that, even for "major" locations, however defined, there is no blanket bar, given the qualification about relevance that precedes the suggestion to avoid (and note "avoid", not "do not, ever").

More specifically, there is no agreement or consensus among editors at large, or among the "experts" on this page, as to whether any and every link enabling navigation to the lengthy and detailed Wikipedia entry on the US - which I believe is one of the site's most visited pages - from genuinely relevant and related pages, should be universally struck out. With personal biogs, as in the example cited above, one could probably argue the toss either way for relevance and hence retention when it comes to countries of birth, although this should be consistent and considered regardless of the actual country in question (personally I'm kind of indifferent); but when it comes to geo-political articles, such as, say those on US states or neighbouring countries, the case for having and/or retaining a link to "US" seems much, much stronger. Almost insurmountable, in fact. N-HH talk/edits 14:36, 28 October 2011 (UTC)

Of course there's no line; I thought the point was that the previous practice of automatically linking all countries is discouraged. Just being a country doesn't automatically make a word more linkable than any other word. And I thought that for all the noise over this issue, there was consensus even from you that most randomly chosen "What links here" links to the U.S. should be unlinked, and those on U.S. states and neighboring countries should be linked. Art LaPella (talk) 23:21, 28 October 2011 (UTC)

A di M, Guyana might list English as its official language, but that doesn't mean that any great proportion of the population are native anglophones, nor that many of them speak standard English. Are you suggesting that Guyana-related articles shouldn't be copy-edited to standard English of a chosen variety, even if written by Guyanans? And I've never seen an India-related article where editors have objected to the harmonisation of the language with one of the major varieties (and not Indian English)—and India has a lot of native anglophones in its middle class. No one has ever suggested that en.WP's rules on ENGVAR, on date formatting, and other related matters are anything but racist, favouring ancestral native-speakers of English over native speakers of currently outlying varieties (Indian English is just one, Guyanan another), and over second-language speakers. But when I've brought this up at the MoS as a matter of interest, there's always an embarrassed silence. I recently saw a film located in Liberia, a so-called anglophone country in Africa (where returnees from the US have ruled the roost, often unfairly, for a long time). You needed subtitles, although repeated listenings to a line of dialogue would have enabled you, often, to work it out. (Such places have an acrolect, mesolect, and basilect.) But on the other hand, I suppose it's natural that the ancestrals might rule the roost at the WP they believe they own, and it's often a matter of developing rules that will keep the peace among them (i.e., US vs UK practice—who cares what date format or spelling they use in Guyana).

On the linking of country-names, it's more than that: it's the reality that almost everyone in the world has a basic familiarity with English-language culture and its knowledge-base, at least enough to think carefully about whether linking these country-names anywhere is useful enough to outweigh the dilutionary and visually disruptive effects. $Tony (talk)$ 03:05, 29 October 2011 (UTC)

My mention of Guyana was in reply to The United States ... speaks the language of English Wikipedia as it that mattered at all when deciding whether to link a country. (Maybe the Isle of Man would be a better example?) But since you mentioned it: India has one billion people, a quarter million of whom are native English speakers, so there's no way it has “a lot of native anglophones in its middle class”, for any reasonable definition of a lot and middle class. And some Guyanese might speak a creole in informal situations, but still English is the only (or main) language they're literate in and so the English Wikipedia is the first one they'll go to; I still think articles about Guyana should be written in standard Guyanese English (essentially British English, as far as writing is concerned). (And I've heard of native English speakers having trouble understanding certain scenes about Trainspotting without subtitles, too, but that doesn't mean Scotland shouldn't count as English-speaking, does it?) I still have trouble with this “ancestral” thing: a minority of Irish or American people have English ancestry; in most cases, their (great-)grandparents had to learn English as a second language to avoid corporal punishment by school teachers or marginalization). As for “almost everyone in the world has a basic familiarity with English-language culture and its knowledge-base”, after the n-th time I told someone I was going to/had been to Dublin and they asked me “That's in Germany, right?”, I'm no longer sure of that. *sigh* ― A. di M. 10:58, 29 October 2011 (UTC)