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.

I've restored the merged version which had consensus. Those of you wishing to dilute the language supportive of linking will have to find something else to do. Perhaps trim your toenails or something. —Locke Cole • t • c 16:21, 9 May 2009 (UTC)

Do you find this tone to be helpful and conciliatory, in general? -GTBacchus^(talk) 16:26, 9 May 2009 (UTC)

If I didn't laugh I'd probably cry, so I'll take humor, thank you. As for the rest, well... it's telling that the usual suspects are all here revert warring to try and force their point of view, again. I don't know, which do you find more helpful: a) making draconian changes with little discussion only a week after a proposed merge was completed, or b) trying to stabilize the page back to the version which had consensus? —Locke Cole • t • c 16:39, 9 May 2009 (UTC)

What I really find helpful is characterizing people's positions in a way that they would agree with, seeing questions from multiple perspectives, and avoid ad hominem comments of any kind. Thanks for asking. I also prefer my on-wiki humor spite-free. -GTBacchus^(talk) 16:42, 9 May 2009 (UTC)

I'll do all the things you mentioned as soon as they start doing them for me. Otherwise, what's the point in trying to see their point of view if they consistently ignore mine? I see I wasn't blindly reverted and there seems to be reasonable editing going on now, so perhaps my revert wasn't as bad as I'd thought it would be. —Locke Cole • t • c 16:49, 9 May 2009 (UTC)

I agree that your revert was fine, and probably a good idea. The notion of "I'll [X] as soon as they do," is why we still have wars. I can't think of a position that I respect less. The best thing to do if people are being unreasonable, IMO, is to go and find reasonable people and bring them to the table. Hi. Nice revert. It's a lovely day. -GTBacchus^(talk) 16:53, 9 May 2009 (UTC)

Locke Cole is, I think, understandably irritated; the revert-warring on this issue has gone to ArbCom. I view this as a tempest about a fundamentally unimportant issue, although one on which several people are disruptive; Locke has an actual policy preference, which is worse. Please bear with him. Septentrionalis PMAnderson 17:08, 9 May 2009 (UTC)

I hope I'm not busting his chops too much. I respect Locke Cole entirely; that argument however, I have issues with. Frustration is always understandable; it's entirely normal and healthy to get frustrated w/ others. Still, we've got to keep our eyes on the prize, namely resolving the dispute, not somehow deepening it. Is that fair? -GTBacchus^(talk) 17:16, 9 May 2009 (UTC)

I don't even think there's a dispute - at least, not one that anyone's succeeded in formulating in any reasoned way. We should be working on improving the guidance here to make it clear and rational. Thankfully we seem to have settled down to that now. I really don't know why the fighting had to flare up in the first place - if people had just come here saying "I think X should be changed to Y because Z", as is the normal WP way, instead of trying to overrule consensus by force and personalize the argument, then we would have got to the same place as we are now with a lot less unnecessary ill-feeling and wasted time. Wishing everyone a pleasant weekend,--Kotniski (talk) 17:28, 9 May 2009 (UTC)

That's not what I expected to hear from the editor who got my attention with the edit summary: "appalling disruption - why do we have to put up with this? can admins really not see what's going on?" By my definition, that's a dispute. -GTBacchus^(talk) 17:34, 9 May 2009 (UTC)

Well, that dispute (complaint) was more about behaviour, and is hopefully now moot. The subject of discussion here now is the wording of the guideline, which can hopefully be kept separate from disputes of that kind. Anyway, I really am off this time, see y'all--Kotniski (talk) 17:43, 9 May 2009 (UTC)

Have a good'n. :) -GTBacchus^(talk) 17:47, 9 May 2009 (UTC)

I've tweaked the section as follows:

• trimmed excess text from the preamble regarding "overlinking"; there is no need to repeat the idea of redundant links three times in the same paragraph
• removed the "screen reader" bit, in the absence of verification of the claim (there are a lot of page elements that muck up screen readers; not sure how links do, though)
• placed "what to link" first, as that is the natural way of describing it (i.e. "do this, don't do that"); it also flows more naturally into the "link density" section

--Ckatz^chat_spy 16:39, 9 May 2009 (UTC)

That's a help. I think it can be condensed further, and I moved the point out from the bottom. Septentrionalis PMAnderson 16:48, 9 May 2009 (UTC)

There are helpful improvements in what CKatz, LC and Anderson just did. Tony (talk) 17:51, 9 May 2009 (UTC)

Anderson, if you persist with your unilateral disruption to this guideline using the merger of BTW into this page by agreement (twice) as some kind of excuse, then unmerge it. Please remember that WP:CONTEXT was also merged into this page as part of the agreement, and significant compromises were made on that basis.

There is no consensus for your edits today. If you're so attached to BTW, bring it back from the grave again. Fine. Your attempts to dilute this page, along with your three-year-long campaign to dilute MOSNUM and MoS main, are as usual transparent. This is not your playground. Please be reasonable and decide which option you want: the agreed-to version (without repetition, please), or the unmerged. I'm sure Kotniski would be upset that his fine work in preparing and negotiating the merger will be wasted, but that is a price we'll have to pay, I suspect, to stop your meddlesome and disruptive playing around with the text and its status. Tony (talk) 14:43, 9 May 2009 (UTC)

There is no consensus for anything on this page; there is only the revert-warring of three extremists. That's why its claim to be guidance is fraudulent.

But I shall indeed unmerge BTW, and see who reverts me there. Septentrionalis PMAnderson 14:47, 9 May 2009 (UTC)

Oh, I wouldn't bother. Besides, the stability of style guides is more important. Tony (talk) 15:04, 9 May 2009 (UTC)

He did bother, and I've reverted it for reasons explained below. Rd232 ^talk 15:22, 9 May 2009 (UTC)

How is it that people are so upset over whether or not BTW is merged here, and how? I've asked, and nobody has yet shown me any practical upshot to the question. Is this dispute really entirely content-free? -GTBacchus^(talk) 15:06, 9 May 2009 (UTC)

No, it is almost entirely about content; WP:Build the Web, now restored makes the case to make links; this page almost entirely omits it, and makes the case against making links. Ideally, this page would include both, and when the merger was made the second time it did; but the three other editors here removed most of that content. Septentrionalis PMAnderson 15:11, 9 May 2009 (UTC)

Thank you. You just did in six words what others have entirely failed to do. -GTBacchus^(talk) 15:28, 9 May 2009 (UTC)

• As long as I have your attention, WP:Build the web (protected) should redirect to WP:Build the Web, not here. Septentrionalis PMAnderson 15:11, 9 May 2009 (UTC)
  • Um... that's two of the same link. I'm not sure what you intended by that. -GTBacchus^(talk) 15:27, 9 May 2009 (UTC)

OK, this is not good. We achieved (eventually) consensus on merging WP:BTW here, taking just the core idea because the related points were mostly here already. At this point there should be no question of one lone editor undoing this (Pmanderson and Septentrionalis are the same editor, which I didn't immediately realise) or slapping on dispute tags. Instead, if there are particular BTW points that people think need to be here, then they should be identified and debated. (More important for me, though, would be overhauling the page to rationalise it and clean it up and make it clearer.) Rd232 ^talk 15:19, 9 May 2009 (UTC)

Hi. What's your response to PMAnderson's claim: "WP:Build the Web, now restored makes the case to make links; this page almost entirely omits it, and makes the case against making links." Do you think that this page, as it currently stands, makes the case for linking and the case against it, or just one or the other? -GTBacchus^(talk) 15:30, 9 May 2009 (UTC)

Both, obviously. I can't understand how anyone could reasonably read it differently. In any case, this page makes the case for making links far better than BTW did.--Kotniski (talk) 15:34, 9 May 2009 (UTC)

Wikipedia:Linking#Overlinking and underlinking. Which overlaps a bit with General Principles - one of the issues I was thinking of when I talked about rationalisation. Rd232 ^talk 15:41, 9 May 2009 (UTC)

(<---)When I look at the page, under "General principles," I read the following statements:

Those statements are followed by a ten-item bulleted list. Five items say when not to link, two items encourage linking, and the other three are more technical, about where a link should go in special cases. PMAnderson, do you consider that to be inadequate, as far as encouraging people to link? -GTBacchus^(talk) 15:44, 9 May 2009 (UTC)

Yes; consider the paragraph to which Rd232 links, and to which he has just reverted: An article is said to be underlinked if subjects are not linked that are helpful to the understanding of the article or its context. However, overlinking[1] is also something to be avoided. A high density of irrelevant links makes it harder for the reader to identify and follow those links which are likely to be of value. Dense and irrelevant links also make pages less accessible for users making use of screen readers. Provide links that aid navigation and understanding; avoid cluttering the page with obvious, redundant and useless links.

It defines underlinking is a single sentence, and then had three sentences of repetitious rant againt overlinking.

He has also reverted WP:Build the Web from the undone merger, which before even the redirect could be fixed. Septentrionalis PMAnderson 15:50, 9 May 2009 (UTC)

Yeah, it's a bit repetitive, but I don't see that linking is discouraged. Only irrelevant links are recommended against, and that recommendation takes more words to unpack than the idea of under-linking, which is more obvious. It certainly doesn't read to me like a rant.

How would you edit that paragraph? -GTBacchus^(talk) 16:02, 9 May 2009 (UTC)

I'm not sure whence this sudden need to have the same number of sentences against underlinking as against overlinking, but I would drawn attention to the whole of the very first paragraph of the page, which says how important linking is.--Kotniski (talk) 16:07, 9 May 2009 (UTC)

• WP:UNDUE. Doubtless I fail to realize how that doesn't count in ocmparison to the Most Important Thing. Septentrionalis PMAnderson 16:31, 9 May 2009 (UTC)

UNDUE applies to articles - it says so right at the top of WP:NPOV of which it is a section. Rd232 ^talk 16:54, 9 May 2009 (UTC)

The idea behind UNDUE applies to any piece of writing in any context. The idea is more important than the written "rule". I think Pmanderson was citing the idea in an abbreviated way by linking the page. He'll probably correct me if I'm wrong. -GTBacchus^(talk) 16:59, 9 May 2009 (UTC)

No it doesn't. NPOV is a specific concept which applies to Wikipedia articles. Policies and guidelines reflect - however loosely - WP:CONSENSUS as demonstrated by discussion and practice, not WP:NPOV, which is based on balancing what different reliable sources say. We don't have reliable sources for what WP policy would be. And if the thought processes behind UNDUE can be applied at all in this context, it means that fringe views of a few editors do not need to be represented in policy - not that competing policy injunctions need to be balanced, BBC-style. Rd232 ^talk 18:03, 9 May 2009 (UTC)

I'm not talking about NPOV. I'm talking about good writing. One tries to maintain a fairness and balance in whatever one is writing, with the exception of writing in praise, or in criticism of something. This is a guideline, and should be written with balanced language. I don't see any fringe views going on here; I see reasonable people disagreeing, and talking past each other, to an extent. -GTBacchus^(talk) 18:10, 9 May 2009 (UTC)

Quite. Unless Pmanderson has concrete suggestions for redefining in minor or major ways what constitutes an irrelevant link, this is just hot air. Rd232 ^talk 16:13, 9 May 2009 (UTC)

Well, I hope he replies to my above question. -GTBacchus^(talk) 16:18, 9 May 2009 (UTC)

(ec)My suggestion was: An article is underlinked if subjects are not linked that are helpful to the understanding of the article or its context, or that many readers will want to see. Each link is a cost, and irrelevant links are pointless costs: many links make it harder to identify the links a reader will want to follow, and make it harder to screen readers to present the text at all. The second is one sentence, and makes all the same points, but this other gallant revert warrior disposed of it before it could be seen. Septentrionalis PMAnderson 16:31, 9 May 2009 (UTC)

That (second) sentence is rather opaque to me. You're introducing a notion of links as "costs", which I don't think many people will find intuitive. Can you do the same thing with more commonly understood language? I hope this criticism isn't off-base... Maybe I'll try and edit that bit. -GTBacchus^(talk) 16:35, 9 May 2009 (UTC)

I see what you mean: An article is underlinked if subjects are not linked that are helpful to the understanding of the article or its context, or that many readers will want to see. An article is overlinked when the number of links make it hard readers to identify the links they will want to follow, and hard for screen readers to present the text at all. Each link should offer some benefit to compensate for its presence - irrelevant links don't. This slightly changes the definition of overlinked, but is the best I can do for now. Septentrionalis PMAnderson 16:45, 9 May 2009 (UTC)

I like that. What do others think? -GTBacchus^(talk) 16:47, 9 May 2009 (UTC)

"or that many readers will want to see" is a meaningless subjective standard which is basically carte blanche to link anything. The rest isn't a substantial change, it's just phrasing (and I don't see it's an improvement). And it feels pointless to argue about phrasing when the guideline needs substantial rationalisation. See eg what was done at WP:EP (on talk and in the history there). Rd232 ^talk 16:50, 9 May 2009 (UTC)

That wording can be easily fixed, I think. If you see a change as unsubstantial, but it helps reduce heat in a conflict, then it's a Very Good Edit. Resolving the dispute is an active goal here.

I'm not sure I know what you mean when you say, "the guideline needs substantial rationalization." Are you saying that it stands in need of rationalization now, or that the fact that we have to rationalize it indicates something, or... what? -GTBacchus^(talk) 16:56, 9 May 2009 (UTC)

I thought I was pretty clear that (a) it needs rationalisation now, and (b) that this means we shouldn't obsess about current phrasing, until after the structure is sorted out. Rd232 ^talk 18:03, 9 May 2009 (UTC)

Well, if it had been entirely transparent to me, I wouldn't have asked, you know? What kind of rationalization does it need? -GTBacchus^(talk) 18:10, 9 May 2009 (UTC)

Yes I know, that's why I answered :) Rationalisation it needs is structural - it's all over the place due to the ad hoc development of WP policy. eg overlap between overlinking/underlinking and General Principles. Rd232 ^talk 18:19, 9 May 2009 (UTC)

I see. It needs organizing and clarifying. Cool. I'm in favor of those things. -GTBacchus^(talk) 18:32, 9 May 2009 (UTC)

• All standards for when to link are subjective; so is "helpful to readers", so is "necessary to readers" as it stands now (as I said above, that's much too narrow). Septentrionalis PMAnderson 17:03, 9 May 2009 (UTC)
  • I think a little bit of subjectivity is not a problem. Since the "rules" here are not meant to be applied mindlessly by robots, including a little subjectivity gives us a comfortable space in which to use judgment, and possibly refine a more detailed consensus over time. -GTBacchus^(talk) 17:06, 9 May 2009 (UTC)

The current standards have subjective elements, but they have a concrete core (see various bullet points). "or that many readers will want to see" is so subjective as to be meaningless. It adds no value, but it does add confusion. Rd232 ^talk 18:03, 9 May 2009 (UTC)

Um... I disagree. I mean, I'm not married to that phrase, but I don't think it's as bad as you make it out to be. How would you phrase it better? -GTBacchus^(talk) 18:10, 9 May 2009 (UTC)

I see no reason to have anything resembling this phrase. I don't see how it adds anything except a certain freedom to ignore everything else the guideline says. Rd232 ^talk 18:19, 9 May 2009 (UTC)

So, with the deletion of that one phrase, is it otherwise good? -GTBacchus^(talk) 18:32, 9 May 2009 (UTC)

No, without that phrase it's still worse than the original. But I'm not interested at this point in fisking the details which amount to issues of clarity and style. Rd232 ^talk 03:22, 10 May 2009 (UTC)

Well, you might as well know that I don't care about the details of clarity and style right now. Those can be worked out. I'm more interested in whether you can separate the content from how it's phrased, and see what's good in it. In particular, what I liked was defining "underlinking" in one sentence, defining "overlinking" in one sentence, and then note that "Each link should offer some benefit to compensate for its presence". That seems to me to be a pretty good way to present the ideas we're talking about. -GTBacchus^(talk) 18:05, 10 May 2009 (UTC)

For general-interest articles, where the links are of the "see also" or "for more information" type, it may be better to not link in the summary, deferring the link until the term is defined later in the article. Numerous links in the summary of an article may cause users to jump elsewhere rather than read the whole summary.

This is somebody's pet peeve. I have indicated that not everybody agrees with this. I certainly don't; if Pontiac's Rebellion makes me want to know who Jeffrey Amherst is, I want to get there at first mention, in the lead, rather than search down to find the actual link. Septentrionalis PMAnderson 18:07, 9 May 2009 (UTC)

I would tend to agree. I like links on first occurrences. Is there really a problem with people surfing away from articles they're interested in, because they're so distracted by the shiny links? -GTBacchus^(talk) 18:34, 9 May 2009 (UTC)

I suspect this is suggesting not to link things which shouldn't be in the summary/lead anyway (like some less relevant See Also links). As it stands it's not really clear what it means though. Links that are "of the "see also" or "for more information" type"? In a way that could mean any link! Maybe it means daughter article links from a section in the article, also being mentioned in the lead, shouldn't be linked? I don't see why. Anyway if it's not clear what it means and nobody can clarify then we're better off without it. Rd232 ^talk 18:47, 9 May 2009 (UTC)

This rule would be a disaster if it were to be misinterpreted (which is really easy). I also agree to removing it. Something seen as unworthy ("See also" type) would not be mentioned in the summary anyway. Nicolas1981 (talk) 09:14, 11 May 2009 (UTC)

Well I don't know about disaster :) but if it can't be clarified let's get rid of it. However there is perhaps a related point worth making: lede sections are more commonly scanned than the body of an article (I think), and overlinking inhibits scanning as much as underlinking (in scanning, the bluelinks act as a sort of emphasis, rather more than in normal pace reading). So perhaps some kind of phrasing would be appropriate to suggest that link density in the summary should consider the effect on scannability. I'm not sure the suggestion can be any more specific than that; it depends so much on context and judgement; but it's something to bear in mind. Rd232 ^talk 03:54, 12 May 2009 (UTC)

hm - well, actually that sounds to me like a personal preference, and it seems too speculative/theoretical. i agree that the "see also type" bit is unclear, unworthy and needs to go. is there a need to single out the lede section for "special treatment"? if so, would "some editors feel there's a particular need to avoid overlinking in the lede section" suffice? Sssoul (talk) 04:42, 12 May 2009 (UTC)

No, because as I said scannability is affected by both overlinking and underlinking. The ideal lead section doesn't (for scannability) have zero wikilinks - it has key words linked, but (as far as possible) only the key ones. This is basically what we do anyway, I think. However you're right in that focussing on scannability will on average reduce linking somewhat, because underlinking doesn't affect scannability as much if the lead is short (and generally if it's long, it's going to have enough wikilinks in it already). The current intro for Wikipedia, for instance, is one that strikes me as reasonably linked for body text (maybe a little overlinked even there), but definitely overlinked for a lead section. Rd232 ^talk 04:57, 12 May 2009 (UTC)

Does this only mean that, eg. on the Military history of Canada page, {The '''[[military history]] of [[Canada]]''' comprises hundreds of years ...}, giving; "The military history of Canada comprises hundreds of years ..." is wrong? Which in an earlier discussion[1] was given as a good example of this working.

Or does it also mean; {The [[Military history of Canada]] comprises hundreds of years ...} which would be rendered, only on the that page as; "The Military history of Canada comprises hundreds of years ..." is wrong?

Note that the wikisoftware doesn't serve the browser a self-referential link, it transforms the text of such links to bold, plain-text. HarryAlffa (talk) 20:45, 7 May 2009 (UTC)

From WP:Lead, "As a general rule, the first (and only the first) appearance of the page title should be as early as possible in the first sentence and should be in boldface"

For the article MediaWiki the first line is, {'''MediaWiki''' is a [[World Wide Web|web-based]] [[wiki software]] application ...}.

I propose that the preferred way of emboldening this first mention is, { [[MediaWiki]] is a [[World Wide Web|web-based]] [[wiki software]] application ...}. Which would be rendered as nolink, but boldfaced text, "MediaWiki is a web-based wiki software application ...." HarryAlffa (talk) 20:45, 7 May 2009 (UTC)

Why would this represent an improvement? The use of actual bolding (three apostrophes) is far more widespread, and seems more logical (better to type what you want to appear, rather than rely unnecessarily on an extra bit of wikisoftware behaviour, which in theory might change).--Kotniski (talk) 08:01, 8 May 2009 (UTC)

I should really have said, shouldn't I? :)

It makes it easier for editors who copy the lead, (as wikicode) into another (perhaps summary) article.

1. They don't have to edit the code to get rid of the bold and replace it with a link back to the original.
2. It automatically has a link back to the original article; more efficient.
3. It eliminates errors wnen linking back; more efficient.
4. At worst, it does no harm.

If I may quote from Cascading Style Sheets (with my emphasis);

CSS is designed primarily to enable the separation of document content (written in HTML or a similar markup language) from document presentation, including elements such as the colors, fonts and layout.

The original idea behind HTML was to do exactly this. HTML was to have no presentational information at all. It was entirely up to the browser how it presented the text. It would do no harm to think this way of self-referential links. The wiki-software programmer has obviously allowed for this, and has taken action accordingly - I happen to think it is both good error catching, and a pleasingly elegant touch to turn this user "error" into a useful feature. HarryAlffa (talk) 12:49, 8 May 2009 (UTC)

If we accept the "seperate content from presentation" philosophy, you would reasonably conclude that self-referential links are actually quite a good idea, because if the community decides to change how the wiki-software presents self-referential links, then voila! The whole of Wikipedia is changed, with no effort from editors! At worst it does no harm. HarryAlffa (talk) 12:49, 8 May 2009 (UTC)

So, is it currently "banned" or not? I read the Guidelines as not banning linking the title to itself (it does no harm), only linking to other articles from within it. HarryAlffa (talk) 12:49, 8 May 2009 (UTC)

I don't know that it's explicitly banned, though it probably goes against widely accepted convention (i.e. if you did it, someone would probably come along and change it back sooner or later). I'm not sure your analysis can be extended to all cases - in many cases the bold reference in the first line is not the same as the article title (for example, in the very frequent cases where the article contains a parenthetical disambiguator). In that case using a link instead would certainly not give the correct result. So we would need to use two different techniques depending on the article title - confusing in itself; and what happens when an article is renamed? You'd have to remember to change from links to apostrophes. All in all I don't see this idea offering enough practical benefit to outweigh the extra difficulties it would cause.--Kotniski (talk) 13:11, 8 May 2009 (UTC)

Of course you'll agree that convention can change - if there is good reason, and I've given a couple at least. There is no difference between the two techniques if the article is renamed; { [[Foo (disambiguator)|Foo} } would appear correctly, and I would say that this amplifies my efficiency points. The article "May contain nuts", uses this style; go there and copy the wiki-code (as I have done here) of the lead, then paste it somewhere harmless, and you will see you have the link back with no effort. HarryAlffa (talk) 13:39, 8 May 2009 (UTC)

Wait! We've been looking at this with both eyes! We need to close one! The MoS Guides are all about how the text is presented to the user. We've been seeing too much, close your editor-eye, and look only with your user-eye. All you see is the first instance of the article title in boldface.

The MoS tells us what the text should look like, not how that look is to be achieved. Therefore this Linking guide cannot make any comment on what wikicode is to be used. Logic's a wonderful thing! Now were do you go to read up on recommendations on formatting wikicode!? HarryAlffa (talk) 15:04, 8 May 2009 (UTC)

Unlike much of what MoS says, this actually has a reason; although not much of one. We began as HarryAlffa suggests; that's why self-links bold. But if the page is moved, a self-link will cease to bold, and anybody who clicks on one will go around in a pointless circle.

A useful page would explain this and let editors choose. Septentrionalis PMAnderson 21:11, 8 May 2009 (UTC)

Yes, explain the advantages and disadvantages of the methods of bold-facing a first-instance of the article title. But where do you put this explanation? I am of the view that a Style Guide should not dictate the markup to be used to express a style, but an example of a couple of methods on some things might not be harmful. HarryAlffa (talk) 16:02, 9 May 2009 (UTC)

It seems to me that page-moves are not very common? Even when they do happen, aren't they likely to lead to a name change of the article? So you'd have to change the bolded first instance anyway? But if page-move is the only reason not to have a self-link, it seems an awfully small one, and needs a rethink! Any edit problems would be ironed out in the usual manner anyway! Is a self-link actually banned? Is it written anywhere in policy etc? HarryAlffa (talk) 16:02, 9 May 2009 (UTC)

Depends what you mean by common; some pages are tweaked routinely. In any case, it is our present custom to bold with three ', not with a self-link. Septentrionalis PMAnderson 18:11, 9 May 2009 (UTC)

I noticed your careful use of "present custom" - not actually banned? If there were advantages you might support a move to at least let this method be a choice for editors? What do you mean by "tweaked routinely"? Surely pages aren't renamed routinely!? HarryAlffa (talk) 18:32, 9 May 2009 (UTC)

Harry asked "is this not actually banned?" A distinction without a difference. Several regulars here believe that MOS is legislation, but they are (fortunately) mistaken: MOS can't "ban" anything. But it can advise against it, and if that actually represents consensus on a given article (and in this case it very well may), that consensus should prevail. Then again, that consensus should prevail even if this ill-begotten page says otherwise; that is policy. Septentrionalis PMAnderson 16:43, 11 May 2009 (UTC)

Just wanted to point out that this issue has come up before, and the consensus seems to be against the proposal: see "Wikipedia talk:Manual of Style/Archive 107#Using self links as bold reiteration". — Cheers, JackLee ^–talk– 18:43, 9 May 2009 (UTC)

From WP:CONS, "Wikipedia does not base its decisions on the number of people who show up and vote; we work on a system of good reasons". Even a simple head-count (I think) doesn't give a great consensus, good reasons were given for self-links (much the same as my own), and the reason against was "Why not just use the markup?". No reason really! HarryAlffa (talk) 19:31, 9 May 2009 (UTC)

There can be no doubt from any general reading[2] on Separate structure and presentation (from W3) that using self-links in this way conforms to the general principles being advocated by authorities on matters of markup languages. HarryAlffa (talk) 19:31, 9 May 2009 (UTC)

Experience has shown that separating the structure of a document from its presentational aspects reduces the cost of serving a wide range of platforms, media, etc., and facilitates document revisions.[3]

It might be worth noting that bold text and self links are coded differently in the generated html. Bold text generates "<b>bold text</b>", while self link generates "<strong class="selflink">self link</strong>". While most browsers render both as bold, this is not guaranteed. In fact, users can and have modified their skin or CSS to display self links differently to suit their preferences for navbox current page appearance. This would actually be a good reason to actively discourage the use of self links to achieve a bold effect. -- Tcncv (talk) 06:49, 10 May 2009 (UTC)

Excellent investigating, you beat me to it! But, you have come to exactly the wrong conclusion! It is this ability for users to have their own preferences is why it's a good idea, users didn't have to modify their CSS, it's an optional extra. The bold tag is deprecated in HTML 4, so the <strong> tag is actually preferred instead, better for accessibility. See general reading[4] and W3Separate structure and presentation. I'd like you to name a browser that doesn't render <strong> as bold! The example you gave has the CSS "class=" modifier, this "may" render it as ... just about anything - that's the design purpose of it. HarryAlffa (talk) 17:22, 10 May 2009 (UTC)

You misunderstood. What is really important is that they produce different code, and this may lead to them appearing differently, which is unacceptable. Class selflink can be redefined at any moment. It may be useful, for instance, if self-links (except those in navboxes) appeared yellow and in italic. Ruslik (talk) 17:34, 10 May 2009 (UTC)

I have to claim a high degree of expertise in HTML & CSS. Without wishing to sound harsh; you are wrong on every level your text touches on! Sorry! Wrong: The different code is not important, it is in fact an advantage. Wrong: Every browser will display <strong> as bold, except at the express design of the webpage creator, or if you, the user, decides it should be different with local CSS files, often for accesibility reasons. Wrong: class="selflink" cannot be redefined at all, it is defined by the CSS when the browser loads the page. Clarify: The navbox different appearance is just an example of the advantages of separating presentation instructions from the document text. I have only read the CSS lead, but you should read the article, it will help you see why the W3 is right! HarryAlffa (talk) 18:18, 10 May 2009 (UTC)

Really? To change the definition of this class I need only to edit MediaWiki:common.css. Ruslik (talk) 19:19, 10 May 2009 (UTC)

Ah! The dynamism implicit in your text I took to mean "re-defined after the page had loaded", which is why I said, "it is defined by the CSS when the browser loads the page", but even so - yes, really. This is exactly the advantage I was talking about! One change to MediaWiki:common.css, by the express wish of the webpage author, will change how the whole site looks, barring local CSS! This behaviour is the whole point of CSS! HarryAlffa (talk) 20:15, 10 May 2009 (UTC)

First, HarryAlffa in incorrect – the <bold> tag is not deprecated in HTML 4. It may be deprecated in future HTML and XHTML standards, and the <bold> tags will eventually need to be replaced with <strong> tags. However, that replacement would still be without a class name, so it would still not be the same as self-links. SO the stated problem still exists.

Excluding first sentence usage under discussion, the only commonly accepted usage of self-links that I am familiar with is the use in nav-boxes or similar constructs (like a shared info-box with a nav-box-like section) transcluded onto multiple pages. (If you are aware of other generally accepted uses, please let us know.) Further, I suspect that those who have modified their CSS have done so with the intent to modify the nav-box appearance and would understandably be surprised if they were to encounter occasional pages where the first sentence appearance is affected. This is why I do not support using self-links in first sentences, as I see it as a misuse of the wikilink syntax having potential unintended side effects. -- Tcncv (talk) 06:41, 11 May 2009 (UTC)

Oops! A bit of loose talk as well as a small boo-boo there! HTML 4.01 says their use is "discouraged". The HTML tag for bold is of course <b> not <BOLD>, it is not to be replaced by <strong> however, it is recommended to be replaced by CSS methods like; font-weight: 900; where the 900 can go down to 100 in century decrements, or be a value word; bold, bolder, inherit, lighter, normal. <strong> & <EM> are recommended used in structured text for emphasis, visually either bold or italics (usually, but it's up to the browser, or local CSS), voice synthesisers give strong emphasis or emphasis, I use either font-weight: or <strong>, depending on context. But we're not here to discuss the intricacies of HTML!

Those who have modified their CSS (to change the colour to red for instance) will have this statement in their CSS file; strong.selflink { color: red; }, I suspect that they will have come across HTML pages with a <strong class="selflink"> tag in the <BODY> of the article in their adventures in wikipedia, and will have made one change of prefixing "td > " in their CSS file giving; td > strong.selflink { color: red; }, so that it's exclusively in a table cell - nav-boxes are rendered in HTML as tables. The addition of "td > " is all that's required as a change, if they haven't done it already - not a problem for those clever clogs! I hope this now means that you are on the way to supporting this usage!

You've said, "...self links to achieve a bold effect", and "misuse of the wikilink syntax". It's not using self-links to achieve any affect - it's using self-links to indicate a self-link, that's all the wiki-software does - it serves a page with class="selflink" on the HTML tag, it is only through purposeful CSS that the tag with this class, is altered in the way it normally displays. So their will be no unforeseen side-effects. Cheers :) HarryAlffa (talk) 16:27, 11 May 2009 (UTC) PS: It's important that everyone realises that the wiki-software doesn't present the browser with a hyperlink, it presents plain text enclosed in the "STRONG" tag.

The current cite note for WP:Link General principles bullet point, "Do not place links in the bold reiteration of the title in the article's lead sentence"

... While linking the article to itself produces bold-face text without a link, this practice is discouraged as page moves will result in a useless circular link through a redirect. ...

This is an editor-efficiency and a user-confusion reason, I think you would agree.

1. It is efficient for editors who move a page; they don't have to re-jig the wiki-code.
2. It is efficient for the user; they aren't left with a link which loops back.

But how much efficiency saving is there really? I think these are both quite thin, 1 for the negligible positive value, and 2 for the negligible negative value.

• Page moves cannot, surely, be so common that cognisance must be given to them at an articles inception, and determine wiki-code for the one-time page move an article might make.
• Assuming a page move takes place, on an article with a self-link, and the editor leaves the self-link as is. So. An editor makes a mistake, another comes along and fixes it. Yes? Is that mistake inherently more likely than any other? How much is a typical user going to be put off by a funny link? Surely the vast majority will take it in their stride, they'll have seen and got around much worse.

So what efficiency savings are gained by leaving self-links in an article? It makes it easier for editors who copy the lead, (as wikicode) into another article, or talk page.

1. They don't have to edit the code to get rid of the bold-face; more efficient.
2. They don't have to edit the code to link back to the original; more efficient.
   1. It automatically has a link back to the original article.
3. It eliminates errors when linking back; more efficient.

That's at least 3 efficiency savings versus 2.
Comparatively arithmetically it seems that using self-links wins out.
But how much efficiency saving is there really? How often do editors copy that bit of the lead? I could be completely wrong, but my suspicion is that it has to be more often than page moves - which means the 3 to 2 count in favour of self-links wins out, even if the point-weight on each side is equal. I think the point-weight of each is greater on the 3 side, which would multiply the 3 to a larger number.
Besides my thought that page moves are rare, what gives rise to my suspicion that editors copy that part of the lead more often than an article undergoes a page move?

1. Summary articles would make a good start that way.
2. Other articles may quote a portion of the lead when explaining the first use of a term in it's own body text.
3. In a talk page recently I brought up the definite article, Set (mathematics), & intensional definition to describe my view. I quoted exactly the first sentence of two of those articles, but had to re-jig the wiki-code to provide a link.

If we can make article talk pages more efficient in coming to conclusions, then this will help the efficiency of the whole Wikipedia project - I think this is quite a weighty reason for the self-link usage.
User changes to Nav-box display have been cited as a reason for not using self-links, however these users have to create their own CSS file with;
strong.selflink { css display commands }
Nav-boxes are served as tables, so these users need only a small prefix for table data cells;
td > strong.selflink { css display commands }
or for table headers;
th > strong.selflink { css display commands }
or for anywhere in a table;
table > strong.selflink { css display commands }
or perhaps
td > strong.selflink, th > strong.selflink { css display commands }
for both table cells and table headers.
Which some would undoubtedly have done already. HarryAlffa (talk) 17:48, 14 May 2009 (UTC)

I seems to me that if the Wikipedia Lead section guideline recommends that the first occurrence of the page title in the lead sentence be "boldface", the KISS principle would strongly favor using the wiki markup that specifically marks text for boldface presentation. I scanned about 250 random articles and did not find any selflinks used in the way you suggest, so the use of the wiki markup for boldface appears to be the overwhelmingly accepted convention – one that seems unlikely to change. And yes, as you point out, <strong> is not an equivalent substitute for <b> (although most or all browsers render them the same) – another reason not to use selflinks to achieve boldface. -- Tcncv (talk) 23:26, 14 May 2009 (UTC)

That's why I've gone to such trouble to explain the advantages of the alternative to convention, an article count measures inhertia, it says nothing about advantages or disadvantages - and a bot was employed to change all articles to this markup. The "another reason not to use selflinks", I've already answered above, and explained why it is not "to achieve boldface", see above. I think it's a pretty weak argument to quote the KISS principle of "unnecessary complexity", of markup that everyone is familiar with, and that lots of people used in this way, before guidelines discouraged it for what is now clearly two very weak reasons. The count is still three clear advantages to two questionable ones. HarryAlffa (talk) 13:09, 15 May 2009 (UTC)

If I might make a suggestion: I think this is the sort of thing that might be developed further at WP:PROJPOL, collaboratively developing a proposal with clear pros and cons. Rd232 ^talk 05:27, 15 May 2009 (UTC)

It may well be the sort of thing, but keep it here for now, the pros and cons have been well laid out - yes?. You may like to bring people here with a more technical slant on the way the wiki-software works, if you know any. I think I've done not to badly in explaining the principles, and some specifics, of the technical side of it, but someone with first hand technical knowledge of the Wikipedia configuration would be very useful to verify or villify my technical apraisal. :) HarryAlffa (talk) 12:06, 16 May 2009 (UTC)

There is no good reason why self-links can't be used as a method to
conform to the style for first mention of an articles title in the body.

Discussion above has shown that;

• previous reasons for discouraging this are now weak.
• nav-box "problems" aren't really problems.
• there are positive benefits to this.

Additionally; using self-links for the first mention of an article's title guarantees that all articles will conform to Wikipedia's style convention, even if that convention changes, by dint of the configuration of Wikipedia.

I think the concensus is that the MoS be changed to indicate that self-links are a good idea! HarryAlffa (talk) 16:42, 19 May 2009 (UTC)

I am happy that your finally reached a consensus with yourself, however, I do not think this is enough to change MOS. Ruslik (talk) 18:47, 19 May 2009 (UTC)

Do you have anything constructive to contribute? I have made reasoned, well founded, clear discourse on this subject, and have shown that all the problems raised by fellow Wikipedians, are not problems at all. Consensus is built on a system of good reasons, not on a simple vote count; for this reason and my system of good reasons I now declare concensus in favour of using self-links! :) HarryAlffa (talk) 19:24, 19 May 2009 (UTC)

I've now changed the WP:Link to reflect this new consensus. HarryAlffa (talk) 13:35, 21 May 2009 (UTC)

See my Essay on self-links. HarryAlffa (talk) 13:41, 21 May 2009 (UTC)

I reverted this change since there is no consensus. Ruslik (talk) 13:48, 21 May 2009 (UTC)

Indeed there isn't; and given how long the existing approach has existed and how widely, a small discussion here between a handful of editors isn't sufficient to overturn that either in policy or in practice. Also despite the pros and cons discussed above (not sure where the weight falls), my gut feeling is still that it's wrong. Links are for linking, not for style. Rd232 ^talk 14:11, 21 May 2009 (UTC)

Your main argument seems to be inertia: we've done it like this for ages. This, and a bad gut, are extremely poor "reasons" for disagreeing with something "new". Your last sentence is answered clearly in the discussion above, and in the links provided there. As you feel unable to weigh the pros & cons, how strongly held can your opinion be? Please assimilate the clear advantages I have given, before hardening your will! HarryAlffa (talk) 14:36, 21 May 2009 (UTC)

There is no consensus for this change (in fact, I'm not sure anyone else agrees with you). A very substantial reason I gave was that a consensus for this change would need to be substantial, eg by listing at WP:VPP and doing an WP:RFC, that kind of thing. And you can't get away from the fact what you are proposing elides form and function in a way which is fundamentally inelegant and confusing. Rd232 ^talk 14:43, 21 May 2009 (UTC)

I have undone Ruslik's revert. HarryAlffa (talk) 14:36, 21 May 2009 (UTC)

You haven't persuaded anyone (yet); you may or may not be an Alfred Wegener ahead of his time, but that doesn't mean you should get into an edit war. Rd232 ^talk 14:43, 21 May 2009 (UTC)

Mr Herr Alfred Wegener was ahead of his time, and it may appear to someone behind the times that I am ahead of mine, but I'm simply in step with current thinking on markup languages. :) HarryAlffa (talk) 15:34, 21 May 2009 (UTC)

Can you think of ONE strong reason why it should be banned, or even just discouraged? Only then should it be not recommended, we can't just say stuff for no reason, or we can but we'd look stupid! HarryAlffa (talk) 14:46, 21 May 2009 (UTC)

Look Harry, due to the community nature of WP, policy on WP is made either (a) reflecting existing practice, to clarify it (b) to change existing practice for reasons agreed by a substantial enough part of the community that the policy change is likely to be reflected in practice. Your proposal currently meets neither. You would need to get more widespread input to make this change to the policy. (I think you'd be wasting your time to try, but I could be wrong.) Rd232 ^talk 15:01, 21 May 2009 (UTC)

Look RD, patronise much? WP:Link is not policy. Now to your other points; wait a minute, your whole argument based itself on WP:Link being policy, so you have constructed no argument here!

Would it be sensible to say that WP:Link offers good advice? I have demonstrated that using self-links is a good idea. Do you totally discount the strength of reason in a debate? HarryAlffa (talk) 15:34, 21 May 2009 (UTC)

So it's a WP:guideline, not a WP:policy. So what? I was using "policy" in a loose sense to cover both, and my comments apply equally. Also, I wasn't intending to be patronising, but I am now: You seem to be under the misapprehension that we need to persuade you that the status quo is correct. We don't. Rather, you need to persuade the wider community that this change to the status quo is desirable - to their satisfaction, not yours. Mmkay? Now please stop edit warring. Rd232 ^talk 16:15, 21 May 2009 (UTC)

The second-last paragraph of WP:CONS, "we work on a system of good reasons".
Ok, let's pretend that you where saying exactly what you meant, instead of the loose expression you chose :)! So. You would only accept a good idea if everyone else accepted it? Not on the strength of the idea itself?

(b) to change existing practice for reasons agreed by a substantial enough part of the community that the policy change is likely to be reflected in practice.

Well, you didn't say "you", you said (or implied) the community. Hmm. The community would only accept a good idea if a substantial enough part of the community accepts it, not on the strength of the idea itself.
So a substantial part starts with one individual. He considers an idea is a good one, he gives reasons and examples of why it is a good idea. Some people say it has some bad effects because of x,y, or z. But the man shows that these are, in fact, minor inconveniences easily overcome, and also shows that the advantages outweigh any perceived disadvantages. No one is offering any counter-reasons to his common-sense appraisal of the idea and it's advantages. He hopes this will show people that it is a good idea (how could it fail to?). Some people say that it is not a good idea because people haven't accepted it, not because it is a bad idea.
WP:CONS, "we work on a system of good reasons". Doesn't give a fork about Status Quo. :)
You seem to be under the misapprehension that you are speaking for a substantial part of the community. Do you have a system of good reasons to support this? The full sentence is, "Wikipedia does not base its decisions on the number of people who show up and vote; we work on a system of good reasons". HarryAlffa (talk) 18:54, 21 May 2009 (UTC) PS: It takes more than one to edit war,and you started it, unreasoningly! i.e without reason.

See RfC below

You're missing the point that yes ultimately policy is supposed to be based on "good reasons", but it's not up to one person to decide that whether there are good reasons for a specific policy change. That requires wider debate, which I guess the RFC will supply. NB I think the RFC text rather long, it could do with a "nutshell" summary at the top for people new to the discussion. Rd232 ^talk 19:55, 21 May 2009 (UTC)

Well, we agree on something! Which is nice! Policy (& guidelines) should, or, is supposed to, be based on good reasons. There shouldn't be a but after that. But even allowing your but ... I am not the only one deciding this. There had been 3 people disagreeing with this (Kotniski, Septentrionalis, Tcncv), I have answered their critiques and they have been silent - from this I MUST infer their agreement, or nothing would ever get done.

Ruslik also disagreed and made two contributions which I answered, and he fell silent - until;

I am happy that your finally reached a consensus with yourself, however, I do not think this is enough to change MOS. Ruslik, 19 May 2009 (UTC)

He ran out of reasons to disagree, so settled on just disagreeing. You (Rd232) have disagreed with changing the MoS because ... em ... it's a change in the MoS! I don't think that counts as a good reason to disagree, your all for good reasons aren't you? So, that's HarryAlffa (me), Kotniski, Septentrionalis, Tcncv - 4 for this change, and Rd232, Ruslik - 2 against this change, but can give no good reasons for being against it. HarryAlffa (talk) 20:44, 21 May 2009 (UTC)

You can't infer agreement from silence, especially when people have previously disagreed. And mischaracterising Ruslik and my position isn't helpful. You've argued very cogently up to this point - though I continue to disagree with you, we'll see how the RFC goes - so don't get tendentious now. Rd232 ^talk 20:58, 21 May 2009 (UTC)

You can infer consensus from silence, sometimes, and I was pushing it a bit to see if anyone objected to me claiming they now thought this a good idea when they had raised previous objections. :) One objector from the 3, so far! HarryAlffa (talk) 12:19, 22 May 2009 (UTC)

That is correct. You (HarryAlffa), I (Tcncv), and others had already stated our opinions, and it was obvious that these opinions were not likely to change. That's when I have the philosophy that it is best to stop debating and listen to what others have to say. That does not mean I have changed my position, accepted your arguments, or otherwise given up. It is more a case that I believe that continuously restating the same position is not productive. Just take a look at any of the date linking discussion as prime examples of where the unending debate between a small number of vocal editors drowns out the greater community – pretty much giving everyone a bad experience and likely dissuading many from contributing their own useful insights. -- Tcncv (talk) 00:49, 22 May 2009 (UTC)

You raised a technical point on navbox CSS, which I explained in the discussion, and also in my essay. Do you not accept my technical apraisal? HarryAlffa (talk) 12:19, 22 May 2009 (UTC)

It is an assessment of the technical implications of it, and the advantages and advantages - so it's arithmetic. The opinion comes in when counting advantages and disadvantages, which of the advantages I claim do you disagree with? HarryAlffa (talk) 12:19, 22 May 2009 (UTC)

Please consider the wisdom of WP:SNOW and move on. -- Tcncv (talk) 22:37, 22 May 2009 (UTC)

OK, here's a suggested restructure for more clarity: User:Rd232/WPL. Comments? Questions? Rd232 ^talk 05:18, 10 May 2009 (UTC)

Are these the only differences? I see no improvement in clarity; but then I see no material differences here. Septentrionalis PMAnderson 05:21, 10 May 2009 (UTC)

No, they're not - I made changes in the first edit; obviously I should have just saved the current WP:linking, and then made the changes, so this was confusing. Bah. Sorry. :( Anyway I've fixed this now; look at this diff to see all the changes. Rd232 ^talk 05:42, 10 May 2009 (UTC)

The change of order seems largely harmless. I dislike the change in the lead; it removes more traces of Build the Web, which is little enough represented now. Septentrionalis PMAnderson 15:17, 10 May 2009 (UTC)

Rd, you're a good writer indeed, but I find it hard to make sense of the diffs you've supplied, since they include the relocating of tracts of text as well as more substantive changes. I wonder whether you could supply a bullet list of the non-trivial changes and relocations here, to help us through? Like

• Text on blah moved from [section] down to a new section [name] and combined there with [blah].
• First para in blah section rationalised ...

Just one point thus far: "These links should be included where it is most likely that readers might want to use them; for example, in article leads, the beginnings of new sections, table cells, and image captions." Why would readers be most likely to want to click on links in the lead, rather than in the body of the article? Does this go further than the recommendation to link only the first occurrence? Bit confusing to me. Tony (talk) 16:03, 10 May 2009 (UTC)

• Because it is first occurrence, where readers who are curious about the term are likely to first see it, and want to know more. But the existing text quoted above offers a opinion the other way; silence may be best. Septentrionalis PMAnderson 16:12, 10 May 2009 (UTC)

PS Why was the page frozen? I thought the recent edits were mostly constructive. Tony (talk) 16:06, 10 May 2009 (UTC)

Because editors were reverting and rereverting without discussion save for edit summaries, and thus necessitating protection until the issue is resolved. --MASEM (t) 16:14, 10 May 2009 (UTC)

It has probebly resolved sufficiently to be unprotected, provided nothing rash is done. Septentrionalis PMAnderson 16:10, 11 May 2009 (UTC)

That's what I thought last time I suggested unprotection (see above), and I was proved wrong a few days later. I was surprised it was re-protected when it was, i.e. after the editwarring seemed to have stopped, but that state does seem to promote calmer discussion.--Kotniski (talk) 16:42, 11 May 2009 (UTC)

It's a pity, because I thought we were all making some progress just before that re-protection. I see that MoS central has been protected again, for no apparent reason. Could be someone at ArbCom sticking their fingers in, but of course I have no proof. Wouldn't be the first time. Tony (talk) 16:49, 11 May 2009 (UTC)

Tony, calm down. There was an edit war there too. But if we can do without continual exact reversion, I think we can try unprotection. Septentrionalis PMAnderson 16:53, 11 May 2009 (UTC)

Are you proposing some modification to BRD that prohibits the R stage?--Kotniski (talk) 17:03, 11 May 2009 (UTC)

No, I am requesting that you (and RD232) actually use the D stage. It would also be nice, and is in the spirit of the essay, if you tried revision rather than reversion whenever possible. Septentrionalis PMAnderson 17:15, 11 May 2009 (UTC)

Unless I'm missing something, you're missing the point: only you at this stage are arguing to unmerge BTW, and doing so unilaterally against consensus is nothing to do with WP:BRD. This argument has been and gone. You want to keep BTW as a concept and phrase? Listen to your own advice ("revision, not reversion") - BRD's still here, so work on improving/clarifying/developing it. Rd232 ^talk 03:45, 12 May 2009 (UTC)

OK, neither of the personal attacklets above makes any sense (I and Rd have been consistently engaged in discussion, and PA isn't doing anything against consensus at this stage). Let's leave it and try not to go down that road again.--Kotniski (talk) 09:07, 12 May 2009 (UTC)

Summary

OK, I'll try and summarise the (proposed) changes:

Structural:

• General Principles: shortened, with most of the bullet points moved elsewhere, and the pipe link example moved to a bullet point. The bullet points are mostly kept as bullets in their new location but in some cases merged into the body - eg "In general, link only the first occurrence of an item. ..." now under Link Density. I would like to get rid of all the bullet points from this section but I can't see where else to put the remaining ones. Possibly they could be rewritten as unbulleted prose.
• "Internal links" added to section title of overlinking/underlinking, since that's what the section refers to (and contrast with External links section below)
• Example subsection of Overlinking/underlinking renamed Link Specificity and refocussed to match
• Chronological items moved down to a Specific cases section (I thought there'd be more, but it still doesn't really fit anywhere else)
• Piped links and redlinks made part of a new Techniques section
• Link maintenance section moved to the bottom
• Interwiki linking renamed interwiki links for consistency

Nonstructural:

• intro: tighten and clarify, noting distinction between roles of internal and external links. Drops some redundancy and avoids unnecessarily shoehorning in the phrase "build the web" whilst retaining the concept. (Phrase is kept in the General Principles section where the BTW shortcut goes.)
• Wikilinked 8 words in the Link Density section, in "Do not link eight words in one sentence ..."

As far as I remember and can see from the diff, that's about it. Very little rewriting apart from the intro - just moving stuff about. Rd232 ^talk 20:13, 10 May 2009 (UTC)

Nothing wrong with bullet points on pages like this; they aren't intended to be read continuously, but to be reserves of advice on particular points.

The rearrangement is harmless, and if it makes more sense than the present layout to one reader, it may to another. RD232 likes superheads with single subheads more than I do, and more than most style-guides, but I don't suppose it matters. I object to the rewriting, however; please leave the intro as it is. Septentrionalis PMAnderson 00:15, 11 May 2009 (UTC)

"Likes superheads with single subheads"? I already explained above that I couldn't see what else to do with Chronological items, and I did this partly because I thought something else might come along for that section. The intro is an improvement for a number of reasons, but I don't want to make that a sticking-point for the restructure, if others agree with Pmanderson (who incidentally hasn't explained why he objects). Rd232 ^talk 02:06, 11 May 2009 (UTC)

• • Also the Piped links section (which should be summarized anyway; we already have guidance on that elsewhere); but all this is a matter of taste. Septentrionalis PMAnderson 16:14, 11 May 2009 (UTC)

the rearrangement looks to me like it's moving in right directions. i'd list these two points together, since they're about the same issue; the point about creating red links should follow them, not interrupt them:

• Think carefully before you remove a link altogether—what may seem like an irrelevant link to you may be useful to other readers.
• If you feel that a certain link does not belong in the body of the text, consider moving it to a "See also" section at the bottom of the article. (Remember that links can also be useful when applying the "What links here" feature from the target page.)

also, while we're at it, i've always had doubts about "aim for consistent link density" - that seems like a highly unnatural aim. it doesn't really jive with the general principles of linking on first mention and to aid comprehension/navigation, which might well result in uneven distribution of links. of course the point about not linking a bunch of words in a row needs to be made, but that's not the same as trying to impose "consistent link density" throughout an article. Sssoul (talk) 05:10, 11 May 2009 (UTC)

Seems fine to me. Nicolas1981 (talk) 09:01, 11 May 2009 (UTC)

Sssoul's "i've always had doubts about "aim for consistent link density" - that seems like a highly unnatural aim.". Yes, when I first saw that years ago, I unthinkingly accepted it. But on closer reflection, you are entirely correct: one paragraph can have more opportunities for high-value links than another. Often. Tony (talk) 16:52, 11 May 2009 (UTC)

It might be worth saying that if one paragraph has many links and another none, one should check to see if they were both done right; that may be the underlying thought. But the point is probably trivial. Septentrionalis PMAnderson 16:57, 11 May 2009 (UTC)

I concur. --Kotniski (talk) 17:08, 11 May 2009 (UTC)

OK; does someone want to try rewriting that Link Density bit in the draft? Rd232 ^talk 03:45, 12 May 2009 (UTC)

Whaddya think? -GTBacchus^(talk) 04:35, 12 May 2009 (UTC)

(outdent) nice start, GTBacchus. to me linking a string of words in a row is something that should generally be avoided regardless of the number of links in other parts of the text; and "link only the first occurrence" is a totally different point that should be separated more clearly from the point about avoiding long strings of links. something more like:

Aim for a sensible link density. Avoid linking strings of words in a row, since that makes reading difficult and undermines the value of each link.

In general, link only the first occurrence of an item. This is a rule of thumb that has many exceptions, including the following:

• If a later occurrence of a link is separated by a long way from the first. Avoiding duplicate links in the same section of an article is generally a safe rule of thumb.
• If the first link was in an infobox or a navbox, or some similar meta-content. The main text of the article should link relevant terms.
• Table entries are another exception; each row of a table should be able to stand on its own.
  

i'm not sure "link density" is the right header - the first-mention-only rule of thumb is not really a "link density issue". Sssoul (talk) 05:02, 12 May 2009 (UTC)

I agree that "link density" isn't the best header. I wasn't really thinking about it when editing the section, but that's a good point. I guess Infobox and Navbox should be linked, too. Good edit, anyway. -GTBacchus^(talk) 05:08, 12 May 2009 (UTC)

instead of making them a separate section, i think these two points can be added to the "linking styles" in the "general principles" section. but since that section is about internal links, i think the first couple of section headers need to be modified to:

1 Internal links: general principles

2 Overlinking and underlinking

o 2.1 What generally should be linked

o 2.2 What generally should not be linked

o 2.3 Link specificity

3 Techniques

o and so on

in fact it probably makes even more sense to rearrange it like this (if it doesn't lead to an overdose of sub-sub-sub-sections):

1 Internal links

1.1 General principles

1.2 Overlinking and underlinking

o 1.2.1 What generally should be linked

o 1.2.2 What generally should not be linked

o 1.2.3 Link specificity

1.3 Techniques

o etc

2 External links

would that work? Sssoul (talk) 06:33, 12 May 2009 (UTC)

That structure looks good on first glance, Sssoul. Concerning the "link density" proposed wording, my first strike-through below is a problem, because it clearly implies that you would, in any situation, link words such as "in" and "one" and "do" and "not" ... even the lexical items such as "words" and "sentence" would not be linked in that context (perhaps in an article on grammar, yes). It's not a good example, and why is one necessary here? Why even raise the issue of the evenness of link density—what should and should not be linked are clearly set out on the page. This seems a needless extra layer.

The "Avoiding duplicate links ..." sentence implies that it is a safe rule of thumb to link the same item again and again in subsequent sections, which, of course, may not be far from each other at all. Want an example of how irritating this could be?

"The main text of the article should link relevant terms." This is the general principle, isn't it? It should not be restated here, and without it, the point about infoboxes etc. has more impact. That is the point of the bullet, I think.

Tables: is the intended meaning that columns should be able to stand on their own, too? Unsure.

Aim for a sensible link density. Do not link eight words in one sentence and then none in the rest of the article. In general, link only the first occurrence of an item. This is a rule of thumb that has many exceptions, including the following:

• If where a later occurrence of a link an item is separated by a long way from the first. Avoiding duplicate links in the same section of an article is generally a safe rule of thumb.
• If where the first link was in an infobox or a navbox, or some similar meta-content. The main text of the article should link relevant terms.
• tables entries are another exception;, in which each row of a table should be able to stand on its own. Tony (talk) 08:31, 12 May 2009 (UTC)

Serial links

Tony, i think the point about avoiding (where possible) uninterrupted serial links is worth making, since one does see strings like "Pattie Boyd, an iconic 1960s English model and photographer" - which has several things wrong with it, but that many links in a row is one of them.

my current proposal for wording that point is: "Avoid linking strings of words in a row, since that makes reading difficult and undermines the value of each link." including all those links in the advice as a self-referential illustration of what's meant was Rd232's idea; i think it's a touch of humour, not to be taken as suggesting that anyone would normally link articles and prepositions. but obviously the links should go if they're more misleading than enlightening.

i'm all in favour of your other suggestions, though! Sssoul (talk) 09:02, 12 May 2009 (UTC)

While I have nothing against humour, I think this example is likely to go over people's heads. What we're cautioning against here is not just absurd strings of links like the "humorous" ones, but really any adjacent links, even if there are only two of them and even if each would seem perfectly justifiable on its own. If we illustrate this using an extreme example, people will think it's only extreme examples that cause the problems.--Kotniski (talk) 09:14, 12 May 2009 (UTC)

I added those wikilinks because the sentence before said "Do not link eight words in one sentence" and that amounted to eight wikilinks with each word wikilinked! It was intended to be humourous. I think such an example could be usefully attention-getting, but I take the point that it would need clarifying with a more real-world example. Rd232 ^talk 12:32, 12 May 2009 (UTC)

i like the self-referential humour, but agree with Tony and Kotniski about it being more of a distraction than a help here.

how about "Avoid linking strings of adjacent words, since that undermines the value of each link"? (It doesn't really "make reading difficult", does it, so i'd like to retract that from my earlier suggestion.) Should a "whenever possible" or "more than two" or something like that be added? and if an example really seems necessary:

Phrases like Pattie Boyd, a 1960s English model and rock muse" should have the less pertinent links removed, and/or should be worded differently.

i'm not really convinced an example is necessary, though ... Sssoul (talk) 15:11, 12 May 2009 (UTC)

• Indeed, I agree that an example is unnecessary; however, there was a reasonable two-word example in the long-standing point in MOSLINK (or whatever is was called before) that cautioned against merely adjacent links, since it would be ambiguous as to whether there was one compound-word link or two single links, etc. I don't see why this advice shouldn't be retained. Tony (talk) 15:57, 12 May 2009 (UTC)

I don't know about anyone else, but for me this (avoiding serial links) is one of the most important pieces of advice for this page, and applies just as much when there are only two adjacent links. I think it deserves not only an example or two, but also hints about how it can be avoided. Eliminating links isn't always the right answer - sometimes it's better to rephrase the sentence so the links are no longer adjacent, or find a single more specific link that does the job of two (like US flag instead of US flag).--Kotniski (talk) 16:14, 12 May 2009 (UTC)

Actually we have all this in the page at present - fifth bullet point under General Principles.--Kotniski (talk) 16:22, 12 May 2009 (UTC)

smile: ah good! so that's settled, then. 8) back to the other ideas on the drawing board, in that case.

does that second reorganizational suggestion make sense to people? Sssoul (talk) 16:34, 12 May 2009 (UTC)

Back to other suggestions ...

I think General Principles does (and should) apply to both internal and external links. If we think it's better to separate them so that General Principles is about internal links only, then internal and external links become so separate that it would be better to handle them on separate pages. That would be an option, I suppose. Rd232 ^talk 02:52, 13 May 2009 (UTC)

well but: everything in the general principles section is about internal links, so there's no reason to separate it from the other sections about internal links. but i see no need to make two separate pages - Linking Part 1: Internal Links and Linking Part 2: External Links seems sensible enough, no?

meanwhile my excuse for not noticing that we already have the point about not linking adjacent items is that i've been looking at the revision on Rd232's page, where that point was omitted. still looking at that page: can that point be restored to the general principles section, please? and can the "first occurrence only" point be rephrased as Tony suggested and moved back to the general principles section? then can we remove the "link density" subsection? and can the points in the overlinking/underlinking section be rearranged so that the redlink advice doesn't interrupt the two suggestions about alternatives to removing links? Sssoul (talk) 05:29, 13 May 2009 (UTC)

I take the "build the web" point in General Principles to include external links to the World Wide Web (though currently this isn't elaborated on there, and it should be). Rd232 ^talk 05:21, 13 May 2009 (UTC)

Conclusion

So can I be bold and put my restructure live? And we go from there based on the further discussion of details which goes beyond restructuring? Rd232 ^talk 13:43, 21 May 2009 (UTC)

Well since previous discussion basically supported the restructure and no-one's objected here, I've implemented it. Also now the discussion about various details, prompted by the restructure which made people take a closer look at the whole thing, can perhaps proceed now. Rd232 ^talk 10:21, 24 May 2009 (UTC)

i thought we had agreed to remove the "link density" section, since the jocular self-referential example is not really communicative, and the notion of aiming for consistent density is not useful. can we go ahead and remove it? Sssoul (talk) 09:30, 25 May 2009 (UTC)

I'm not sure the argument is that clear; it got a bit lost in the big discussion. And the basic idea of "sensible" link density makes sense to me. Could you start a new section to discuss it? Rd232 ^talk 11:10, 27 May 2009 (UTC)

Since the topic has been brought up, should we consider clarifying the current wording with a more explicit statement along the lines of, "self-links should not be used to achieve boldface in the lead sentence"? -- Tcncv (talk) 19:15, 24 May 2009 (UTC)

considering the weight of consensus, it would seem helpful to record that briefly in the guideline. The RFC hasn't been open that long, but without any support apart from the originator of the idea, WP:SNOW surely applies by now. Rd232 ^talk 20:06, 24 May 2009 (UTC)

Yes. Part of the problem with the MOS is that it is not clear enough sometimes; let us not make that mistake here. Dabomb87 (talk) 00:37, 25 May 2009 (UTC)

It may be useful to change the appearance of the selflinks (except those in navboxes) to avoid any future confusion. For instance, changing their color to green may work. In addition, they will be easier to find if their color is different. Selflinks in navboxes can remain bold. Ruslik (talk) 08:24, 25 May 2009 (UTC)

Why not change self-link display globally? Wouldn't it be equally a navigation aid in navboxes? Rd232 ^talk 11:56, 25 May 2009 (UTC)

I am not against, I am only trying to create a proposal that will not be opposed. Ruslik (talk) 12:42, 25 May 2009 (UTC)

Well, OK. But I'm not sure how easy it is to make a different display behaviour for navboxes. (Possible, sure.) Rd232 ^talk 14:19, 25 May 2009 (UTC)

You can look here. Ruslik (talk) 15:02, 25 May 2009 (UTC)

OK. But I'm a bit weary of this topic at the minute... Rd232 ^talk 15:14, 25 May 2009 (UTC)

We should be cautious here. There are some nav-box like constructs out there whose selflink behavior is likely expected to be the same as in nav boxes. For example the ((WP help pages (header bar))) template used on the WP:Help and related pages. Navbox like constructs also appear often in infoboxes, such as Everyone_(song) (at least before a bot came through and indiscriminately cleaned these out). -- Tcncv (talk) 15:56, 25 May 2009 (UTC)

The code in my simple.css file is actually limited to the main (ns-0) namespace. So, it won't affect the help namespace. Namespace other than main can be treated seprately, in my opinion. As to infoboxes, an exception can be made for them much like for navboxes. Ruslik (talk) 18:33, 25 May 2009 (UTC)

I don't think we should legislate against a bad idea simply because a single user is violently pushing it. I would consider that policy creep. --Hans Adler (talk) 18:13, 26 May 2009 (UTC)

I agree. Instead of legislating, it is better to change the style of selflinks to preempt any confusion. Ruslik (talk) 09:42, 27 May 2009 (UTC)