Wikipedia talk:Disambiguation/Archive 52

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.

In rare cases where a qualified title is still ambiguous and has a

primary topic

, it should ...

More often, a qualified title that is still ambiguous will have no

primary topic

, and ...

I do not see why a partial disambiguation should be less likely to have a primary topic than would an undisambiguated title. What’s the thinking here? —В²C ☎ 21:52, 6 September 2019 (UTC)

I agree that, as the guideline is written, they are not likely to be rare. Stating they are rare certainly doesn’t make them so.

However, as stated in the close of the above RFC, there was consensus that the standard should be higher for finding that a parenthetically disambiguated title has a primary topic than normal. That’s what would make it rare.

I propose adding the following sentence to the start of INCDAB to better reflect the consensus:

A parenthetically disambiguated title may have a primary topic, but the threshold for identifying one is significantly higher than for a title without parenthetical disambiguation.

I’m not clear how the closer above found consensus for Option 2, given the nearly even split among participants. Leaving that aside for the moment, there was a fair amount of middle ground. Whether you get there by saying that primary topic takes precedence, but with a significantly higher threshold than normal, or by saying that parenthetically-disambiguated titles shouldn’t have primary topics, but allow for rare exceptions, the result is about the same.--Trystan (talk) 22:47, 6 September 2019 (UTC)

Very interesting.

There are two types of disambiguation, natural and what I'll call formal disambiguation.... parentheses, commas and any other conventions we use (are there any?). And there are significant differences.

The purpose of any article title is twofold... to make the article easy to find, and to make it easy to wikilink to it.

Now we don't require that a formal disambiguation should be used in sources (it may be, but we don't care), or that it be a likely search term (and those two things are much the same thing really). We do of a natural disambiguation. And similarly, nobody is likely to want to wikilink the formal disambiguation, instead they wikilink the base name by using the

So the bar for Primary Topic might not even be the same for formal as opposed to natural disambiguations. We should not assume that it is. Andrewa (talk) 02:58, 7 September 2019 (UTC)

• The purpose of any article title? What about the purpose of titling the article? To state what the topic is. In journalism, the style is for a title to be a hook (like DYK). In fiction, the purpose of the title is to tantalise. In scholarship, the purpose of a title is to simply and directly state the topic of the content below. Easy wikilinking can be achieved with redirects, like shortcuts. If
  Paris, France, editors can easily wikilink Paris, no need to compromise the article title. If the purpose is for logical searching, if you want to complement the superior AI search engine algorithms, then what you are looking for is keywords. —SmokeyJoe (talk
  ) 06:18, 7 September 2019 (UTC)
  • So you're saying that one purpose of the article title is to give the article a title? Yes, that's trivially true... sort of, like x=x, which is a profound truth in some contexts... but I don't see the significance of the statement that an article title is an article title in this context.
  • This is not journalism or fiction. Scholarship, hopefully yes. And yes, the purpose of a title is to simply and directly state the topic of the content below. But I'd add clearly, and if the title is ambiguous, then its meaning is unclear. That's the whole problem. And yes, wikilinking can be by shortcuts.
  • For the umpteenth time, nobody is proposing to move Paris! So agree no need to compromise the article title.
  • Good point about keywords. We should probably make more use of them. Perhaps include the non-hidden category names in the HTML metadata of an article page? Or do we already do this? It seems logical to do so.
  • The Wikipedia article title has two consequences of which I'm aware in the outside world. It appears in the URL, and it appears in the title HTML tag for the page (and in both cases it's the whole title, not just the base name). I'm guessing that's not likely to change, and that it's the title tag that has most impact on external searching. Our internal search of course uses the article title directly. Andrewa (talk) 12:53, 7 September 2019 (UTC)
• I was the one who added "rare" since it was clear from the consensus and the closing statement that PDABs have a higher threshold than base name primary topics. I would note that naturally disambiguated topics can have primary topics in the same way as other articles at their best title since natural disambiguation is use of another term. Comma disambiguation for settlements in the US (and to a lesser extent in Australia and Canada) is often considered natural disambiguation (and sometimes for UK churches and streets) but for other topics like settlements in other countries (and similar) comma disambiguation is the same as using brackets, see Harris, Outer Hebrides for example. Crouch, Swale (talk) 08:46, 7 September 2019 (UTC)
  • Yes, comma disambiguation can be a form of
    natural disambiguation
    if it appears in sources, as in theory could parenthetical but that is less likely.
  • And that's not at all surprising, is it? These conventions weren't pulled out of the air, they were and are themselves natural English similar to what we'd seen elsewhere, and which we hoped and believed that English speakers would understand.
  • But the point is that we don't require any sources for these (full) titles anyway, because of our own naming conventions. In this sense they are treated as
    formal disambiguations, and so that's what we should call them whenever they could be either formal or natural and are therefore in a sense both. Andrewa (talk
    ) 22:33, 7 September 2019 (UTC)
• There's no reason to say they're rare. That gives the impression the practice is discouraged, which it isn't. If they're rare then let them be rare, but we don't have to explicitly say so. I suggest this be reverted.  — Amakuru (talk) 22:51, 7 September 2019 (UTC)
  • It is discouraged to the extent that the RFC consensus was for a higher standard than for regular primary topic. Without that, the guidance is not reflective of the consensus.--Trystan (talk) 23:39, 7 September 2019 (UTC)
  • I agree that there is (and should be) a higher standard for PDABs (particularly talking about parenthetical disambiguators here). I believe that has also been evident in the RM discussions in recent years, not just in the mentioned RFC. —BarrelProof (talk) 21:59, 13 September 2019 (UTC)

Why a higher standard?

I don’t understand. Let’s say there are two uses of Foo and they’re both films, but one is a well known American academy award winner and the other an unrelated obscure Korean film. The page views are about 4,000/day vs three a day. So the well known one is the primary topic and is at the base name while the foreign obscure one is disambiguated. Okay, but now say there is a third use of Foo, say a major well known city with 12,000 views per day, and that is the primary topic at the base name. So now the American film does need disambiguation and the decision is about whether it’s the primary topic for Foo (film). Why should the standard be higher than in the previous scenario in terms of how it was decided that it was the primary topic of Foo? —В²C ☎ 06:32, 8 September 2019 (UTC)

Because PDABs add an unnecessary layer of complication and confusion and there was a clear consensus that the standards are higher. Crouch, Swale (talk) 07:45, 8 September 2019 (UTC)

With a an undisambiguated title, the main consideration is how to direct the significant number of users that search for and end up at "Foo". But if you have "Foo" (the city and primary topic), "Foo (American film)", and "Foo (Korean film)", virtually nobody will search for or end up at "Foo (film)". So where it redirects to doesn't really matter much.

Let's say "Foo (US film)" gets 4,000/day and "Foo (Korean film)" gets 1,000/day. That meets the standard primary topic test; the first film is significantly more likely to be sought than the second. So we move the US film to "Foo (film)", which is a slightly shorter title. It doesn't help anyone get to their article any faster, and in fact will mean some increased portion of the 1,000 people seeking "Foo (Korean film)" will now end up at the wrong article. (For example, "Foo (film)" will now show up in the search box autocomplete. Users looking for the Korean film would likely click that, where they would not have clicked "Foo (US film)".)

So in choosing a primary topic for "Foo", the question is about how best to direct people searching for "Foo". In choosing a primary topic for "Foo (film)", the question is whether slightly shortening the title for one topic is worth the increased number of users reaching the wrong article. Totally different considerations, hence totally different standards.--Trystan (talk) 15:46, 8 September 2019 (UTC)

Trystan put it well, in my opinion. —BarrelProof (talk) 21:59, 13 September 2019 (UTC)

This is probably (I hope) a very quick discussion (for a change).

It's claimed here that We disambiguate article titles based on the articles that exist and could exist.

I don't think that's even remotely true. Very interested in other views.

It's now a moot point at the RM, as the contributor has agreed to create the stubs they say they have the sources to support. Just interested in the principle. Andrewa (talk) 21:09, 29 August 2019 (UTC)

Huh? We do have guidance on putting red links in disambiguation pages (

WP:REDHAT

), so what is supposed to be done when an existing article is accumulating links meant for a topic that does not yet have an article?

WP:SPECIESOUTCOMES). I'm generally loathe to confer primary topic status when genus names are shared; most genera of plants and animals are niche topics. I suppose a butterfly may more likely to be primary than a moss, and a tree more likely to be primary than an insect. But we still lack articles for many genera, and articles for a moss or fly may well be written before articles for a buttefly or a tree. I don't see any reason to leave the moss at the base title, accumulating links meant for the butterfly just because the butterfly article hasn't been written yet. Plantdrew (talk

) 23:09, 29 August 2019 (UTC)

Thanks, very interesting! So there are circumstances in which it's best to anticipate that these other articles will be created (perhaps despite

wp:ball

). Good point.

And that could arguably be the case in the RM in question, in that the supporters had undertaken to create the other articles, and claimed to have sources to support these articles. And we assume good faith in this of course.

I still think the RM is out of process, and the argument invalid. These other articles should be created first. Andrewa (talk) 22:57, 30 August 2019 (UTC)

I've seen something very like the sentence you quoted in a guideline somewhere. Naturally enough, I can't find it again.

I may come across five or ten redlinks a day on DAB pages which are orphans or which link to deleted articles, and delete most of them with enthusiasm. However, per

WP:IAR

I do not delete orphan redlinks where the DAB page itself makes a plausible case for notability (e.g. a river or mountain which does exist, but doesn't seem to have an article in any Wikipedia).

I've seen numerous DAB pages which included a redlinked genus. On almost all of them (or nearly), the genus entry could be justified per

WP:DABMENTION by a bluelink to a family (biology)-type article which mentioned it; which I added if it was lacking. I don't recall seeing a DAB page which included two redlinked genera each in a different kingdom (biology)

; though it wouldn't surprise me if such exist.

There is also the issue of unaccepted names, which are present in the older literature and are therefore both plausible and very necessary search terms. I have seen DAB pages which contained both an accepted and an unaccepted name within the same kingdom. As an unusual example, the DAB page Baeria contains two unaccepted genera, one a plant and the other an animal; and a see-also from a misspelling, which was how I came across it. Narky Blert (talk) 19:48, 31 August 2019 (UTC)

Also very interesting! Yes, I think it used to be in policy or guidelines but may have been removed. With so many policy and guideline pages, and so many editors discussing and changing them, and no easy way of searching the histories that I've found (suggestions welcome), it's a configuration management nightmare.

The problem I see is, while we assume good faith we can't always assume competence. The contributors concerned in this RM seem to have limited English and I suspect they don't have a good grasp of what sources are required to support an article... I've tried to gently question this. I haven't even suggested that the article instead be rescoped to include all the various meanings of the current title, but it does seems a possible

BCA

candidate.

Of course we're here for all English speakers, not just native speakers, and sources don't need to be in English. But we haven't got that far at the RM either, and I don't see that we will. I still think that the obvious course of action is, create the foreshadowed articles (good non-deletable stubs will do) and take it from there. Andrewa (talk) 20:15, 31 August 2019 (UTC)

An article that could exist but does not yet is fairly likely to meet

Yes, but sometimes those topics may be mentioned in two or three or even more articles. With the one bluelink per entry "rule" stringently enforced by a lot of editors, I think it's usually better to let the search engine do its job. Station1 (talk) 02:51, 1 September 2019 (UTC)

For quite a few terms, particularly multi-word phrases, the search engine does not do it's job at all. It turns up many irrelevant results and buries the relevant ones. bd2412 T 12:15, 6 September 2019 (UTC)

Tell me more. An example would be good. Andrewa (talk) 06:19, 7 September 2019 (UTC)

The contributors concerned in this RM seem to have limited English : it is not my case. For lots of contributors, if the article could be created, there are no reason for using a title without a parenthesis, despite the fact the others articles does not exist at this time. --Panam2014 (talk) 16:33, 1 September 2019 (UTC)

despite the fact the others articles does not exist at this time... I would say despite the fact that the other articles do not exist...... Do you get my meaning? There are many similar mistakes in the discussion.

My point was not that you're in any way disqualified from contributing here. But where it comes to understanding the finer nuances of policy, you're at an obvious disadvantage, and I think that shows in the discussion too. Andrewa (talk) 11:34, 4 September 2019 (UTC)

I think that if someone involved only in this disambiguation discussion, and not in that of the article under RM discussion, considers him/herself sufficiently uninvolved in that discussion, then it would be helpful to everyone if s/he could close the RM for those of us who are involved, rather than let it drag on infinitely... :) Boud (talk) 03:04, 6 September 2019 (UTC)

• This is a version of the old discussion, "does the world outside Wikipedia exist?"... which is answered each day when the content contributor slaves below decks chained to the oars create new notable articles from the real world of reliable sources. In ictu oculi (talk) 09:34, 6 September 2019 (UTC)

Very interesting! Discuss the impact of this at User talk:Andrewa/Primary Topic RfC#Articles that exist and could exist. It might surprise you as much as it does me. Andrewa (talk) 17:00, 10 September 2019 (UTC)

Not disambiguation in that case; "Canadian" is part of the name. "The Canadian Parliamentary Press Gallery (CPPG) is a self-governing corporation that consists of accredited journalists who cover Parliament and other Ottawa-based governmental organizations and institutions." [1] [2] - Station1 (talk) 04:57, 11 September 2019 (UTC)

Disagree... "Canadian" is part of the name... Well, yes, in the sense that any

natural disambiguation is part of the name. But Parliamentary Press Gallery is also a perfectly good name for the article, except for being (potentially) ambiguous, as the existence of the redirect shows. Andrewa (talk

) 10:41, 12 September 2019 (UTC)

No,

Canadian Mounted Police would be a good name for Royal Canadian Mounted Police, even though it redirects there. It's not a description of a press gallery that happens to be at the Canadian parliament, it's a proper noun name of an organization. Station1 (talk

) 20:28, 13 September 2019 (UTC)

• I disapprove of all the talk above of "Articles that could exist". That set is too large, and includes articles that could, but should not, exist. Instead, much of the above is agreeable if read as "Articles that should exist". The standard for articles that should exist is written into WP:Red link. --SmokeyJoe (talk) 06:35, 11 September 2019 (UTC)
  • Agree that Articles that should exist (my emphasis) is a better phrase. But disagree that the two phrases differ significantly in meaning in this context. By an article that could exist we mean one that would survive AfD, and that article should exist as well. Wikipedia is incomplete without it, so the set is no larger or smaller. And IMO that's exactly what WP:Red link says too. And on the other hand, could carries the extra connotation that these articles don't currently exist, which is the intended meaning, while should doesn't IMO carry that connotation quite so strongly. (And just BTW, could wasn't my choice of term, it was the term used in the RM comment that started all of this.) Andrewa (talk) 10:29, 12 September 2019 (UTC)
• I'd say we don't disambiguate articles, we disambiguate topics. Most of the time the topic will have its own article, sometimes it won't have an article yet, and at other times it will never have one because it's not notable by itself, but it's still treated at sufficient length within another article. – Uanfala (talk) 12:16, 12 September 2019 (UTC)
  • Interesting observations. I'd say that strictly we disambiguate article titles. A topic that's not notable enough, or for which we just don't have enough material, for an article still needs to meet a sort of notability standard, otherwise we are giving it
    undue weight by having a redirect by its name. I very much like the idea that we disambiguate titles that are potentially ambiguous, that is there are other articles that should exist and hopefully will, someday, but don't yet, or even redirects by that name that should exist. But I thought that was contrary to our practice and policy. It seems it's not contrary to current practice and consensus. I'm unconvinced about policy, so far. Andrewa (talk
    ) 13:08, 12 September 2019 (UTC)
  • It's the title that's ambiguous but this is really about relationships between titles and topics. If they corresponded 1:1 then life would be easy. Redirects mean that having many titles for one topic is no problem either. The problem we're solving occurs when a reader searches (broadly defined) using a title which has multiple title:topic relationships. Certes (talk) 13:44, 12 September 2019 (UTC)
    • Exactly. (Although I suggest you avoid describing a
      surjective
      ... maybe that's a quibble and perhaps unusually IMO this is the fault of my beloved mathematics rather than just another illogical quirk of English, but I have often seen these terms confused.)
    • More specifically, the problem occurs whenever a reader seeks (is that a better term than searches?) a topic using an ambiguous name which they do not consider ambiguous (that is, it has multiple title:topic relationships but they regard one of these as what we call a Primary Topic, which trivially includes the case that they think that there is no other possible meaning but isn't restricted to that case), and we have (at some point in history) chosen a different topic to be primary. Andrewa (talk) 21:34, 12 September 2019 (UTC)
      • Not just readers but editors too. We have too many articles about a primate paid a guinea to celebrate mass in a hamlet by a ford. Certes (talk) 21:52, 12 September 2019 (UTC)

@Andrewa: for all of the others articles, all should exist. So I will create the others articles but it is not necessary to create the others articles to move the first article. --Panam2014 (talk) 19:36, 14 September 2019 (UTC)

This discussion continues at #Pre-emptive disambiguation below. Andrewa (talk) 19:54, 14 September 2019 (UTC)

The case in point

The RM that started this is closed leaving the article at the undisambiguated name. So no pre-emptive disambiguation has occurred.

And nor have any of the other articles been created. Sovereignty Council of Sudan (1956–1958), Sovereignty Council of Sudan (1964–1965) and Sovereignty Council of Sudan (June–July 1965) remain redlinks at this time.

And nor has the scope of the article at

Maybe somebody will find something relevant in the guidelines, but in my opinion it's all down to common sense. If you have one article with a potentially ambiguous title, the question is how likely is it for another article to be created. If 1) the topic area is one that's relatively saturated with existing articles (pop culture), 2) there is a very large number of similar entities that are judged not to be notable (songs), then yes, it's not usually likely that disambiguation will become needed at some point. If on the other hand, you're creating an article in a topic area with relatively few articles (say, geography of Southeast Asia), and there is a very large number of notable entities that dont' yet have articles (say, populated places), then pre-emptive disambiguation might be a good idea. Another factor is how likely the name is to be distinct. Long name with unusual consonant sequences or short name made up of common sounds?

As for the case in the RM discussion, I think it's a matter no so much of pre-emptive disambiguation as of the title's appropriateness with respect to the scope of the article. If the title were "Retirement compensation arrangements" (a generic term), but the article were about

proper name for a particular kind of savings account used in Canada), then there would be a mismatch in scope. In such cases I think it's best to have the scope clearly declared with a parenthetical disambiguator (like "(Canada)"), rather than rely entirely on the surreptitious use of title case. – Uanfala (talk)

00:55, 14 September 2019 (UTC)

This is somewhat off-topic, but I feel obliged to say that Wikipedia uses sentence case, not title case, for the names of its topics, and

proper name. The mere fact of it being plural is a strong clue about that. As stated in the article on the subject, a proper name identifies a single entity, and there are many retirement compensation arrangements, just as there are many individual retirement accounts. A class of entities, such as the class of retirement arrangements that are defined under subsection 248(1) of the Canadian Income Tax Act or the class of retirement accounts defined in subsection 401(k) of the U.S. Internal Revenue Code, is not a proper noun. —BarrelProof (talk

) 00:48, 16 September 2019 (UTC)

I also thought that pre-emptive disambiguation was explicitly prohibited, which is why I thought that the discussion at #Articles that exist and could exist would be short and simple. This is exactly the question i was asking there.

But can't find where. One of the things missing from Wikipedia is the ability to search previous versions (or if it's there I haven't found it, either). I guess we'd want to throttle such a search carefully to avoid colossal server loads (and creating a DoS paradise!).

But the more I think about it the more I think pre-emptive disambiguation is a good idea, anyway. If you think that an article title is ambiguous in the most general sense, you shouldn't use it. That was the letter of the very first cut of WP:AT and perhaps has never changed.

Wikipedia:Unnecessary disambiguation links to at least one other relevant page and the related talk pages might be of interest too. And RMs regularly remove unnecessary disambiguations. Perhaps we are not entirely consistent in this. Andrewa (talk) 20:01, 14 September 2019 (UTC)

@Andrewa: for some subjects, pre-emptive disambiguation is really necessary when both articles have the same importance. --Panam2014 (talk) 16:03, 15 September 2019 (UTC)

Agree. But I don't think that's the question here. The questions are, what is our current policy and practice when (for example) only one of two articles currently exist, and what should it be? And there seem to be various opinions on both of these questions. Andrewa (talk) 20:33, 15 September 2019 (UTC)

Yes, I don't think the term "pre-emptive" applies when there are already two articles on Wikipedia. Pre-emptive disambiguation is about what to do when only one of the hypothetically confusable topics is currently discussed on Wikipedia. —BarrelProof (talk) 21:25, 15 September 2019 (UTC)

Exactly. It's a very good term, and that's one reason this section is so much more productive than my attempt at #Articles that exist and could exist above. Andrewa (talk) 21:53, 15 September 2019 (UTC)

• Preemptive disambiguation can be useful when all relevant topics have the same lack of notability. For example, redirects tagged {{R from song}} are helpful because they can be used to populate categories such as Category:xxxx songs, Category:John Doe songs, Category:Songs written by John Doe and Category:Songs produced by John Doe. In such a case, one or more of "Snafu" (Snafu song), "Snafu" (Fubar song) and "Snafu" (Bohica song) might be bluelinked redirects on the DAB page Snafu, and Snafu (song) a redirect to it tagged {{R from incomplete disambiguation}}; even though none of the songs has an article.

(I have been involved in exactly such a case. There was a

WP:NSONG; and it is now a redirect.) Narky Blert (talk

) 17:08, 17 September 2019 (UTC)

I'd like to see the specific example, but yes, it sounds like a good use of preemptive disambiguation.

The reason for wanting the specific example is, I'd like to consider whether it was within current guidelines to preemptively disambiguate in this case, and if not whether it's adequately covered as an example of the occasional commonsense exception. Andrewa (talk) 10:48, 18 September 2019 (UTC)

My memory was imperfect. It was Tomorrow's Dream (song), where the RM was to remove the qualifier. There may or may not be preemptive redirects relating to entries in the DAB page Tomorrow's Dream (I haven't checked everything), but all seven entries could have a useful qualified and categorised redirect (four songs, one album, two novels). Narky Blert (talk) 19:06, 18 September 2019 (UTC)

• There is an elephant in the room which was never noticed in the discussion (
  WP:SONGDAB

That discussion was all about readers coming in blind using the searchbox, trying to find an article.

It did not address the problems caused by bluelinks which point to the wrong place. (This is a widespread and well-known problem with

WP:SIAs and {{surname

}} pages, among others.)

A reader who clicks on a bluelink in an article should be taken to the right place, not to an ambiguous or – still worse – a misleading and wrong place.

An editor, an enthusiastic fan of the band Tarfu, adds a link in that article to "Snafu" (song), their best-known number. However, the band Fubar recorded a completely different song called "Snafu" – and "Snafu" (song) already exists, either as a standalone article or as a redirect to Fubar. Confusion worse confounded. Narky Blert (talk) 21:57, 18 September 2019 (UTC)

Songs are somewhat of a special case, because they often share titles with other songs and are sometimes only mentioned very briefly in articles (e.g., in a list of tracks on some album). I'm trying to bring up here the wider question of pre-emptive disambiguation as a more general issue across Wikipedia. Anyhow, that RFC has been archived, and is found here. The headline for the discussion is misleading, because it assumes that any song that does not have a stand-alone article written about it is therefore a non-notable song. In some cases, a song might be discussed extensively in an article that is about a band or about an album, perhaps because the song was particularly important to the history of the band, and we should not jump to the conclusion that the lack of a separate article about the song means, ipso facto, that the song is not notable. I don't think a separate (largely duplicate) article should be created about the song just to prevent it from being considered non-notable on Wikipedia. That creates a maintenance headache. But I have no problem with the idea that if only one song with a particular name is really discussed to any significant degree anywhere on Wikipedia, then it should be considered the primary topic among the other candidates where a song might just be mentioned once somewhere in a list of tracks in some discography. —BarrelProof (talk) 00:00, 19 September 2019 (UTC)

An actual example. I've just cleaned up the mess which was

99.9 to 99.9 (Mar del Plata) for precision. The frequency is used everywhere, with many countries having several stations. (2) Pop (Argentine radio station) and Top (Argentine radio station). Even though Pop (radio station) and Top (radio station) are redlinks, I could foresee possible problems with them; those names look very generic. Narky Blert (talk

) 07:21, 19 September 2019 (UTC)

A reader who clicks on a bluelink in an article should be taken to the right place, not to an ambiguous or – still worse – a misleading and wrong place. Exactly. And surely that's a good reason not to have articles at ambiguous (or just plain wrong) titles? A topic that's primary for some is not primary for others, however we choose P T. Editors who assume a different P T (including those who don't even think that the term is ambiguous) will of course innocently wikilink to the wrong article. If the mislinking is to a DAB we can warn them and we do and mostly they'll fix it themselves, and even if they don't fix it the wikignomes can easily find and fix mislinkings to DABs and they do, and the

pipe trick makes it easy and natural to do so. But on the other hand, if the mislinking is to a wrong article then nobody gets any warning. Andrewa (talk

) 23:24, 20 September 2019 (UTC)

MOS:DABMENTION the other topics. Incorrect blue links in the article space just get fixed like any other incorrect blue links in the article space. -- JHunterJ (talk

) 19:38, 23 September 2019 (UTC)

Incorrect blue links in article space do not get fixed. They tend to remain for ever, unless a DABfixer with a suspicious mind investigates them. As an example from my to-do list, Orel, Russia: 13 bad incoming bluelinks.

An editor once posted on my Talk Page, requesting that I refrain from posting {{

dn

}} in a particular infobox because they made the article look ugly, and members of a specialist WikiProject would spot and fix such errors soon enough. Two days later, I checked and corrected a bluelink which had triggered my spideysenses in a different instance of the same infobox. Guess who the most recent editor had been, a week or two earlier?

It may be about time I checked Tetrahedron again. It usually has a few links-in intended for Tetrahedron (journal). Narky Blert (talk) 04:13, 25 September 2019 (UTC)

Although it has little to do with pre-emptive disambiguation, I've listed some of the most commonly mislinked articles in User:Certes/misdirected links#Examples. Certes (talk) 11:02, 25 September 2019 (UTC)

Incorrect blue links do get fixed. They may tend one way or the other, but editors edit. Regardless, errors elsewhere are a bad reason to make errors here. -- JHunterJ (talk) 17:25, 25 September 2019 (UTC)

I assume this is in reply to this edit above by Narky Blert. Yes, of course they do sometimes get fixed, but only when editors stumble upon them by accident. On the other hand, links from articles to DAB pages are detected by DPL bot, and are far more likely to be fixed as a result. Andrewa (talk) 00:49, 26 September 2019 (UTC)

Incorrect links do not get fixed unless someone takes specific trouble to go looking for them.

Earlier this year, Certes and I ran a project to repair links to surname articles for taxonomists in articles about species. There were about 1,500 of them, and it took two weeks and one trout. They included several explicitly linked to (surname) and (disambiguation) pages through those qualifiers. The task included a fair amount of multilingual searching. As a bonus, we were able to find by eye and repair some other bad links which turned up in those articles and elsewhere; e.g. to Amman, Baker, Chiron, Hedge and Jordan. I don't know how long those links had been around; but on looking at our worklist again, I found one error which had been introduced in 2011, and I doubt that was the oldest. We didn't even attempt to look at mononymic footballers, where the problem may be at least as bad.

That was the tip of only one iceberg. Narky Blert (talk) 10:22, 26 September 2019 (UTC)

If you're just going to go with hyperbole, nothing ever gets improved on Wikipedia unless someone takes specific trouble to improve it. We call those someones "editors". But we don't degrade the encyclopedia for the users to benefit the editors, and pre-emptive disambiguation is a degradation of the encyclopedia. -- JHunterJ (talk) 13:45, 26 September 2019 (UTC)

Which element of my post was hyperbole? Narky Blert (talk) 21:28, 26 September 2019 (UTC)

that incorrect link do not get fixed unless someone fixes them. Perhaps "tautological" would be a better description. Wikipedia does not get edited unless someone edits it. -- JHunterJ (talk) 12:36, 27 September 2019 (UTC)

Incorrect links get fixed when someone notices them. This can happen in (at least) two ways. Links to dabs get noticed because DPL bot flags them, both in talk page messages to the adding editor and in a convenient daily report. Links to the wrong article don't; we have to go looking for them. This is harder and takes longer than going through the DPL list. I've fixed quite a few recently but each group means thinking up a new search criterion: I'm currently fixing links to surnames from templates which don't list surnames. I'm sure I've missed far more than I've fixed, and I fear that no one else has found them either. Certes (talk) 14:01, 27 September 2019 (UTC)

Disambiguation isn’t needed unless there is another topic dealt with on Wikipedia to distinguish. [3] (my emphasis) Interesting comment.

Precision is defined at

WP:AT as "The title unambiguously identifies the article's subject and distinguishes it from other subjects." The implied scope must be something like "the set of subjects covered on Wikipedia" or "the set of subjects that could be covered on Wikipedia, according to current content guidelines." It becomes meaningless if the title needs to be sufficient to unambiguously identify the topic among all possible subjects in the world.--Trystan (talk

) 01:56, 26 September 2019 (UTC)

Exactly. There are in fact three sorts of topics...

• Those that have articles (and presumably deserve them).
• Those that don't deserve articles.
• Those that do deserve articles but don't yet have them.

The question is, where and how do we discuss that third sort? I think if there's evidence that a topic is of the third sort, then that evidence should be enough to create a non-deletable stub, and that's what we should do rather than produce that evidence at an RM. It doesn't take appreciably longer, and improves Wikipedia. And it helps to weed out claims that the evidence exists by people who are guessing or worse (but possibly in complete

here) they can instead do something more useful. Andrewa (talk

) 04:38, 26 September 2019 (UTC)

WP:DISAMBIG defines as its scope "there is more than one existing Wikipedia article to which that word or phrase might be expected to lead". DMacks (talk

) 05:03, 26 September 2019 (UTC)

Exactly.

WP:DISAMBIGUATION are the policy and guideline against pre-emptive disambiguation (not just an interpretation of them). Otherwise we have William Shakespeare (English playwright born 1564) just in case another English playwright named William Shakespeare explodes on the scene next year. Future Wikipedia can handle its issues, and moves for changes in ambiguousness and primaryness are cheap. -- JHunterJ (talk

) 13:45, 26 September 2019 (UTC)

I think it would lead to endless unresolvable conflict regarding article titles. It’s relatively easy for editors to agree when a title is ambiguous between two or more existing articles. When one editor can say “We need to disambiguate this title, because it conflicts with some non-notable topic I am familiar with”, there is no common foundation on which editors can build consensus.--Trystan (talk) 14:05, 27 September 2019 (UTC)

That case is relatively easy... the non-notable topic doesn't need an article, so we just need rough consensus that it is non-notable. The problem is when there is a notable topic that doesn't yet have an article but should have.

And it seems to me, rereading the above, that then the question is not so much how we decide this (as in what evidence we consider) as where we decide (as in, is an RM a suitable place for the notability discussion even if the article in question doesn't yet exist). WP:DAB seems clear on this... Disambiguation is required whenever, for a given word or phrase on which a reader might search, there is more than one existing Wikipedia article to which that word or phrase might be expected to lead. (my emphasis) But

WP:AT#Precision

is not nearly so clear... Usually, titles should unambiguously define the topical scope of the article, but should be no more precise than that. No mention of whether this ambiguity arises whenever notable topics exist, or whether it's only if the articles on them exist. Does the policy at WP:AT override the guideline at WP:DAB? Or does the fact that the policy fails to say let the guideline override it? Do they need to be completely consistent? Why have guidelines at all if we're going to put everything in the policy?

I still think that, on practical grounds, these discussions shouldn't be part of WP:RM. If the evidence exists that there's another notable topic, then create that other article before going to WP:RM. Neatest procedure by far, and what WP:DAB already says. Andrewa (talk) 16:56, 27 September 2019 (UTC)

Most wikilinks use the middle initial but that may just be because it's the article title. If the common name omits the E. then the page should move. It will need a

WP:RMTR because the redirect has history as a stub duplicate article. Certes (talk

) 08:06, 3 October 2019 (UTC)

I've

WP:BOLDly moved the article to James Gunn (astronomer). I don't see many sources using the middle initial.  — Amakuru (talk

) 13:07, 4 October 2019 (UTC)

You have missed nothing; you have made a gud catch.

As Certes says, the first place to post is at

WP:PAGEMOVER, and post a reasoned request at Wikipedia:Requests for permissions/Page mover. Even if you're granted PAGEMOVER rights: if you're in any doubt, post at RMTR. Narky Blert (talk

) 22:54, 4 October 2019 (UTC)

What you have prepared so far looks very reasonable to me. Some comments:

• Re #1: No, typically not. Ptopics are ultimately chosen to help the reader get where they want to go. If an overwhelming majority of readers think of term T as referring to topic X, then T should at least be a
  WP:PRIMARYREDIRECT
  to X, if not the title of X. This is true even if there are some RS that say that actually by definition T refers to topic Y.
• Re #2: The best other example I can think of is
  Holocaust
  .
• What about just
  WP:NCPLURAL
  , but hey, it'd be a natural way to disambiguate.
• The number of mistargeted links to Climate change is IMO the strongest evidence that the current primary topic is badly chosen. Using Special:WhatLinksHere has the issue of bringing in a bunch of navbox inclusions, but these search results paint an even clearer picture. (I used User:PrimeHunter/Source links.js to generate that query)
• At
  User:Femkemilene/sandbox#First_rename_climate_change
  you say The climate change page will be either a disambiguation page to both this article and global warming at the end of step 1, or a redirect to global warming.. To be clear, I think it's important that this be decided as part of that RM (and ideally your nom should take a stance on which option you're advocating for).

Colin M (talk) 15:46, 7 October 2019 (UTC)

Thanks for that list of mistargeted links and the rest of the comment :). Really hope this will work out with the sometimes heated discussions around this topic.

Climate variability in scientific terms means something else, so doesn't really work in terms of precision. I think I'll go for climatic changes. I fear there will be consensus to move, but no consensus how. There's even people not wanting the topic as a separate article at all.

talk

) 21:06, 8 October 2019 (UTC)

1. Talk:Union Station#Requested move 14 August 2019
2. Wikipedia talk:Disambiguation/Archive 51#Application of WP:DIFFCAPS to WP:DABNAME
3. Talk:Friendly Fire#Requested move 22 August 2019
4. Talk:Absolute Zero (disambiguation)#Requested move 11 September 2019
5. Talk:Dark Matter (disambiguation)#Requested move 11 September 2019
6. And of course, the requested move directly above this section

In discussions 1-3, consensus was that capitalization alone was sufficient, while in discussions 4-6, there was no consensus either way and from what I see, very noteworthy opposition. If there's a given article

• Freezer Burn
• Special Effects
• Warning Shot
• Union Station
• Central Station
• Friendly Fire

Disambiguation pages carrying "(disambiguation)" where all non-primary topics are capitalized

• Sea Urchin (disambiguation)
• No Man's Land (disambiguation)
• Blue Moon (disambiguation)
• Trade Secrets (disambiguation)
• High voltage (disambiguation)
• Event horizon (disambiguation)
• And many more in the previous section

Comments

• It appears that right now, there is no standard for disambiguation pages of this nature. I was originally on the side of "(disambiguation)" is required to differentiate DAB pages from articles. However, after the recent discussions and reading over the arguments for not having it, I'm unsure. Nevertheless, I believe that not having a general rule and having each DAB page be evaluated individually is a horrible option. Usually with naming conventions, there is a generally accepted way to name every different kind of page, with IAR applying for exceptional cases. If there's no guideline here, we'll just keep debating this forever. I'd very much like to see what the community thinks here. Eventhorizon51 ^(talk) 23:35, 16 October 2019 (UTC)
  • I think you missed the prevailing argument in the discussion above which can be summarized as, "it depends". A large number of participants expressed the opinion that some of those primary topic subjects are much more likely to be sought using title case than some of the others. This opinion being as prevalent as it is suggests we're stuck with looking at these on a case-by-case basis. In fact, that's what the closer essentially said. Our only hope is maybe we can convince a consensus that it's okay to send title case searchers to the dab page. Then they'll learn. Not sure why there is resistance to this. Like you said, the primary topic will be easy to find as it will be listed at the top. --В²C ☎ 01:57, 17 October 2019 (UTC)
• "Do differences in capitalization alone sufficiently distinguish"?

A. It depends. Homoglyphs (0 and O; 1, l and I) are not good enough. DIFFCAPS where the difference is between title case alternatives is not good enough. --SmokeyJoe (talk) 02:43, 17 October 2019 (UTC)

• On

  "(disambiguation)" is required to differentiate DAB pages from articles

  what are the counter arguments? --SmokeyJoe (talk) 02:47, 17 October 2019 (UTC)
• Why, Eventhorizon51, did you create so many editable section headers. Is each meant to be a thread starter? --SmokeyJoe (talk) 02:49, 17 October 2019 (UTC)

• Yes, this is almost unreadable. The close asked nominators to nominate RM's on each item separately, not to continue debate on this topic on this page. Randy Kryn (talk) 03:05, 17 October 2019 (UTC)

• Removed some section headers. There is that better? Eventhorizon51 ^(talk) 03:10, 17 October 2019 (UTC)
• Just because the closer stated that there is no restriction on making individual RM proposals immediately does not mean general discussion here cannot continue. That said, I don't think there is much interest in a general solution because of a desire to support title case search for at least some but not all primary topic topics even though they don't have proper names. --В²C ☎ 17:22, 17 October 2019 (UTC)

• My opinion, which I've stated in many of the recent discussions, is that title difference in capitalization, is not sufficient. Creating two different targets for basically the same title is a wrong approach. Whether the target should be the PT or a dab page, is less important to me (I do have an opinion on that as well though), as long as both result in the same target. I'll also point to the RFC from least year at DIFFCAPS which while resulted in no consensus, had a majority of editors in favor of its removal, so I find it hard quoting a policy which does not even have majority community consensus. --Gonnym (talk) 17:31, 17 October 2019 (UTC)
• Like so much on WP, "it depends", as SmokeyJoe observed. There are various cases where upper-casing is arguably sufficient, and cases where it is clearly not. The most common type of the latter case is when a term is sometimes capitalized by some people (usually in a
  AReaderOutThataway ^t/_c
  10:44, 22 October 2019 (UTC)

How much more likely is "much more likely than any other single topic"? In a recent RM discussion (Talk:Kawésqar#Requested_move_4_September_2019) with a TWODABS situation in a dispute about whether there was a primary topic both sides acknowledge that one topic was favored 1.64/1 in terms of page views. In your experience and opinion is that always/often/usually/sometimes/rarely or never enough of a ratio to meet the "much more likely than any other single topic" criteria? In my opinion it's always, and it seems like the community agrees much more often than not. Am I wrong? --В²C ☎ 18:58, 2 October 2019 (UTC)

For example, if one topic is getting 60% of the hits, and the others are getting 40% all together, doesn't that make the one topic "much more likely" to be sought "than any other single topic"? Maybe this should be clarified at

WP:PRIMARYTOPIC. --В²C ☎

19:30, 2 October 2019 (UTC)

Well let's consider the extreme case where there's only one other topic. If I mix up 6 red balls and 4 green balls in a bag, and pull one out without looking, would I be "much more likely" to pull out a red ball than a green ball? I probably wouldn't use that language. I would agree with Certes that 2:1 is about where that phrase starts becoming appropriate. Colin M (talk) 19:34, 3 October 2019 (UTC)

I am dissatified with the decision at

WP:NCLANG ("Where a common name exists in English for both a people and their language, it is most often the case that neither is the primary topic"). Narky Blert (talk

) 21:37, 4 October 2019 (UTC)

• Colin M, great analogy and analysis! In that 6:4/red:green mix, red is 1.5 times as likely to be picked as is green. If we have 62 green and 38 red then we're at a 1.63 ratio. With a distribution of 51/49 (1.04) certainly red is is merely more likely. But I think anything above 1.5:1 crosses the line into much more likely. I mean, is 62 merely more than 38, or is it much more than 38? If you had pay of 38/hour and it was bumped to 62/hour, would that much more pay, or merely more? --В²C ☎ 21:53, 4 October 2019 (UTC)

• The words "likely" and "probably" do not have precise quantified numerical definitions. They are words used when not only the answer is uncertain, but so is the question. Is it likely that a meteorite killed the dinosaurs? Trying to quantitative these adjectives is to miss the point. Read more at https://www.mathsisfun.com/definitions/probability.html --SmokeyJoe (talk) 05:22, 3 October 2019 (UTC)
  • I’m aware of the ambiguity in the current wording. That’s the point. Is there consensus about a more precise meaning that the wording should reflect? That’s what I’m asking. —В²C ☎ 06:38, 3 October 2019 (UTC)
    • It's better to first consider page views of all involved as illustrative, and then to look closely at long term significance. Page views are very easy to obtain, and often correlate with long term significance, but where they don't its probably because of recentism or bias due to certain users having cheaper, easier and more frequent access. Then, if there's any doubt, there's no primary topic. I see you are still pushing the argument that no primary topic means the DAB page is at the base name and the DAB page is the wrong page for many. I agree, and the answer is to repudiate MALPLACED, never put a DAB page at a base name but instead PRECISEly title DAB pages with a "(disambiguation)" suffix. No more unexpected DAB pages from Go box autocomplete selections! Do this and the problems causing you angst will go away. --SmokeyJoe (talk) 08:00, 3 October 2019 (UTC)
      • The weighting of the long term significance criteria relative to the likelihood of being sought criteria is a separate issue from the question of exactly how likely is "much more likely than any other single topic"? It's also separate from the issue of what to do once no primary topic is established. --В²C ☎ 16:28, 3 October 2019 (UTC)
• I know most people are wont to avoid giving specific numbers here, but I think it's good to have some sort of benchmark. for me, if one of the articles in a twodab situation receives twice as many views as the other one, then it's likely to be a primary topic. I wouldn't consider a ratio of 1.65/1 as sufficient. – Uanfala (talk) 10:27, 3 October 2019 (UTC)

  That sounds about right. If only two topics have significant interest then I'd put the dab at the base name unless the ratio is at least 2:1; perhaps more. Reading

  WP:PRIMARYTOPIC literally, a ratio of 1.01:1 can suffice if the 49% is spread thinly amongst several other topics, but I'd be reluctant to follow that advice. Certes (talk

  ) 10:43, 3 October 2019 (UTC)

  Interesting. In my experience the community tends to be more lenient with the likelihood criteria in a two-dab scenario than in multi-dab scenario. After all, even if the split is 51/49, at least 51% are taken directly to the article they're seeking, while the remaining 49% are only one hatnote link click away from the article they are seeking. But if we find no primary topic in that situation and put a two-dab dab page at the base name, then 0% are taken directly to the article they're seeking, and 100% are one dab page link click away from the article they are seeking. Everyone is worse off or neutral; no one is better off. Why do that? --В²C ☎ 16:28, 3 October 2019 (UTC)

  Because being taken to an article on a different topic is worse than being taken to a dab. The dab is shorter and clearly a dab, whereas the "wrong" article may well be long and look superficially similar to the article sought. To take the hackneyed example of New York, a reader seeking the city who is sent to the state will load 2.4 MB of text and graphics and might well pluck out a statistic believing it to apply to the city. Suppose, by way of analogy, that someone asks me directions but I'm a tourist here myself. A dab is like admitting that I don't know and pointing at a nearby map; a bad primary topic is like sending them in the wrong direction. Certes (talk) 16:42, 3 October 2019 (UTC)

  Because being taken to an article on a different topic is worse than being taken to a dab. So, I think our disagreement comes down to that. I understand your position in theory, but I can only imagine it happening in practice. Never actually experienced it. Even with New York City/state the first lines are "The City of New York, ..." and "New York is a state in ..." respectively. The language in most of our articles requires a level of intelligence and comprehension much higher than what is required to make the distinctions necessary for this. Even if I granted landing on a different topic is a little bit worse than being taken to a dab, which I would only do for the sake of argument, that wouldn't come close to making up for the cost of sending more than half to the dab page instead of directly to the article they are seeking. --В²C ☎ 17:35, 3 October 2019 (UTC)

  I agree with the first sentence. If an article on a different topic is barely worse than a dab then we should put borderline primary topics at the base name. If there's a bigger gap then we should be stricter about only doing this for very clear primary topics. Certes (talk) 17:42, 3 October 2019 (UTC)

  Agree but note the if in If an article on a different topic is barely worse than a dab.... And that antecedent is not generally true. NYS/NYC may be hackneyed but it's a spectacular example of what can go wrong. B2C says I can only imagine it happening in practice. Never actually experienced it. But in practice that's exactly what happened at New York. Readers wanting NYC were being taken to a long article. They are now taken to a much shorter DAB, and that's an improvement. But under our P T guidelines they should be taken straight to NYC, which is almost certainly the P T. This would then send those wanting NYS to an even longer article. Having the DAB as the

  base name is the current and best solution. But is this an occasional exception, or does it (as I have concluded) indicate a more general flaw in the way we have long treated ambiguous article titles? Andrewa (talk

  ) 18:26, 11 October 2019 (UTC)

  First of all, no, the city is not the primary topic under our P T guidelines. Despite the significantly higher page views there is no evidence that those numbers reflect relative likelihoods of being sought between the city and the state. Since any searcher for the city is likely to know "new york" is ambiguous with the state name they are likely to search with "new york city" or "nyc". But if it was the P T, if most people searching with "new york" were looking for the city, then yes, the minority looking for the state with "new york" would be penalized by being taken to the large city article rather than to a dab page. Once. Presumably they'd remember to type "new york state" next time... And the one time they had to wait to load the city article, at least they'd have the hatnote link to take them to the state article once it did load. BFD. Furthermore, if anyone is on such a slow connection that this even matters (a cheap motel in Wyoming with SnailPace WiFi?) they should be clicking Search, not Go, after entering "new york" or whatever their search term is. That way they get to see search results and decide what to do from that instead of htting Go and loading the matching page. This whole large article loading consideration that you and especially SmokeyJoe love to bring up sure seems overblown to me. --В²C ☎ 19:33, 11 October 2019 (UTC)

  I'm not sure I agree with you on the hypothetical scenario of 51% vs. 49% thinly sliced. At least in the case where the 51% topic is something that people are widely aware of. If all the other competing topics are very obscure, then readers and editors will generally know better than to think that they can search for or link to them without a disambiguator. (I'm not sure how often scenarios like this come up in practice though. John Adams (disambiguation) has a lot of links, but their summed views are more like 15-20% rather than 50%. Curious if someone can think of a better example - maybe an acronym?) Colin M (talk) 19:50, 3 October 2019 (UTC)

  Kawésqar, Kawésqar language. --В²C ☎ 20:27, 3 October 2019 (UTC)

  That looks sensible as the article at the base name is obviously not about the language and has a direct hatnote to the

  one other, closely related, topic. However, several similar cases have changed recently to a two-entry no-PT dab. There's also at least one bad incoming link (from KBF via the questionable redirect Kakauhua) that I'm not sure how to fix (Spurious languages#kbf?) Certes (talk

  ) 20:35, 3 October 2019 (UTC)

  But that's just two pages. The scenario I was talking about (which Certes raised as an aside above) was one article having a little over half of the overall views, and the remaining 50%-ε being spread thinly over many other topics, each having a very small number of page views. Colin M (talk) 00:43, 4 October 2019 (UTC)

  C. elegans (disambiguation) could be an example. It lists a primary topic and about 1750 alternative meanings. Certes (talk) 11:28, 4 October 2019 (UTC)

That’s a great example of needing to look at introductory no-context use versus repeated in-context use. In every case, “C.” refers to the C. things being talked about, and elegans (a common Latin word, elegant) is one of many of these things. —SmokeyJoe (talk) 12:22, 4 October 2019 (UTC)

Surprisingly, the

primary redirect C. elegans has incoming links from only 17 articles. Seven mention Caenorhabditis; ten have no explicit context saying what C. stands for. All links seem correct. The worm is the standard organism for many types of research, and may have well over 51% of C. elegans references. Certes (talk

) 12:39, 4 October 2019 (UTC)

Other candidates are Ace, Atlas, Herald, Star, Telegraph and popular names such as James Brown and John Williams. Certes (talk)

• New to this thread, but to answer the original question the simple answer is that there is no fixed page-view percentage at which primary topic is asserted. Page views are just one of many things we consider when determining what, if anything, is primary. Each situation is assessed on its own merits and
  Talk:Fernandinho_(footballer), I find myself arguing a PTOPIC on 90% pageviews while others seem to prefer a dab. But each case is individual.  — Amakuru (talk
  ) 12:55, 4 October 2019 (UTC)

  So, in other words, the answer to the question — How much more likely is "much more likely than any other single topic"? — is.. "it depends"? That is, in some cases it might be 50% more likely and in other cases it has to be twice as likely or perhaps even ten times more likely to be considered "much more likely"? So, then, what determines that in each case? Note that the question here is not asking about primary topic - it's a more specific question; a sub-question if you will. Are you saying that you want flexibility on this question so that you can justify your sense (based on who-knows-what) about whether a given topic is or is not the primary topic? Wouldn't it be better to identify exactly what that who-knows-what is, document it, and just go by that? --В²C ☎ 17:15, 4 October 2019 (UTC)

  WP:NWFCTM notwithstanding, since editors are human we are apt to intuitively decide an issue, and then use reasoning to justify our intuitive decision. That applies just as much to deciding primary topic as to anything else. That's just how the human mind works. The best way to counter this tendency is by tightening up the decision-making criteria to discourage post hoc rationalizing. I'm not saying we need a specific fixed page-view percentage, but I do think it would be helpful to come up with as tight a range as we can achieve consensus for. That said, in the real world many decisions are made by choosing arbitrary lines. Most notably the voting ages, driving ages, draft ages, drinking ages... They are all arbitrary lines that are within the "about right" range. Having specifics helps make decisions about which the correctness does not matter nearly as much as having some clear line. Imagine if we did say it's 60% of all page views. Would that be so bad? Maybe that would mean some topics would be primary that we felt shouldn't be, and some wouldn't be that we would feel should be, but so what? We would agree because the page views would dictate whichever it was, not our subjective opinions. And so all these cases that we go on and on about despite there being no one right decision would be resolved very simply. Wouldn't that be better? --В²C ☎

  18:53, 4 October 2019 (UTC)

Well, if we introduced that rul then one of the first moves would be Apple Inc. to Apple, and the fruit article to

WP:5P1, the most fundamental rule we have. It might just be my liberal western bias coming to the fore, but I feel like there's something empirically right about prioritising the fruit over the tech company, and systemic bias aside, I don't actually think that Wikipedians are that bad at making these calls.  — Amakuru (talk

) 19:14, 4 October 2019 (UTC)

I don't think so; certainly not necessarily. That's the point. We would still have the long-term significance criteria to consider. I'm just saying that long-term significance should not affect how we interpret how likely "much more likely than any other single topic" is. Even if we recognize that Apple Inc. is much more likely to be sought than is the fruit (and surely it is), we can still decide to override that with long-term significance. But wouldn't it be better to at least make the likelihood assessment less subject to personal and systemic bias? Let's be clear that that is what we're doing, overriding with long-term significance, rather than redefining the likelihood of "much more likely" at every turn. --В²C ☎ 19:24, 4 October 2019 (UTC)

...closely followed by the likes of Prince (musician), Monopoly (game), Billboard (magazine) and Arsenal F.C.. Certes (talk) 19:29, 4 October 2019 (UTC)

A separate issue is the guidance we provide about how to decide what to do in all such cases when the two criteria conflict, leaving that decision to the whims of whoever shows up, every time. In this supposed guideline, we provide practically no guidance on this. But, that is a separate issue to the question I'm asking in this discussion. --В²C ☎ 19:59, 4 October 2019 (UTC)

WP:PTOPIC

is not clear; but IMO the two tests (usage and long-term significance) should be cumulative not alternative. You have to pass both to be a PTOPIC.

There was a recent discussion/kerfuffle at Wikipedia talk:Article titles#RfC about articles on three digit numbers on this sort of topic. My conclusion from it: if a three-digit number has a PTOPIC, it is the AD year; if the year is not PTOPIC, then there is no PTOPIC. Narky Blert (talk) 23:20, 4 October 2019 (UTC)

Narky Blert, by that interpretation nobody searching with “Boston” (for example) would be taken to the article they seek (most likely the Massachusetts city, Boston) but to the dab page. That’s not the community consensus. —В²C ☎ 03:14, 8 October 2019 (UTC)

@Born2cycle: IMO too much emphasis is placed on reading rather then editing. People typing into the search box will always get where they want, sooner or later. One more or fewer click is neither here nor there. The danger and damage arises from bad bluelinks. Readers who click on bluelinks should always be taken to the right place – and that's the end of the matter.

UK readers who type 'Boston' into the search box might be mildly surprised to find themselves looking at an article about

Perth, Australia; but they can deal with it. However, bluelinks like this one are intolerable. Narky Blert (talk

) 04:27, 8 October 2019 (UTC)

Perth, Western Australia, is a perfectly good title for the large city, Perth is not perfectly good. --SmokeyJoe (talk

) 00:01, 9 October 2019 (UTC)

In the last year or so, I've corrected 133 articles which linked to the wrong Perth. Clearly I'm not the only editor fixing such problems, and there will be more that we've not spotted. Certes (talk) 00:11, 9 October 2019 (UTC)

That itself is a compelling reason to put Perth WA at a PRECISE title. Another rationale is that editors mislinking to Perth are an indicator of many readers who are similarly confused. --SmokeyJoe (talk) 00:23, 9 October 2019 (UTC)

Narky Blert, thank you. This may be of interest to SmokeyJoe and Andrewa as well. This really helps understand your perspective. I think it comes down to a difference in priorities. You're primarily concerned with bad bluelinks, and believe reducing the chances of bad bluelinks should be a higher priority than maximizing the chances users get directly to the article they seek when searching with a given term. Right? I too am concerned with bad bluelinks, but I think existing primary topic criteria already handles it well enough, and do not think polishing the bad bluelink turd (pardon the expression) is worth reducing the chances users will get directly to the articles they seek when searching with a given term. There is no way to eliminate all bad bluelinks, and we have practices that mitigate the problem. First and foremost, since anyone can edit WP, anyone who encounters a bad bluelink can fix it, just like you did with Perth in the diff above. Secondly, hatnote links, like the one to Perth, Scotland at the top of Perth, allow users to get back on track quickly when misguided by the occasional bad bluelink. WP is by definition a constant work in progress, and nobody should be surprised by hiccups like occasional bad bluelinks. Trying to reducing the incidence of bad bluelink occurrence is certainly admirable, but adding measures that do so marginally with the side effect of making navigation to countless highly sought articles less efficient is, IMHO, not a tradeoff beneficial to most users. --В²C ☎ 16:17, 8 October 2019 (UTC)

@Born2cycle, Certes, SmokeyJoe, and Andrewa: This is a hypothetical example, expanding on my idea of searchers and readers. I haven't seen an actual example; it would be difficult to locate one; but it's pounds to pennies that they exist.

Imagine that we have several articles on people called

WP:PTOPIC; all the other articles have qualified titles, and are on the DAB page. Readers clicking on links to Joe Schmuck from non-political articles might be surprised to learn that this multi-talented man also was a moderately successful boxer, played two seasons of professional baseball, starred in silent films, wrote erotic novels, described a species of beetle found in the Amazonian jungle, and was an early advocate for the legalisation of cannabis. They are likely to assume that his other talents have been omitted from the PTOPIC article, and to go away with totally the wrong idea. Narky Blert (talk

) 19:33, 11 October 2019 (UTC)

The whole concept of primary topic is designed to avoid exactly such a scenario. That's not to say something won't ever slip through the cracks, but, in general, if primary topics are selected correctly, per the criteria, it shouldn't happen. Also, I think we should give our readers a little more credit. If they only know of a Joe Schmuck as a politician in the early 20th century, and when they click on his linked name they land on an article with "(actor)" (or whatever) in parentheses, plus nothing about him being in politics, and dates likely not lining up, we have to assume they will figure out that's a different Joe Schmuck. --В²C ☎ 19:59, 11 October 2019 (UTC)

"we have to assume they will figure out that's a different Joe Schmuck". Why? we are an encyclopaedia, not a party game. Narky Blert (talk) 22:11, 11 October 2019 (UTC)

James Brown was born in 1862 after a brief career as a newspaper editor. Although best known for his musical career, he was also a notable basketball player and gridiron quarterback. Certes (talk) 20:18, 11 October 2019 (UTC)

He also acted in The Dirty Dozen. Narky Blert (talk) 22:07, 11 October 2019 (UTC)

At least that one links to the right guy. The articles I mentioned link incorrectly to the singer at the base name. (Yes, I should just

fix them.) Certes (talk

) 23:08, 11 October 2019 (UTC)

• My gut instinct is to refer to the 68–95–99.7 rule. 50.1-68% is more. 68-95% is much more (and is the PT discretionary range). 95-99.7% is very much more. >99.7% is almost certainly. A PrimaryTopic may be found at 68%, and may not be found at 95%, Pageviews are an unreliable easy statistic and other things need serious considerations. —SmokeyJoe (talk) 22:04, 4 October 2019 (UTC)

“More” means more than all independent others combined, not just the next strongest. Non-independent others may be excluded from consideration, or may be taken as contributing the claim of the strongest. —SmokeyJoe (talk) 22:07, 4 October 2019 (UTC)

In general, yes, but the clause in question here says "much more likely than any other single topic". So it's explicitly excluding any others. --В²C ☎ 22:21, 4 October 2019 (UTC)

• Regarding the 68% rule... So 67% is just "more", not "much more"? If one has 67% the most the other could have is 33%. So 67 is not much more than 33? If I said a bike ride is 33 miles and it turned out to be 67 miles, would that be much more, or just "more"? If your pay jumped from 33 units to 67 units would that be more, or much more? If the price for a meal of fish is 33 and for lobster it's 67, does the lobster dish cost much more than the fish, or merely more? Maybe it's just me but in every other context I can think of, 67 seems to be much more than 33... If one dog weighs 33 lbs and the other 67 lbs, does the larger dog weigh much more, or just "more" than the smaller one? --В²C ☎ 22:21, 4 October 2019 (UTC)

I think it is not reasonable to compare quantities and probabilities in terms of the same “more”. Consider uncertainties. 33 miles is much more than 32 miles if you are talking about measuring 32 miles to the nearest foot. Probabilities, or likelinesses, don’t work like quantities. 33 miles will be larger than 32 miles everytime. If Pageviews A-B are 67-33, the probability of the next view being A is only 67%, there is a 33% chance of being wrong. That’s pretty bad. —SmokeyJoe (talk) 12:56, 5 October 2019 (UTC)

Sure, SmokeyJoe, but the likelihood of being wrong is 100% if you send 0% to the article they seek. I'll take being wrong 33% of the time over being wrong 100% of the time. Every time. --В²C ☎ 16:15, 7 October 2019 (UTC)

Could you please rephrase that? Sending no one somewhere is like a division by zero error. Is your wrong article the imprecisely titled disambiguation page? Who is "sending"? Why not empower the reader to make informed choices, by having PRECISE titling? --SmokeyJoe (talk) 00:57, 8 October 2019 (UTC)

Every time a user enters a search term and clicks Go we send them somewhere by virtue of how we arrange our articles and choose our titles. That somewhere is either the title they’re seeking (right place), or somewhere else (wrong place). Our goal should be to arrange our articles and select our titles to send the most people to their right place the highest percentage of times we can. If using a particular base name search always results in the user being sent to a dab page, then none (0%) are sent to the right place. That’s not a good result especially if it’s possible to send more than half to the right place by locating the most likely to be sought article at the base name. —В²C ☎ 03:02, 8 October 2019 (UTC)

The "wrong place" can only happen, without the reader knowing, if the title is ambiguous, and is a problem if the reader has a different PT assumption to editors. I disagree with your "Our goal should be to arrange our articles and select our titles to send the most people", that's presumptuous and patronizing, the goal should be to provide titles that inform the reader of what page they are selecting. That means PRECISE titling, including for DAB pages. I think you are mostly referring to DAB pages being the "wrong page"? If so, I agree, I personally dislike going to a DAB page that I anticipated being an article. "If using a particular base name search always results in the user being sent to a dab page" is obviously a failing of using a basename for a DAB page, the solution is to not use basenames as DAB pages, the solution is not to force a PT decision where there is no PT. --SmokeyJoe (talk) 03:42, 8 October 2019 (UTC)

Yes. The problem occurs when the reader has a different PT assumption to editors. This happens more often than you might think. Having heard that a concert featured Madonna, Prince and Fish, it's easy to presume that those might be the titles of musicians' biographies. The first one is a dab, with the entertainer listed prominently. The others have non-musical PTs which are obvious when you stand back and think but are not what you might presume when your mind is preoccupied with the context. Certes (talk) 10:29, 8 October 2019 (UTC)

SmokeyJoe, how is that goal presumptuous and patronizing? We simply do our best to determine what anyone searching with a given term is most likely seeking. If there is a topic that is most likely sought with a term, then searching for that term should take users to that topic. What's the alternative? Sending them all to a dab page. How is that better? It's certainly no better for the majority who could have been taken directly to the article they're seeking, and it's no better even for the minority that will end up one click away from their destination either way (by either dab page link or hatnote link). A much smaller minority will be an extra click away - but isn't it better to add clicks for a small minority than for the majority of those seeking with the term in question? How is it patronizing or presumptuous to send people searching with "boston" to Boston? And I don't see how not using basenames as DAB pages is a solution to anything. So we move Mercury to Mercury (disambiguation). What does that solve? What do we do with Mercury? We must redirect it to Mercury (disambiguation), right? So what difference does it make to not have the basename Mercury as a DAB page, but rather a redirect to the DAB page? --В²C ☎ 16:17, 8 October 2019 (UTC)

• Links to ill-chosen PTOPICs damage the encyclopaedia, are difficult to find, and rarely get fixed. Saving even a majority of readers one click from the search box is as nothing compared with the confusion or misleading information imparted by clicking on a bluelink which points to the wrong article. It's all about the readers. I generally look for a ratio of at least 8:1, and preferably at least 9:1, before agreeing to a PTOPIC.

Some examples. (1)

Salt Lake City, Kolkata

. The only reason that such bad links are few is that a couple of editors have undertaken the mindblowingly tedious task of checking and correcting those bluelinks.

I have corrected well over a hundred thousand (possibly close to one hundred and fifty thousand) links to DAB pages. They are easy to find;

Spidey-Sense suggested that something was wrong: most obviously, things like bluelinks to newspapers called Star or Sun

. Errors like that only get fixed if an editor spots one of them, and takes the trouble to look and see what others there might be.

I remember sorting out two articles about people from India which were differentiated only by the presence or absence of a fullstop after a personal initial; inevitably, the links-in were a mess. I have a bookmark folder in my browser called 'To-do' which contains potential problems of that sort which I haven't got round to looking at yet. (It currently includes, among others,

Tiempo de Amar and the redirect Tiempo De Amar, where the capitalisation may be a false way of distinguishing between two topics; but until I've dug into Spanish Wikipedia, I just don't know.) Narky Blert (talk

) 22:40, 4 October 2019 (UTC)

1. It implicitly assumes that the % of pageviews that come from a search is the same across all articles. This is certainly not true. For example, in this RM, I argued that articles on TV episodes are much less likely to be arrived at via search (rather than from wikilinks, or links from outside Wikipedia, including search engines), for the simple reason that the titles of individual TV episodes are not widely known (even among people who have watched the episode).
2. It assumes that the likelihood that someone looking for information about X will use T as their search term is the same as the likelihood that someone seeking Y will use T as the search term. In some cases, this is far from true, and leads to inferences of relative probabilities that may be off by one or more orders of magnitude from reality. For example, close to 100% of readers looking for information about
   Stevedore
   via search, perhaps 1% or less will have used the search term "Dockers" to get there.

Pageviews are a useful piece of evidence when evaluated in their proper context, but I think it's a bad idea to try to uncritically plug them into some math formula to make a ptopic decision. Colin M (talk) 03:25, 5 October 2019 (UTC)

• To answer the question of how we should interpret "much more", I think we need to understand what the purpose of that requirement is. And now that I think about it, I don't know. A primary topic needs to be searched for more than all the others combined, and much more than any one other topic. The first part isn't ambiguous, and just requires a simple majority. Consider two potential primary topic discussions, A & B. In both, the potential primary topic is likely to be sought 55% of the time, so it meets the first part of the test. In A, the remaining 45% is split between 9 other topics at 5% each. In B, the remaining 45% is all for a single other topic. So A clearly passes the primary topic test, while B (arguably) does not. But why treat these cases differently? In both A & B, the same percentage of people will end up incorrectly at the primary topic. The potential for incorrect links appears to be similar in both cases.--Trystan (talk) 13:46, 5 October 2019 (UTC)

  A difference is that in the 55%/45% case, some readers (and editors) may assume that the 45% topic is primary and expect to find it at the base name (or mislink it using the base name). In the 55%/5%×9 case, although it's unclear whether there is a PT or not, there's only one contender for the role. Certes (talk) 14:00, 5 October 2019 (UTC)

  Good point, Certes. A 55/45 situation is more likely to create bad bluelinks than is a 55/20/20/15 situation. However, I don't think an occasional bad bluelink is reason enough to not send the 55% to the article they seek, which is the result of putting the dab page at the base name. And, yeah, in the 55/45 situation anyone seeking the 45% one is more likely to expect it to be at the basename than is someone seeking a 20% or 15% target in the other scenario, but, again, I don't think the occasional sending of a user to the unexpected article is reason enough to not send the 55% to the article they seek. --В²C ☎ 16:23, 8 October 2019 (UTC)

Perspectives

I think Born2cycle has hit the nail right on the head with this edit in which I was pinged (thank you).

Narky Blert, thank you. This may be of interest to SmokeyJoe and Andrewa as well. This really helps understand your perspective. I think it comes down to a difference in priorities. You're primarily concerned with bad bluelinks, and believe reducing the chances of bad bluelinks should be a higher priority than maximizing the chances users get directly to the article they seek when searching with a given term. Right? I too am concerned with bad bluelinks, but I think existing primary topic criteria already handles it well enough, and do not think polishing the bad bluelink turd (pardon the expression) is worth reducing the chances users will get directly to the articles they seek when searching with a given term. There is no way to eliminate all bad bluelinks, and we have practices that mitigate the problem. First and foremost, since anyone can edit WP, anyone who encounters a bad bluelink can fix it, just like you did with Perth in the diff above. Secondly, hatnote links, like the one to Perth, Scotland at the top of Perth, allow users to get back on track quickly when misguided by the occasional bad bluelink. WP is by definition a constant work in progress, and nobody should be surprised by hiccups like occasional bad bluelinks. Trying to reducing the incidence of bad bluelink occurrence is certainly admirable, but adding measures that do so marginally with the side effect of making navigation to countless highly sought articles less efficient is, IMHO, not a tradeoff beneficial to most users.

The fundamental disagreement here is, they regard landing on a DAB as a loss, even if that DAB leads explicitly and obviously to the wanted article, with one more mouse click. I regard it as a win.

And that underlies all the rest. To me the that one mouse click is a trivial task that takes almost no time and no thought, and follows loading a DAB that takes minimal time and bandwidth. But to them it is an unacceptable burden.

I doubt that we can ever resolve this. It reflects a fundamental difference in our ways of reading Wikipedia, and our expectations of the way others do so. But we can at least understand it. I think I now do. Other comments? Andrewa (talk) 20:44, 8 October 2019 (UTC)

For a DAB page that’s just a few lines I suppose it’s not unreasonable to characterize the cost of having to land on it as “one mouse click [that] is a trivial task”, though I, for one, still find it to be annoying and definitely not a win. I’m much less bothered by landing on another article with a hatnote link to my desired article at the top because it’s just one line to scan; that truly is a trivial task, every time. But wading through the larger DAB pages riddled by categories with seemingly random headings organized in some unfamiliar fashion is a much more significant task which can be very annoying and time consuming. I think it’s important to spare our users from that as much as reasonably possible. And that is arguably the main purpose of identifying primary topics and arranging our articles accordingly. —В²C ☎ 15:53, 10 October 2019 (UTC)

You don't seem to be consistent in this. The DAB at The Americans (disambiguation) (formerly at The Americans) has nine entries in all, and one subsection, a See also with four entries. It all fits on my screen without scrolling. To call it one of the larger DAB pages riddled by categories with seemingly random headings organized in some unfamiliar fashion would be ridiculous.

Yet you argued forcefully (and successfully) that it was better to have

base name

. As I said at the time, this seems rather bizarre to me. It benefits very few readers, and makes things more difficult for many more.

I can now see why you like it that way, but I don't see any reason to think that many readers had any problem before. Most of those who will now find the article at The Americans would have found it just as easily or even more easily at The Americans (2013 TV series). Andrewa (talk) 02:46, 11 October 2019 (UTC)

We agree that the Americans DAB page falls into the category that you characterize as requiring "one mouse click [that] is a trivial task". But how am I inconsistent in not wanting that DAB page at the base name? Remember, I characterize landing on it as "still ... annoying and definitely not a win". There is no way the article about the epic 2013 TV series could be easier to find than it is now: enter "the americans" and click Go. When doing that took you to the (even relatively small) DAB page, that made it harder. And annoying. Which is arguably why it was fixed. --В²C ☎ 19:44, 11 October 2019 (UTC)

Because you seem to be relying on assumptions on the way people use Wikipedia. There is no way the article about the epic 2013 TV series could be easier to find than it is now: enter "the americans" and click Go. Yes, that's your way of getting there. But you're a bit unusual in that you already know exactly what the article you want is called, and have returned to it frequently but can't be bothered bookmarking it, even when you find it so annoying to be taken to a DAB. Most unusual in fact.

But the most significant thing there is that you already know what the article you want is called. We also need to deal with the more common situation, that people don't know the article name. They may have a guess at it, and if for them the meaning of the base name is the same as Wikipedia's decision on Primary Topic, they may get it right. At the moment, you get it right for The Americans, not surprisingly because it was your idea to move the TV series article there. But if ever the P T changes (as was foreshadowed as an obvious possibility in the RM close) then you'll be taken to the wrong article. If it's a long, wrong article, you say that won't matter any more to you than being taken to a short DAB, but for most users it will.

Another way of getting there is to start to type The Americans and look at the drop-down list to see whether there's an article name that matches what you want. Previously, people who did this got The Americans (2013 TV series) on the list, a name which they were most likely to recognise as the article they want, but now they don't.

So it has become more difficult for these users. Do you see that now?

We could fix the "search Wikipedia" box (which some I know say isn't a search box at all - whatever) to always show redirects, and to always show DABS, and have hovertext or the like to explain what (disambiguation) means for those who may not know. But we have no control over other search engines etc.. The other way is, have the article at an unambiguous, recognisable name. That solves it all, except for those with a strong and to me unreasonable aversion to DABs. Andrewa (talk) 08:58, 14 October 2019 (UTC)