Archive 45 ← Archive 49 Archive 50 Archive 51 Archive 52 Archive 53 → Archive 55

Guidelines for title-and-name disambiguation pages

@Anthony Appleyard, WPGA2345, Spinningspark, Hugsyrup, and Lightburst: There are currently 1,148 pages in Category:Title and name disambiguation pages. Pursuant to the precedent established at Wikipedia:Articles for deletion/Capt. Price, I propose that we delineate some specific guidelines for these pages.

As I noted in the AfD, it is common practice in the real world for sources to refer to holders of certain offices or stations by their title and surname. These references frequently make their way into articles, complete with links. Having disambiguation pages at these titles makes it easier for readers to find particular individuals who might be known by this combination, while also helping to eliminate incorrect links. These pages provide better results than Wikipedia's search function, which will sometimes return every article containing the title and the surname, even where the words have no relation to one another in the article. Jumbled results from the search function bury those most relevant to a search for people who might actually be called by that title and name.

In terms of guidelines, I think that we should have a requirement that there be a reasonable propensity for the subject to be referred to by the title and name combination. Thus, as noted, people who are merely a sports team captain should not be listed with people titled "Captain", and people who are merely the president of a company should not be listed with people titled "President". In some cases, people have given names that in other contexts are a title (like Justice Smith). My thinking is that a person who is so named will usually be the primary topic of the term over someone who has a different actual name, but may be known by the title and surname. I would welcome any thoughts or refinements to what I have come up with so far. Cheers! bd2412 T 23:56, 16 July 2019 (UTC)

Hey, thanks for pinging me on this. I'm totally in agreement with what you propose, and I think the way to establish the 'reasonable propensity' that you suggest is simple enough - do sufficient reliable sources refer to them in that way, or not. I think there could be a sort of grey area with people who, say, held a military rank when they were younger but are rarely if ever referred to by that rank now. For example, I probably wouldn't put John McCain on a disambiguation page for Lieutenant Commander McCain. But again, that's easy enough to decide based on whether enough reliable sources refer to him in that way, or not. Hugsyrup (talk) 08:11, 17 July 2019 (UTC)
I definitely agree with your Lieutenant Commander McCain example, and would incorporate into the guideline that just because a person has held an office or rank or station does not mean that they would be known by it. I think that it bears noting that the higher the rank or the office, the more likely that the association will hold, irrespective of the time the position is held. For example, Benjamin Harrison, the first President Harrison, only held the office for a month, but will forever be known by it. bd2412 T 12:54, 17 July 2019 (UTC)
I've added entries to Captain Price (disambiguation) where the highest rank mentioned in the article is captain. I left in one officer promoted from captain to rear admiral: I wouldn't have added him, but I suppose he would still be the captain of his ship. Certes (talk) 13:12, 17 July 2019 (UTC)
"Lieutenant Commander McCain" actually does generate quite a number of hits in newspaper and book sources, in reference to activities he underwent at that point in his life. So that's a tricky one, as although he wouldn't typically be known as that now, it is a plausible link target for him. In his case he's probably the only notable McCain to have held that title so it could just be a straight redirect, but I wouldn't bar him from a list of other Lieutenant Commander McCains, if there were any...  — Amakuru (talk) 13:25, 17 July 2019 (UTC)
I think there is ample reason not to exclude McCain from such pages. In fact, this is just the kind of thing I was referring to in the AfD. Sure, just about all recent sources that refer to "Lieutenant Commander McCain" will go on to say he had a political career and became a senator. However, I am seeing sources from the 1960s and 1970s that refer to "Lieutenant Commander McCain". Anyone reading those sources, who don't know his story already, will not get it from the source they are reading because it is too early. So what are they going to search for in Wikipedia to see if he is notable? Also, I saw a source from the 1920s for "Commander McCain". Presumably this is John S. McCain Sr. and it is not beyond the bounds of possibility that an earlier source might refer to him as "Lieutenant Commander McCain" when he held that rank. SpinningSpark 14:05, 17 July 2019 (UTC)
I feel like we might be going down a rabbit hole based on a brief example that I didn't give a lot of thought to. My point was really that, as a general principle, merely having held a rank at one point in your life shouldn't be sufficient for default inclusion on these disambiguation lists. The criteria should be whether reliable sources refer to you as such. If there are sources that refer to McCain by his military rank then I'd have no issue with him being on the hypothetical list, and I apologise for choosing a bad example! That said, I do see another potential question which is that after giving the example I re-checked the John McCain article and it looks as if the highest rank he held was actually Captain. It's not impossible, then, for him or other similar individuals, that we could find sources referring to Lieutenant Commander McCain and Captain McCain, and quite possibly other ranks he held as well. How should those be handled? My view is, again, so long as there are sufficient sources using a rank then he can go on that list, even if it means he ends up on three or four different lists with different ranks. Hugsyrup (talk) 14:20, 17 July 2019 (UTC)
I think perhaps a useful rule of thumb would be, "what would people call them after they have died". The captain of the Titanic will always be "Captain Smith"; John McCain will probably always be Senator McCain. Someone like William Howard Taft would likely be called President Taft or Chief Justice Taft depending on the context, but it would probably be less common for him to be referred to as Governor Taft. bd2412 T 15:49, 17 July 2019 (UTC)
...except in sources that predate him becoming President Taft etc. My consideration here is for people trying to identify him from a mention in an old source. If there are some sources that use "Governor Taft" and fail to mention "President Taft", even if the majority of sources talk about President Taft, then it is useful to include him on a Governor Taft disambiguation page. SpinningSpark 15:56, 17 July 2019 (UTC)
I'm in broad agreement with everything that has been said; the points of difference are in detail.
I propose this test, which has been suggested above: was the person ever notably known by that title and name? (Which comes back again to reliable sources.) For example, there is no redirect from Lieutenant Kennedy to JFK, and I can see no reason why there should be. On the other hand, Taft would have been notable had he never been other than a state governor.
There's an example of promotion through the ranks with guaranteed notability at every stage in the English and Welsh legal system: Mr/Mrs Justice X (High Court) to Lord/Lady Justice X (Court of Appeal) to Lord/Lady X (Supreme Court), where there should be DAB pages or redirects for the earlier titles; because that's how they're named in the law reports. (Northern Ireland is, I think, the same. Scotland is different: all Scottish judges have the honorary title of Lord/Lady.)
(As an aside, I keep coming across pages which are fully or partly tndis wrongly categorised as hndis, often with the wrong sortkey (i.e. Name, Title instead of Title Name). They're worth watching out for.)
Major Major Major Major is, of course, a special case. Narky Blert (talk) 19:25, 19 July 2019 (UTC)

Adopting Narky Blert's language, I would propose wording this as:

A name and title combination should only be included on a page if the subject was notably known by that title and name, and this is reflected in reliable sources. For example, there is no redirect from Lieutenant Kennedy to John F. Kennedy, even though Kennedy at one point in his military service held this rank. Although many notable people with the surname Kennedy may have at some point held the rank of Lieutenant, none of them were notable for having held the rank. By contrast, William Howard Taft can be listed at Governor Taft and Justice Taft as he would have been notable even if that was the only office he had held.

bd2412 T 18:55, 21 July 2019 (UTC)

I endorse bd2412's suggested wording. I chose my words for the purposes of argument. bd2412's wording strikes me as a sound proposal for a WP:CONSENSUS both for DAB pages and for redirects. Narky Blert (talk) 21:11, 21 July 2019 (UTC)
Thanks - here is what we have so far, then:

Title-and-name disambiguation pages

It is common practice for sources to refer to holders of certain offices or stations by their title and surname. These references frequently make their way into articles, complete with links. Having disambiguation pages at these titles makes it easier for readers to find particular individuals who might be known by this combination, while also helping to eliminate incorrect links. These pages provide better results than Wikipedia's search function, which will sometimes return every article containing the title and the surname, even where the words have no relation to one another in the article. Jumbled results from the search function bury those most relevant to a search for people who might actually be called by that title and name.

A title-and-name disambiguation page should only be created if there are multiple individuals who can be included on the page. Furthermore, an individual should only be included on a page if the subject was notably known by that title and name, and this is reflected in reliable sources. For example, there is no redirect from Lieutenant Kennedy to John F. Kennedy, even though Kennedy at one point in his military service held this rank. Although many notable people with the surname Kennedy may have at some point held the rank of Lieutenant, none of them were notable for having held the rank, and it is therefore inappropriate to have a disambiguation page at that title. By contrast, William Howard Taft can be listed at both Governor Taft and Justice Taft (and can be the redirect target of President Taft) as he would have been notable even if any one of those was the only office he had held.

Having a title in some capacity is also not, by itself, sufficient to merit inclusion on the page. There must be a reasonable propensity for the subject to be referred to by the title and name in combination. Thus, people who are merely a sports team captain should not be listed with people titled "Captain", and people who are merely the president of a company should not be listed with people titled "President". In some cases, people have given names that in other contexts are a title (such as actor Justice Smith, politician Major Owens, and musician Earl Hines. Where this is the case, the person with such a name will often be considered the primary topic over any person with a corresponding title and name.

Cheers! bd2412 T 03:22, 23 July 2019 (UTC)
I like it; except that I would prefer "will often" to "should usually" in the final sentence, as less prescriptive.
The title in Duke Ellington was a nickname (as it was also in Count Basie and King Oliver); but, Earl in Earl Hines was a given name. Another example designed to confuse us is Lord Kitchener (calypsonian), whose performing name bore no relation to the little-known name he was given at birth, but was adopted from the most usual designation of Herbert Kitchener, 1st Earl Kitchener, as in the famous (in UK) WWI recruiting poster Lord Kitchener Wants You. Narky Blert (talk) 08:04, 23 July 2019 (UTC)
Sorry, returning to this after a bit of time away, but I endorse bd2412's wording, and nice work to you and Narky Blert and others on finding what seems like a neat compromise on a slightly tricky problem. Hugsyrup 12:46, 23 July 2019 (UTC)
I have implemented Narky Blert's proposed changes. Should this be added to Wikipedia:Manual of Style/Disambiguation pages#Specific entry types? bd2412 T 23:22, 23 July 2019 (UTC)
It needs to go somewhere, and in the absence of a ((tndis)) template I can't think of anywhere better. Narky Blert (talk) 07:30, 26 July 2019 (UTC)
Not quite. Regarding the Governor Taft example... That qualifies as "the subject was notably known by that title and name, and this is reflected in reliable sources", but I still don't think it merits inclusion. Here's why. Anyone in 2019 or later reading an old source published prior to his presidency that refers to him as Governor Taft knows who that person is and who he ended up being and would not search WP with "Governor Taft". So I'm not sold by the "the subject was notably known ..." wording. I think another qualifying clause should be there, something like, "...and is reasonably likely to be searched for today by that name name and title". --В²C ☎ 19:54, 26 July 2019 (UTC)
Why try to (a) overcomplicate a straightforward proposed guideline and (b) make life difficult for readers? There are already too many difficult-to-understand or ambiguous guidelines without adding another one. The fewer words and provisos, the better. Narky Blert (talk) 22:28, 26 July 2019 (UTC)
We are not solely concerned with what a person will search for. Editors writing on the period of Taft's governorship are likely to find sources from this period containing references to "Governor Taft", and could very well link Governor Taft with the expectation that it leads to the subject relevant to that period. bd2412 T 22:52, 26 July 2019 (UTC)
What about Governor Reagan? That redirects to Governorship of Ronald Reagan, not to Ronald Reagan. Seems like a reasonable handling of title/name; one that the the guideline should cover. --В²C ☎ 23:19, 26 July 2019 (UTC)
What about it? Unambiguous titles are not particularly relevant to discussions on the contents of disambiguation pages. Is there another "Governor Reagan" against which the current redirect should be disambiguated? 00:03, 27 July 2019 (UTC)
Governor Reagan? Stop farting about with 1980s programming games. It should be redlinked, and any attempt to land on that title auto-sent to the search engine. Don’t confuse autojavatitlecomplete with search. Some time around 2007, Wikipedia got a real internal search engine. —SmokeyJoe (talk) 13:49, 27 July 2019 (UTC)
Wikipedia's internal search engine does more harm than good when trying to find a specific office-holder by their title and surname. Look for "Justice Johnson", for example, and the results legitimately reflecting people who might actually be called this are swamped by irrelevant returns like Lyndon B. Johnson judicial appointment controversies, Johnson & Johnson, Texas v. Johnson, Lynching of Ed Johnson, and Social justice warrior. bd2412 T 18:53, 28 July 2019 (UTC)
User:BD2412, I don’t see the harm. The top result both with the internal search engine and the google search of Wikipedia is Justice Johnson. 2nd-3rd are not irrelevant, Lyndon B. Johnson judicial appointment controversies is a completely sensible result for the search term. Johnson was an important president involved in controversy over justice appointments. Are you trying to psychoanalyse the random Wikipedia user for what they want based on a two word search? I think you are overweighting word order. In any case, the search engines are dynamical learning algorithms, unlike the very primitive notion of search by title auto completion. —SmokeyJoe (talk) 23:01, 28 July 2019 (UTC)
It makes sense to the extent that Justice Johnson exists, and is the first result. Lyndon B. Johnson judicial appointment controversies is a completely sensible result for "Justice Johnson"? Not according to the way titles and names are used in the English language. Given that there are 32 people who might historically be known as "Justice Johnson", one would have to go deeper than that to determine whether the search is providing all of the likely results. Does "Johnson Justice" yield a different result? bd2412 T 00:50, 29 July 2019 (UTC)
a completely sensible result for "Justice Johnson"? Not according to the way titles and names are used in the English language.. You seem to working under the assumption that readers put “titles and names” into the search box as a search query. I don’t think that is normal, and I certainly don’t. I am personally frequently annoyed that the search box tries to auto complete to a title my search terms, and worse, in doing so it generates a list that hide the actual search option (the “containing ...”), pushing that below the fold. I have learned that pressing the search button *before* entering the search terms takes you to a search page with better functionality, and one that has recently been much improved in ease of use, but I think few are aware. Does “Johnson Justice” yield a different result? That shouldn’t be a concern, if that search query doesn’t return useful results, the searcher returns and searches something more detailed, the dynamically learning search algorithm notes this and learns and in future presents better results. Adjusting titles for searching seems to me to be a 1980s computer programming mindset. —SmokeyJoe (talk) 00:00, 30 July 2019 (UTC)
If all that you know of someone is that their last name was Johnson, and they served on the Supreme Court of a U.S. state, how would you find them using the search box? bd2412 T 00:45, 30 July 2019 (UTC)
What’s your point? I think I have forgotten why we are having this discussion. I often search for things like that, partial information. It is always a challenge. Johnson is a very common name. I don’t know which state. The Go Box would probably be awful, I think it feels better than it is because it takes you to other interesting things and you forget your original purpose. I entered your question into google and got William Johnson (judge) as the third hit. On a mobile device, searching Wikipedia directly tends to be frustrating. —SmokeyJoe (talk) 00:59, 30 July 2019 (UTC)
Perhaps we are talking about different things. The discussion here is about having guidelines for structuring title-and-name disambiguation pages. There was recently an AfD proposing to delete one, with the implication that they should all be deleted (we have about a thousand of them). The discussion resulted in a fairly resounding collection to keep these pages, but with some limitations on their content (the specific issue raised in the AfD was the inclusion of sports team captains, not commonly referred to as "Captain Foo", on a page of people titled "Captain"). That being the case, I thought it prudent to consolidate the rules governing these pages. I think the issue that brought you here was the redirect of Governor Reagan to Governorship of Ronald Reagan. Although this is not a disambiguation issue, I don't see how the Governorship of Ronald Reagan would not be the primary topic of the term Governor Reagan, absent any other governors named Reagan. bd2412 T 01:42, 30 July 2019 (UTC)

Is there anything else before I add the language above to the MOS? bd2412 T 01:13, 9 August 2019 (UTC)

Was added here. —SmokeyJoe (talk)
Sorry, I mentioned the specific intended location a few paragraphs above, but it is indeed better to specify that here as well. Thanks. bd2412 T 04:49, 9 August 2019 (UTC)

How much hatnote is too much?

Guidance per WP:D#Primary topic with two or more other topics:

If there are two or three other topics, it is still possible to use a hatnote which lists the other topics explicitly, but if this would require too much text (roughly, if the hatnote would extend well over one line on a standard page), then it is better to create a disambiguation page and refer only to that.

Is this too much, or okay:

This article is about the 2013 TV series. For the older series, see The Americans (1961 TV series). For the Robert Frank book, see The Americans (photography). For other uses, see The Americans (disambiguation).

See: The Americans.

--В²C ☎ 18:30, 8 August 2019 (UTC)

Hatnotes clutter the intro, per WP:1HAT usually only 1 or 2 should be used. If there are more then a DAB page should usually be created. Crouch, Swale (talk) 18:35, 8 August 2019 (UTC)
In many cases, including this one, there is a dab page. That's not the question. Hatnote links to the most likely alternate destinations in the hatnote still save clicks for users. That's why the template supports links to multiple destinations plus to the dab page for the rest. --В²C ☎ 18:50, 8 August 2019 (UTC)
I would say that is about the maximum length of hatnote where there is no dab. In this case I would just link to the dab, which lists the 1961 series and the photo book (along with The Americans (gang) which gets more pageviews than either of them). Certes (talk) 18:56, 8 August 2019 (UTC)
If you have to pause and think, there are too many links. Hatnotes are supposed to help readers, not to distract them. In the example case, I might link the other TV series (because a reader might have landed on the wrong page by mistake); but per WP:NAMB I would not link the book, and might not even link the DAB page. Narky Blert (talk) 18:59, 8 August 2019 (UTC)
Link to the dab page is essential (we're talking about the primary topic article, the one at the undisambiguated title The Americans). I think all we need is:
This article is about the 2013 TV series. For other uses, see The Americans (disambiguation).
PamD 19:34, 8 August 2019 (UTC)
I wrote too rashly, and you're right of course. A link to the DAB page is essential from a WP:PTOPIC. Narky Blert (talk) 08:54, 9 August 2019 (UTC)

How to disambiguate the Phases in Marvel Cinematic Universe?

Please see and participate in Talk:List_of_Marvel_Cinematic_Universe_films#Naming the phases. AngusWOOF (bark • sniff) 15:21, 14 August 2019 (UTC)

I've replied on the Talk Page. Narky Blert (talk) 19:50, 14 August 2019 (UTC)

Ridicule page

We could use more opinions at Talk:Ridicule#Requested move 14 August 2019. A permalink for the discussion is here. Flyer22 Reborn (talk) 20:31, 16 August 2019 (UTC)

Application of WP:DIFFCAPS to WP:DABNAME

I recently started an RM discussion over at Talk:Union Station thinking the move would be relatively uncontroversial. However, I was shortly met with opposition from every editor joining the discussion. Their argument seems to be that according to WP:DIFFCAPS, Union Station (capital S) and Union station (lowercase s) are two distinct terms and the Union Station page should not carry the "(disambiguation)" tag in the title as it has no primary topic. After looking into it a bit more, it doesn't seem that this rule is applied uniformly across the site. As I've listed at the RM, here are just a handful of pages that would need moving if that argument is applied uniformly.

However, it seems like there are also existing disambiguation pages distinguished by a single letter only, like Freezer Burn.

I think making a capital letter the only difference between an article and a disambiguation page is a terrible idea as it's bound to confuse readers looking for generic terms rather than proper names. But it seems that a good number of people disagree with me here. The way I see it, there are two options:

I'd very much like to see what others' opinions are on this issue. As it stands now, there's no uniformly applied rule here. Eventhorizon51 (talk) 00:49, 15 August 2019 (UTC)

I don't understand why Friendly fire is even a potential primary topic for Friendly Fire. If WP:DIFFCAPS is followed in the way everyone seems to be suggesting, then isn't "Friendly Fire" by definition not "Friendly fire"? The policy considers them different, so how could one act as the primary topic of the other? Eventhorizon51 (talk) 00:47, 16 August 2019 (UTC)
Now you're presenting your argument in sensible terms. The key issue here is this: if someone enters "Friendly Fire", with two capital Fs, how likely is it that they're looking for Friendly fire vs any one of the other uses of Friendly Fire listed at Friendly Fire (disambiguation)? Frankly, I think there is a good argument to be made that Friendly Fire (two cap Fs) has no primary topic. But that's to be decided at a specific RM for that title. If that's how it goes, then indeed Friendly Fire (disambiguation) would be moved to Friendly Fire. But the main point remains: to decide whether a given "Foo Bar" with multiple uses should redirect to "Foo bar" or hold a dab page depends on these primary topic questions. --В²C ☎ 01:10, 16 August 2019 (UTC)
If you think there's a good case to be made for moving Friendly Fire (disambiguation) to Friendly Fire, then surely that same case can be made for all the other potential moves I listed above (in the case of Dry Cleaning (film), Jury Nullification (book) and Mob Mentality (album), the parenthetical qualifiers can be dropped per DIFFCAPS). I would honestly love to see those RM discussions take place if you would be inclined to initiate them. I'm not since I don't want to come off as WP:POINTy. However, I should tell you that I think fixing this issue across the site would take a while and potentially bring opposition as my list above is by no means exhaustive. Eventhorizon51 (talk) 02:54, 16 August 2019 (UTC)
Just because I think there's a good case to be made doesn't mean the community agrees, or even that I think it's important enough to be worth pursuing. To anyone to whom it does matter, each case must be considered individually with an RM proposal. But thanks for compling this list - worthy of consideration in the future. --В²C ☎ 20:09, 16 August 2019 (UTC)
Capitalization differences may give rise to a new primary topic, but they don't always do so. It's similar to pluralization. In some cases...
  1. The plural creates a new primary topic (Window vs. Windows)
  2. The plural leads to a dab page (Raven vs. Ravens)
  3. The plural is a primary redirect to the ptopic for the singular form (Penguin, Penguins)
As with Friendly Fire, one might ask "Why is Penguin even a potential primary topic for Penguins?" given that the latter would never be considered as a title for the former article. As B2C says, it's a question of what will, on average, be most helpful to the reader. If the ptopic of the "base" (singular/uncapitalized) form is a very popular/significant topic, and the article(s) under the variant (plural/capitalized) form are marginal, that makes it very likely that users searching for the variant form will be looking for that base ptopic. Colin M (talk) 15:39, 18 August 2019 (UTC)
@Colin M:"Most helpful to the reader" is not an objective standard. I think we can all agree that different readers have different needs and will sometimes enter things into the search box differently. What you're essentially doing here is creating three separate levels of primariness.
  1. No primary topic at all (ex. Keep Cool)
  2. Primary topic at the "base" form but not under variants (ex. Central station vs Central Station)
  3. Primary topic at the "base" form and under variants (ex. Short Circuit redirects to Short circuit)
There is a fourth case, where capitalization indicates a different primary topic entirely (Iron maiden vs Iron Maiden), but that's not relevant here. I don't know about you, but I think we have enough of a primary topic debate with article titles as it is. This just opens up another nasty can of worms. Not only would we have to argue about if there is a primary topic, but even after we decide there is, we also need to argue about whether it's primary enough for the variant capitalizations. How is that productive? RM is backlogged enough without this vague mess. Implementing this would just cause more problems. Eventhorizon51 (talk) 21:05, 18 August 2019 (UTC)
The criteria at WP:PTOPIC are not objective either. If they were, RMs would be a lot easier. Implementing this would just cause more problems. You say that as if I'm proposing some novel scheme. I'm just describing the way things are already operating. How is that productive? I think allowing some nuance into the discussion, rather than applying a uniform rule across the board, allows us to improve reader experience in some cases. Jury nullification gets around 750 views per day, whereas Jury Nullification (book) gets less than 4. If even 1% of users search for "Jury nullification" (the concept, not the book) using the capitalized form, then making Jury Nullification a WP:PRIMARYREDIRECT is the superior choice for helping readers to get where they want to go. OTOH, the band Iron Maiden gets 5x as many views as Iron maiden, so the mathematics of getting readers where they want to go is vastly different. Absolute rules like "always PREDIRECT capitalization variants" or "never PREDIRECT capitalization variants" are unhelpfully procrustean. Colin M (talk) 22:52, 18 August 2019 (UTC)

People ignoring WP:PRIMARYTOPIC

I created a move discussion for the disambiguation page CBD, since Cannabidiol is unquestionably the primary topic (with 20 times more pageviews per day than the next most popular topic). However, the people involved in the discussion don't seem to care and want to keep CBD as a disambiguation page anyway. What can I do to convince them that this guideline exists for a good reason? Kaldari (talk) 04:24, 29 August 2019 (UTC)

For every person who ignores PRIMARYTOPIC, hundreds obsess excessively from a narrow perspective.
In some parts of the world "CBD" *only* means "central business district", which corresponds to "downtown" in American English, and many of the people in those parts of the world do not obsess, or have routine business with cannabinoids. "CBD" for "central business district" is routinely used as if there is no other meaning eg. This breaks your PRIMARYTOPIC argument.
Pageviews are unreliable data. Sure, look at them, everything with a lot of pageviews needs consideration, but pageviews are very poorly used as a decision making criterion in isolation. Pageviews reflect systematic biases, and may be completely at odds with long term significance. --SmokeyJoe (talk) 05:42, 29 August 2019 (UTC)
Very well said. I'll link to that post from User talk:Andrewa/Primary Topic, statistics and reasonableness, the talk page of a foreshadowed user essay that includes that very point. Andrewa (talk) 21:17, 29 August 2019 (UTC)
Cannabidiol has more views but we should only count the subset which result from searching for the synonym "CBD". That may or may not be far more than search for "CBD" meaning central business district or something else. We can't tell. Certes (talk) 00:18, 30 August 2019 (UTC)
We can't tell with certainty, but considering that CBD averages 232 hits per day while Central business district averages 425 and Cannabidiol averages 8836,[1] it does appear more likely than not that a majority are searching for cannabidiol. Station1 (talk) 08:40, 30 August 2019 (UTC)

A proposed new disambiguation template

This discussion is intended as a continuation from Wikipedia talk:Disambiguation#Guidelines for title-and-name disambiguation pages. That discussion seems to have reached a conclusion, and it seems better to open a new one rather than to start another hare in that older one.

I propose Template:Title and name disambiguation, as a parallel to Template:Human name disambiguation. The template would populate Category:Title and name disambiguation pages, with any other side-effects thought necessary. The most common shortcut would be Template:Tndis, by analogy with Template:Hndis.

A link to the new template would need to be added to Template:Disambiguation#Variant templates.

I can see two main advantages:

  1. The instructions about how to use Category:Title and name disambiguation pages would all be in one place, and where you'd expect to find them. The instructions might be linked to from Wikipedia:Manual of Style/Disambiguation pages#Specific entry types; but having the complete instructions there doesn't feel right: other specialised DAB templates don't.
  2. Many DAB pages with tndis links also contain other types of link. For example, King Charles contains several links to articles which are very properly listed there, but which are not about kings called Charles. It is correctly tagged ((dab|tndis)). Others, however, are pure tndis pages; King Joseph is an example. According to Wikipedia:Categorization#Categorizing pages, "each categorized page should be placed in all of the most specific categories to which it logically belongs. This means that if a page belongs to a subcategory of C (or a subcategory of a subcategory of C, and so on) then it is not normally placed directly into C". King Joseph is currently tagged as ((dab|tndis)), and so goes against that guideline.

As to point #2: I have created at least one tndis page which I put into Category:Title and name disambiguation pages, only to have another editor recategorise it as ((dab|tndis)), on the basis of either WP:DBC or some discussion or other on some loosely-related topic. Apparently Category:Letter-number combination disambiguation pages is acceptable as sole category, but Category:Title and name disambiguation pages is not. The proposed dedicated template would be (a) easier to type and (b) avoid such wastes of time.

As to point #1: there are rules about how to sort articles about titled people at WP:PEER, but they look inappropriate for DAB pages. Why should the DAB page King Charles be sorted as 'King Charles' and the DAB page Prince Charles (disambiguation) as 'Charles, Prince'? (IMO correctly, Prince Charles (disambiguation) is sorted as 'Prince Charles'.) That's inconsistent and counter-intuitive (as well as messing up the sorting in Category:Title and name disambiguation pages).

I don't see a need for any parameters on Template:Title and name disambiguation. Sortkeys can be handled by ((DEFAULTSORT)), which needs more typing and is therefore less likely to be added unnecessarily or wrongly. I suggest that the Usage description on the proposed new template should include:

Title and name disambiguation pages are sorted by the basename, i.e. 'Title Name' not 'Name, Title'. A sortkey should only be added to remove a (disambiguation) qualifier, or diacritics or camelcase, from the base name.

I have no knowledge of template construction, and no inclination to learn. If this idea is supported, someone else will have to implement it, so... Narky Blert (talk) 19:37, 26 July 2019 (UTC)

Buck Peak

What's the best thing to do with a DAB page like Buck Peak? I came across it in List of highest points in Oregon by county, where I redlinked to Buck Peak (Oregon) in the hope that there's only one in the state.

Courtesy ping to Asiaticus, who created the DAB page (and thereby alerted me to the peak in the Oregon article, which may have been linking to the Diablo Range article for a good long time). Narky Blert (talk) 16:15, 1 September 2019 (UTC)

I've added an entry for your redlinked Oregon peak. @Asiaticus: I've been bold and deleted your final entry of "Or one of 23 other summits with that name in the United States.": almost every dab page could have something like that "Or one of the other towns / people / albums etc with that name", and we just don't do it, as I understand WP:MOSDAB etc. PamD 17:01, 1 September 2019 (UTC)
It's rather disappointing that Multnomah County, Oregon#Geography lists its highways but not the highest point in the county! I'd have linked to that from the dab page if it had a mention, but can't do unless it does appear there, so the link is to the list of highest county points. PamD 17:04, 1 September 2019 (UTC)
And I found another to add to Buck Peak. PamD 17:12, 1 September 2019 (UTC)
All mere hillocks compared with Buck Peak (New Mexico), in the Magdalena Mountains. Certes (talk) 17:22, 1 September 2019 (UTC)
But should they be on the dab page as red links (as you've done), when the articles which mention them don't link them? I'd prefer to just list them in black with the blue link. What do other people think? Are there likely to be future articles on all these? PamD 17:26, 1 September 2019 (UTC)
A good question. My understanding of MOS:DABRED is that we should either have an entry of the form Red Peak, part of Blue Range or no entry at all, depending whether the entry passes WP:GEOLAND. I'm not sure whether these peaks qualify. Certes (talk) 19:33, 1 September 2019 (UTC)
IMO redlinks on DAB pages fall into the 'mostly harmless' category, provided that/especially if there's a bluelink pointing to an article which mentions the topic (i.e. WP:DABMENTION). That way, readers can find the best information WP has. Trying to prove lack of notability can be hard work. A redlink which might sit there forever does little harm to the project; and a well-formed redlink suggests to an editor thinking of writing the article (or to an editor thinking of adding a redlink) what its name should be. Unifying a variegated collection of redlinks if you find one and can do it feels worthwhile, but it isn't easy. Narky Blert (talk) 19:58, 1 September 2019 (UTC)

I'd like to combine Buck Peak with Buck Mountain to create a mountain index article titled List of peaks named Buck, with the usual mountain index article formatting and with redlinks for future expansion. Is that OK with other editors? —hike395 (talk) 01:00, 2 September 2019 (UTC)

Would Buck Mountain and Buck Peak redirect to List of peaks named Buck? If they would, how would you to plan to monitor the list for any bad links it might collect (such as the link from List of highest points in Oregon by county to Buck Peak (Diablo Range), California mentioned in my original post)? Narky Blert (talk) 01:38, 2 September 2019 (UTC)
@Narky Blert: Could you help me understand what a bad link to a list article would be? An editor intending to link to a peak that isn't in the list? —hike395 (talk) 03:32, 2 September 2019 (UTC)
@Hike395: A 'bad link' is any link which points a reader to the wrong or an ambiguous place. In the case of a list article, that would be a link to the whole list when the intended target is a specific entry in the list. Unhelpful links like that are extremely difficult to find and to fix. Redlinks are fine! they tell the reader that something exists, but that there is no article yet (and perhaps you would like to write it...).
By way of example, biologists customarily refer to binomial authorities by their surnames alone. At the suggestion of our colleague Certes, a couple of months ago he and I ran a project to turn such links to surname pages (most of which are list pages, with no discussion of the surname as such) into links to the relevant articles. He wrote the searching bots, and we worked together on the fixes. There were about 1,500 such bad links, and the project took about two weeks (probably something like 200 man-hours). The relevant article was often in English Wikipedia; sometimes it was in another Wikipedia; in a few cases, there was no article anywhere, so that the original link to the surname page had been totally and utterly useless.
We chose not to extend the project to football (soccer) players, where the problem may well be at least as bad. Narky Blert (talk) 04:21, 2 September 2019 (UTC)
Thanks for doing that hard work! I can certainly sympathize with spending hours on WP drudgery. I'm currently slowly going through Category:Pages with malformed OS coordinates, because people sometimes enter unusual or incorrect map coordinates into infoboxes.
For cases like the binomial authorities or football player list articles you mentioned, or for broad-concept articles, what tools are available for detecting ambiguous inlinks? Anything other than "What Links Here"? —hike395 (talk) 13:09, 2 September 2019 (UTC)
Good work on fixing bad coordinates! I know how tedious that can be. I've fixed a few, when I was trying to track down an ambiguous obscure village or the like, and found that the coordinates in an article about an allegedly nearby place were wrong (on one occasion, in the middle of the Atlantic Ocean). Narky Blert (talk) 09:08, 3 September 2019 (UTC)
It's awkward, which is why it hadn't already been done. PetScan gave me a list of articles in Category:Surnames, which I pasted in reasonably sized chunks back into PetScan's "Links to Any of these pages" whilst setting "Has any of these templates" to ((Speciesbox)), ((Taxobox)), ((Automatic taxobox)), etc. A few global edits in a text editor produced User:Certes/Taxa linked to surnames. (The current version is the rump of pages we missed with the first pass; earlier versions were much longer.) The Surnames links there call PetScan again to find the name(s) in each article. We repeated the exercise with redirects such as Melvill, and again with titles of the form X (surname) and X (name) both with and without the qualifier. (Yes, someone actually attributed a species to Gray (surname).) Gory details here and here. Certes (talk) 14:54, 2 September 2019 (UTC)

Idea:/ Question: I just thought of a hack that might make everyone happy. What happens if we make a List of peaks named Buck, have Buck Mountain and Buck Peak redirect to that, and then mark the redirects as disambiguation pages? No editor is going to mistakenly link directly to List of peaks named Buck. Essentially no one will notice if the redirects are marked as dabs. But all of the automated "link to dab" tools should work, right? Editors would get notified if they link to Buck Mountain or Buck Peak. Narky, Certes, Shhhnotsoloud, what do you think? —hike395 (talk) 08:34, 3 September 2019 (UTC)

That sounds highly irregular! (Which means I doubt it's ever been tried before.)
Under this scheme, Bear Mountain (disambiguation) and Bear Peak (disambiguation) would be WP:DOUBLEREDIRECTs if they linked to the base names. But, neither redirect would be needed at all! The sole purpose of (disambiguate) redirects is to enable a clean link to a DAB page when the whole page is the target (usually from a hatnote or a see-also on another DAB page). The list article itself would act as the valid link, and have to be deliberately chosen.
This idea is worth thinking about. Narky Blert (talk) 09:39, 3 September 2019 (UTC)
We currently have two redirects marked as ((Disambiguation)): PPK and Kotla. Both were made into redirects in the last week without removing their previous content and look highly irregular (i.e. wrong). Certes (talk) 10:47, 3 September 2019 (UTC)
 Fixed these two obviously bad cases. For the "good" case, I can make a category of disambiguation pages which are "Redirects to SIAs that should not be linked to", if that helps. The notice on such redirects are real -- no one should link to Buck Mountain or Buck Peak (in the hypothetical case). Is there anything un-encyclopedic or un-WP about this? Would love to know before we do anything.
I think this hack can generally solve the multi-year problem of DPL and SIAs, and lead to harmonious editing. —hike395 (talk) 14:51, 3 September 2019 (UTC)
I suggest a new template ((R to SIA or list)) (n.b. ((R to list)) is already occupied) along the lines of ((R ambig)). It should carry a similar must-be-fixed warning, and a new one saying the redirect must be in at least one DAB category, and also populate a new hidden category of the type you suggest. It should be easy to pick up redirects which have the template but haven't been placed in a DAB category. Narky Blert (talk) 19:44, 3 September 2019 (UTC)

If you're interested in this, you might also be interested in this: Wikipedia:Articles for deletion/List of lakes named McArthur. Shhhnotsoloud (talk) 16:47, 2 September 2019 (UTC)

No consensus in primary topic discussions

One change to primary topic discussions I would like to see considered is that when a discussion involving a primary topic results in a no consensus, that should be taken as evidence that there is no primary topic, regardless of what articles were initially under discussion. For example, the move discussion about moving Avatar to Avatar (Hinduism) and Avatar (disambiguation) to Avatar, there was no consensus, which defaulted to no move. Instead, I'd like to see such discussions focus primarily on whether or not there is a primary topic. This way, if there is no consensus, that would mean there is no primary topic at all, and the result would be that the DAB page would displace the article occupying the primary topic. Thus in Avatar case, the pages would have been moved if there had been no consensus that there was a primary topic. I realize this would be a major change to how consensus involving primary topics is decided. However, currently, it takes a clear move consensus to move an article away from a primary topic title, which means an article will stay at the primary topic simply because there's no consensus to move it, when consensus points to there being no primary topic. By focusing the discussion on whether or not there is a clear primary topic instead of whether or not to move, we have a better chance of actually determining if there is a primary topic or not. - BilCat (talk) 07:21, 31 May 2019 (UTC)

I agree that should be the case (at least up to a point) I think the main argument against it is that it breaks external incoming links and means editors have to fix many DABlinks. But yes I think while the burden does generally lie on those wanting to move an article from its long-standing title, it should always the responsibility on those wanting a PT and a no consensus should actually mean there's no consensus to have a primary topic and thus the DAB should be at the base name, see this comment and this. Likewise when there is a consensus that there is no primary topic as here what should be done? In this case there was clear consensus that the place in Scotland wasn't primary for "Lewis" but no clear consensus if "Isle of Lewis" was actually the best title, arguably one could suggest that by "defaulting to the status quo" we should move it to Lewis, Scotland since that would reflect the article still being called "Lewis" but simply having a qualifier however as noted there was some support for the new title anyway. See also the open RM at Talk:Mull#Requested move 27 May 2019 where there is a proposal to put the island at "Isle of Mull" per COMMONNAME and the DAB at "Mull" per NOPRIMARY. Crouch, Swale (talk) 09:43, 31 May 2019 (UTC)
"No consensus" implies that two or more options are well supported, and the best outcome depends on what those options are. If it's "X is the PT: put X at the base name" versus "no PT: put dab at base name" then, although I usually !vote for the dab, the present process is probably correct. If "X is the PT: put X at the base name" and "Y is the PT: put Y at the base name" are both credible options then that indicates strongly that there is no PT and the dab merits the base name. That's not even a compromise solution: neither readers seeking X nor readers seeking Y deserve to be taken to the wrong article, especially when they could easily mistake it for the right article. (What's the population of New York?) Even as a resident of Scotland, I'm surprised to see [Isle of] Mull as PT given its more general use for a headland (Mull of Kintyre etc.) I'm also delighted that I couldn't find any bad incoming links, though I suspect some kind gnome may have fixed them once the debate started. Certes (talk) 10:39, 31 May 2019 (UTC)
In principle, this is probably right... but I think we might run into problems. Imagine if someone just really thinks a particular topic shouldn't be primary. Since "No Consensus" results are fairly common (and potentially easy to force), they could keep requesting RMs over months/years until they get one. Seems weird that, even though their RM lost, they won by getting their desired outcome of NOPRIMARY. -- Netoholic @ 10:47, 31 May 2019 (UTC)
That's not always a bad thing. Two years ago, the article on NY state was titled New York. Several of us went into its RM believing the city was a PT and came out with the dab at the base name, which I think was the best result for the readers. Certes (talk) 11:02, 31 May 2019 (UTC)
Yes, it was a good result, but it was also the one that challenged my whole thinking on Primary Topic. It shouldn't have been so hard to get a good result! We always had consensus that New York State was not the Primary Topic. But it took eleven years and multiple RMs to convert that consensus to a page move. Reread wp:NYRM2016#Discussion by panelists, especially the two who found no consensus to move. There needs to be a better way. Andrewa (talk) 13:13, 31 May 2019 (UTC)
"We always had consensus that New York State was not the Primary Topic. But it took eleven years and multiple RMs to convert that consensus to a page move. " Exactly! - BilCat (talk) 18:50, 31 May 2019 (UTC)
Are there any cases where it benefits readers to have a recently created article at an ambiguous name? Andrewa (talk) 13:13, 31 May 2019 (UTC)
You mean have the new article take over the base name? Only when a new topic becomes primary, such as next week's pop sensation sharing their name with someone who has a stub for playing one cricket match in 1832. Certes (talk) 17:50, 31 May 2019 (UTC)
I mean any time a new article is created at a name that we know to be ambiguous... that is, there is another encyclopedic topic that could be at that article name. I was hoping for an actual example.
But let's take that hypothetical example... suppose Hairy Harry and the Half Headless Horsemen were a new band, named after a children's book on which we already have a stub. They both appear to be notable. I say, we have no idea which is P T now or in the future. So, what do we do? Currently, there's nothing to stop the book being boldly bounced to a disambiguated name, and the new article being created at the base name over the redirect (hopefully with a hatnote). Or if the band was given a disambiguated name but then got lots of traffic, there would be an RM and it would be moved to the base name, and again the book disambiguated. Both of these seem bad outcomes to me. Andrewa (talk) 20:30, 31 May 2019 (UTC)
@Certes: the Mull move was successful, I fixed all but 1 links and they were all for the island. The last is at Battle of Loup Hill which is probably for the island but since it mentions Kintyre its not clear. There was 1 link for the geographical term in 2010 though. Crouch, Swale (talk) 05:51, 4 June 2019 (UTC)
Thank you. I've fixed Battle of Loup Hill as I'm confident it refers to the Isle of Mull. Certes (talk) 09:45, 4 June 2019 (UTC)
They could be clearly marked something like "Title discussion" or "Primary topic discussion", and by separating the issues, we don't have an RM that has to try to do both at the same time, and we might see fewer "no consensus" results. -- Netoholic @ 18:59, 31 May 2019 (UTC)
  • That said, I am also not really on board with changing anything explicitly that would result in a bunch of new arguments over the border between "not moved" and "no consensus to move", both of which used to frequently be called "no consensus to move" before the supplementary RM page added the part about "three outcomes". Dekimasuよ! 23:10, 3 June 2019 (UTC)

The two ways to see "no consensus" regarding Primary Topic RMs

There are two ways to look at a discussion outcome of an RM that proposes moving an article to a disambiguated title and the dab page to the basename when it results in no consensus.

  1. The current way is to presume community consensus previously established that the topic of the article currently at the basename is the primary topic, and that the RM discussion did not change that. Furthermore, just because the tiny fraction of the community participating in that RM did not form a consensus that the topic is primary does not mean that reflects community consensus. Only when there is a clear local consensus based in policy do we overturn previously established community consensus.
  2. The proposed way is to decide that no consensus by the participants in the discussion reflects a community consensus that there is no primary topic.

I think the second/proposed way gives far too much weight to WP:LOCALCONSENSUS over presumed previously-established community consensus. --В²C ☎ 19:42, 31 May 2019 (UTC)

My advice is still, solve this completely by getting rid of P T, but grandfathering old articles as described in User:Andrewa/Primary Topic RfC. Andrewa (talk) 21:27, 31 May 2019 (UTC)

BilCat, did you get a chance to read User:Andrewa/Primary Topic RfC over the weekend as planned? Any comments?

See

https://en.wikipedia.org/w/index.php?title=User_talk:Andrewa/Primary_Topic_RfC&diff=901309146&oldid=899873781

for my latest thoughts. Andrewa (talk) 01:54, 11 June 2019 (UTC)

Where to

There seems some dissatisfaction with P T as it currently is, but as yet no consensus on a way forward. It's not even clear to me whether there is yet consensus that any change is required. (Although I supported and still support the original proposal as a step in the right direction.)

Some feel that we should raise the bar in some way, but keep P T for clearcut cases. My view is of course that we should raise the bar so high that the only P Ts remaining are existing well-established articles (but be prepared to make the occasional exception, as with all rules). I think it's the simplicity of this that most appeals to me. But I also like SmokeyJoe's suggestion if there is no clear PrimaryTopic then there is no PrimaryTopic. [3] Definitely another step in the right direction.

And we have the Yogurt principle and Andrew's Principle which both go outside the square. I admit I was a bit surprised at the support expressed above for the Yogurt Principle, should we perhaps reinstate at least one of the seven deleted or retargeted Project Namespace redirects? [4]

My proposal is now backed by a great deal of case evidence and is looking better and better. But a detailed analysis of this evidence is lacking. When I find time to do that (or someone else does) then we maybe can think about an RfC along those lines. I think there's still a long way to go but that we'll get there. Happy to be proven wrong. Andrewa (talk) 03:37, 6 June 2019 (UTC)

Some editors are dissatisfied with PT as it currently is, and there has always been disagreement over where to "draw the line" (or even if a line should be drawn) between primary topic vs. no primary topic. As I noted above, the way forward for this particular (partial) definition of the line would be at WT:CONSENSUS. -- JHunterJ (talk) 13:18, 6 June 2019 (UTC)
Good point that we would want consensus at wt:consensus to make an exception for article title discussions. RM discussions are already explicitly a special case at wp:consensus.
But both the original proposal here [5] and my one under development completely avoid that complication. Mine in particular would remove the need for the vast majority of P T discussions.
Another point I'd make is that the problem with NYRM2016 (which motivated another editor to start this particular discussion) wasn't so much achieving consensus as assessing consensus. Consensus was assessed as achieved by one of the panel, and in glorious hindsight their assessment stands up very well (just read it). But not the other two panelists, unfortunately... they seem to have been swayed by some patently discardable !votes and arguments. I doubt I will ever understand why. Andrewa (talk) 01:24, 7 June 2019 (UTC)
New York is a closer case than this analysis would suggest. There were reasonable arguments for the state having primacy which would be impossible to make for titles like John Smith or Scale or even Washington. There are some articles, like Apple and Mouse, for which similar arguments could be but probably would not prevail. bd2412 T 02:30, 14 June 2019 (UTC)
reasonable arguments for the state having primacy... Very interesting suggestion. Such as? Andrewa (talk) 03:26, 14 June 2019 (UTC)
Such as the city being a fully contained component of the state, so that everything that happens in New York City technically happens in the State of New York. By definition, the state covers a higher population and has vastly more land area, and has a history that includes the history of the city plus numerous events throughout the state that had little or nothing to do with the city. Also, the proper full name of the city is The City of New York, which, if applied rigorously, would be self-disambiguating from the state. I was one of the strongest proponents of disambiguating the names, after I went through all the links, but frankly if there had not been a substantial number of hits intending specific sports teams, the magazine, and occasionally the song, I wouldn't have been as confident in supporting the move. bd2412 T 03:55, 14 June 2019 (UTC)
That first argument seems to be the HLJC. It has no basis in policy, and failed to gain any support for adoption when proposed. The page currently reads in part This essay emerged from July 2016 discussions of the New York case, as an effort to find compromise over primary-topic status of New York State vs New York City vs absence of primary topic for this name. Numerous counterexampleswere provided, making this criterion impossible to apply without contradicting extant titling practice, WP:PTOPIC and WP:AT policies. This page is retained for historical purposes only and should not be invoked in titling discussions (except to prove its impracticality). So that argument is discardable.
The second is also discardable. The fact that there are other names available for the city does not make the name New York available for the state.
These arguments and many others were proposed in NYRM2016 of course. But no valid ones. There was a basis for the claim that if there was no consensus, then there should be no move. That logic is sound. The only problem is that (apart from that argument) there was consensus, strong consensus in fact, so it doesn't apply. You can't pull yourself up by your own bootstraps.
Or are there others you'd like to propose as valid? Perhaps the argument that Wikipedia will be irreparably harmed by moving the State article away from the base name? (;-> Andrewa (talk) 10:08, 14 June 2019 (UTC)
Don't forget one billion Africans! But seriously, the main arguments against were status quo/no consensus, higher-level jurisdiction, and (to some Americans) NY=state NYC=city. All were reasonable, but were rightly considered less compelling than arguments in the proposal such as no PT. Certes (talk) 10:18, 14 June 2019 (UTC)
Agree that these were the main arguments. I guess you could call them all reasonable, but not sure that this is helpful... in terms of our long-established (by consensus) policies and procedures, both the status quo argument and the HLJC are discardable, and should not have been considered in determining whether there was consensus to move. The status quo argument is logically fallacious, by circularity... It's only relevant after it's established that there is no consensus, so you can't use it in determining whether or not that consensus exists, any more than you can pull yourself up by your own bootstraps. The HLJC has been rejected, again by (strong) consensus.
But that last one is the one that is very interesting, and was discussed in the NYRM discussions, but not nearly so much as it should have been in hindsight (but the signal to noise ratio was daunting). New York means different things to different people. To most people it means the city, and that's both the most significant topic and the one most people want. But is that good enough? I now say, no. That's what we have established at NYRM, and why I eventually decided (to my surprise) that Primary Topic had reached its use-by date, rather than proceeding to NYRM2018 as I originally proposed. Wave means one thing to a surfer or sailor and probably also to any primary school student, another to a physicist and probably to anyone who has progressed to western secondary education or beyond, and still a third thing to the Pope or any royal princess. Macedonia means different things to different groups, and here there is a strong POV in supporting any of them, but there's also a POV in denying any or even all of them primacy. We are saying, in Wikipedia's voice, that the (various) nationalists are all wrong.
So... it's Primary Topic that is busted. Not consensus. Andrewa (talk) 17:45, 14 June 2019 (UTC)

It seems inevitable that this will go to archive with no immediate action.

But I'd like to thank User:BilCat for raising it. It seems to have stirred some discussion at User talk:Andrewa/P T examples and scenarios which is helping me refine my own proposal still further. I'm hoping that someday, some action will come of it all. Andrewa (talk) 14:11, 19 June 2019 (UTC)

IMO its too easy to keep a primary topic, those such as Lewis and Washington should have been moved years ago, primary topics should usually be obvious and clear to the vast majority of people, such as London, Paris and Energy. Crouch, Swale (talk) 20:58, 19 June 2019 (UTC)
primary topics should usually be obvious and clear to the vast majority of people??? Where do you get this idea? The policy pretty much says the exact opposite. See Wikipedia:BUTIDONTKNOWABOUTIT. --В²C ☎ 22:20, 19 June 2019 (UTC)
Exactly. But that's not good. Is it? Andrewa (talk) 00:11, 20 June 2019 (UTC)
I think we have long had rough consensus that there is a problem, and the NYRM fiasco has just highlighted it. Talk:Isle of Lewis is a similar sad story to NYRM, just not quite so newsworthy, and there are sure to be others. Not sure what the point is about Washington.
To me the best example of P T we don't want to touch is Mathematics. Again, there are others. But very few of these are going to be created in the future, because if they're that generally known we already have an article! One every ten years? Or for a topic as clear cut as Mathematics, maybe even one every few centuries? Even if it's one a year, these can and should be dealt with case by case, with commonsense and allowing the occasional exception.
But just how to best raise the bar, and how far? I'd like to do something very simple and low-cost, but a bit radical, and so far there have been many contributions that have led me to refine the proposal (including three complete rewrites from the ground up, so far) but none that have seriously challenged its basic premises. And so rather than do either you'll do something else that neither likes at all - Enry Iggins. Andrewa (talk) 00:17, 20 June 2019 (UTC)

Here's another thought... is Primary Topic even consistent with article title policy? Article titles should be recognizable...... Isn't an unambiguous title intrinsically more recognizable than an ambiguous one?

This has been addressed in past discussions I think, and consensus was that ambiguous titles are OK (I'm looking at archives to see where, help with that would be appreciated). So essentially I'm wanting to revisit that in line with more recent experience. Andrewa (talk) 23:57, 21 June 2019 (UTC)

"Florida" is technically ambiguous, because there are other things called that besides the state, but "Florida (state)" would not be much of an improvement, because it could be confused with Florida State. Of course, we could call it Florida (U.S. state), but that seems excessive. bd2412 T 01:11, 22 June 2019 (UTC)
Yes, Florida (disambiguation)#Places alone has eleven subsections each with multiple entries, so Florida is highly ambiguous... as one would expect, it's Spanish for flowery so it's similar to George Street in the formerly British colonies only worse, in both cases there are lots and lots of others for which we don't yet have articles and may never have.
I wouldn't touch Florida. Moving it is perhaps still more damaging than leaving it where it is, and just pity the mobile or slow connection users who load that large page in error. The generation after next may do something more logical (;-> but I can't think of what, so for now grandfather it and just fix what we can.
We should definitely have a redirect from the unambiguous Florida (U.S. state), and do, it was created about ten years ago. It would be interesting to know exactly why, but the creator is inactive. Andrewa (talk) 02:14, 22 June 2019 (UTC)

It's an important point, so I'm going to spell out the explanations I offered to B2C (diffs above).

They described their way of working, and they're right, if we were to do it my way, and if other users follow B2C's way of working (and I'm sure some do), and if they agree with us on the primary topic, they'll load a DAB page, and will then need to identify the article they want on it, and click on that link. And they will then be taken there, reliably and naturally. That's acceptable IMO.

The problem is, as we now do things, if other users don't agree with us on the primary topic, these others will be taken to the wrong article. This may be much longer than a DAB, and pose a problem for slow connections and mobile devices. But worse still, this user expects to be take to the topic they're looking for, because that's what they think the primary topic is. So when they don't find any content there that matches their use of the term, it's logical for them to conclude that Wikipedia doesn't cover the topic at all, and give up. Both of these outcomes are avoidable and unacceptable IMO.

Those are not the only harm of the way we now do things. Just two examples.

And of course B2C's is not the only way of working. And new ones will continually arise, as the web etc changes. That's another reason to be as logical as we can be in deciding article titles. And deliberately choosing ambiguous article titles is not in any way logical. It is contrary to the spirit of WP:AT, and can be expected to cause problems. And does.

Primary topic served us well in the early days. But all the evidence indicates that it is now well past its use-by date. Andrewa (talk) 01:40, 23 June 2019 (UTC)

It gets worse than that. When New York led to the state article, a reader might assume that its information referred to NYC, and vice versa if New York were a primary redirect to the city. Certes (talk) 09:42, 23 June 2019 (UTC)
Good point. And it's a no win scenario. If (as previously) the destination of New York is the State, then the demographics in the infobox, for example, might well be logically assumed to apply to the city by someone who regarded the city as the Primary Topic. But it cuts both ways. If we were now to make the city the destination of New York (as current policy would have us do IMO, and NYRM2018 was to test this, as well as debunking that Condorcet nonsense as a byproduct), then someone who regarded the State as Primary Topic (and the many RMs showed that there are many such people too) would (wrongly but logically) assume that they referred to the State. It doesn't matter who is right, having an ambiguous name point to an article rather than a DAB is a bad idea in any case. Andrewa (talk) 12:05, 23 June 2019 (UTC)
Oh, poppy cock, Andrewa. First of all, it’s not merely some users that are put at an inconvenience when a dab page is at a term despite that term having a primary topic, it’s the vast majority of those who search with that term with WP search/GO. Users using Google are taken directly to their article regardless of what the title is (by selecting their desired article from the search previews). Users with slow connections and loading issues can quickly learn to use Google or WP Search (to see article previews from which to select their desired destination) rather than GO. And, we’re not talking about cases like “New York”, for which there is no reason to believe that the significant majority searching with that term is seeking either topic. And even if someone is taken to the wrong article in any such case if they can’t immediately tell from the hatnote, intro text and captions that they’re on the wrong article, they’re not going to get much from any article. An ambiguous term with an undisputed primary topic is a better example, like “Paris”. Consider someone searching for the Texas town with that term. They can use Google or WP search and find it in the search results and click on it, which would take them directly to the article they seek, or they can enter “Paris” in the WP search box and hit GO, which will take them to the French city. Oh, the horror! And you think the tiny number looking for a relatively obscure use of a topic with that term and GO won’t realize they’re on the primary topic article and figure out to use the hatnote or re-search? This is the “problem” you’re trying to address? Primary topic has not outlived its usefulness. It provides the exact same utility today as it did in the early days. It generally improves the experience of those using WP search/GO. —В²C ☎ 15:31, 23 June 2019 (UTC)
Andrewa is quite right, and this obsession with the Go box is seems to be causing you a blindness to what makes a good title for an article. Google does the Go box concept way better, Wikipedia should get rid of it. —SmokeyJoe (talk) 03:07, 24 June 2019 (UTC)
Perhaps just redirect to a search engine such as Google? Allow users to select their search engine in their user profile, and prompt them to do so the first time they search. But the simple, fast search we do have works really well for articles at unambiguous names. Andrewa (talk) 11:12, 24 June 2019 (UTC)
The Go function originated when the Wikipedia internal search engine was really crap. The internal search engine used to return pages seemingly carrying the most occurrences of the words search for regardless of where or context. The Go function was just go to the title entered, with auto complete, and then with options 1 or 2 character misspellings accommodated. It really is a high school sophistication level, and it works really well if you know what it is doing, and everyone knows and agrees that titles list the most important words in sequence order. This often works, but can also lead to very tortured title phrases, and in response to that Wikipedia has moved in the direction of extreme parenthetical titling. The worst thing about the Go box is that it hides access to the actual internal search engine, although hidden in plain sight, as all you have to do is hit the go button with no characters entered to get the search page. --SmokeyJoe (talk) 03:32, 25 June 2019 (UTC)
I use the search box many times on a typical day, and it almost always gets me to what I want which is why it is first stop. Second stop is the search page, which probably should be easier to get to. Third stop is either Google or DuckDuckGo depending on how political I feel (I like DDG in theory, but Google delivers more often). The external search engines with site:en.wikipedia.org do sometimes deliver what our internal searches have not, so maybe there's room for improvement. I'm skeptical, it's not too bad now, and we're primarily about content, so I say let Google do what they do best.
But the more of our articles are at unambiguous names, the better the search box will work. Andrewa (talk) 03:58, 25 June 2019 (UTC)
Sticks and stones, В²C. You don't seem to have addressed the issues. But you raise a good point about hatnotes. Where they exist (and in many cases they don't) they help recover from the problem of being taken to the wrong article, which can happen for many reasons, most if not all of them caused by an ambiguous article name. But it would be even better if readers weren't taken to the wrong article in the first place. And getting to the wrong article may itself give the impression that we don't have the right article at all, particularly if the hatnote is just the otheruses template. Andrewa (talk) 11:12, 24 June 2019 (UTC)
If I didn't address something, it was inadvertent. What did I miss? Anyway, a dab page is never the right article, by definition; it's always the wrong page. The idea of primary topic is that sending a majority to the sought article and only a minority to a wrong page is preferable to sending everyone to a wrong page. And that's just as true and valid today as it was in the early days. --В²C ☎ 18:18, 24 June 2019 (UTC)
There are degrees of wrongness. Sending a reader to the wrong article is worse than sending them to a dab page, just as giving someone wrong directions is worse than admitting I don't know the way. Suppose it's three times as bad, to pluck a number out of thin air. Then we break even if two out of three readers want the PT: putting the PT at the base name would inconvenience one reader three times as badly as a dab would inconvenience all three. Maybe that number is two or ten but there's a threshold somewhere. Certes (talk) 18:36, 24 June 2019 (UTC)
Okay, there are degrees of wrongness. Fair enough. So let's consider someone searching for the town in Texas called Paris. They search with "Paris" and are sent to Paris, the article about the city in France. How wrong is that, really? Assessing the level of wrongness about this stuff is highly subjective, but I think the wrongness of being taken to a dab page is comparable to the wrongness of being taken to an article about another subject with the same name. If the result of a given search term was rated on a scale of 1 to 10 where 10 is landing on the article being sought, 1 is plainly wrong (article topic doesn't even have the name, e.g. searching with "cat" takes you to Abraham Lincoln), I would rate being taken to a different article with that name, or to a dab page, as both being about 5. But even if you're more tolerant with dab page landings, and give them say a 7, you still have to do the calculations based on likelihoods. If 80% searching with a given term are looking for article A, 10% for B and 10% for C, if we take them to a dab page we get an average of 7. If we take them to A, then 80% are 10s and 20% are 5s... that's an average of 9. Much better overall. --В²C ☎ 19:29, 24 June 2019 (UTC)
Nobody is proposing moving Paris. Andrewa (talk) 20:57, 24 June 2019 (UTC)
That was just an illustrative example and besides the A,B,C point I just made, but I also don't see why the reasoning presented above wouldn't apply to Paris. --В²C ☎ 21:50, 24 June 2019 (UTC)
You deliberately gave an illustrative example that is not an example of what is under discussion? Why?
The A, B, C point seems to be attempting an objective criterion for deciding P T, as has been attempted unsuccessfully and repeatedly both by yourself and others. Andrewa (talk) 02:59, 25 June 2019 (UTC)
A DAB is not a wrong page in anything like the same sense that the wrong article is the wrong page. A DAB is an excellent navigation aid. It's fast, reliable and convenient, and should take the reader immediately to the right article, with one mouse click.
If technology ever changes this (and maybe it has, I'm not a mobile user and may never be as my home is in a satellite-only area of the Australian outback) then we need to revamp them, maybe for example give mobile users a different view of DAB pages (not hard to do). So your input as a mobile user is valued.
But I'm not convinced that it has yet changed or ever will, and if and when it does, having articles at unambiguous names is likely to help. It is logical, it is commonsense, and it is the spirit (but not yet the letter) of our existing policy. Andrewa (talk) 20:57, 24 June 2019 (UTC)
Like I said, it's highly subjective. But when I search with "the americans" I find landing on the dab page that is currently at The Americans to be just as wrong as landing on The Americans (1961 TV series) would be. --В²C ☎ 21:50, 24 June 2019 (UTC)
I'm not sure exactly what you think is highly subjective. I think Primary Topic is sufficiently subjective that it is reaching and in some cases exceeding its limits of usefulness. That New York (as an actual example) means different things to different people was the reason that we had the wrong article at that name for many years, and as a direct and predictable result thousands of mislinkings were created. That's the example that started this whole discussion.
I agree that Mathematics should stay where it is, but part of that is I can't think of an acceptable disambiguator anyway. Long term we may want to refine the criteria for articles that are genuine Primary Topics. Not for the fainthearted. There's also the downside of moving well-established articles. So for now, just grandfather them all.
But what we can and should do is to stop regularly putting new articles at ambiguous names. Immediately. And similarly with RMs.
Agree that if it's unacceptable to load a DAB, then we shouldn't have a DAB at the base name. In fact, we shouldn't have DABs at all. But I think most of us find them useful. (;->
How about a quick straw poll? B2C was searching for The Americans (2013 TV series) and got the DAB at The Americans instead (or try The Americans (disambiguation) in case B2C's RM has been done). For me that would have been a win. For him that seems to have been a loss. Other opinions? Would getting to the DAB be a win for you, or a loss? Andrewa (talk) 03:40, 25 June 2019 (UTC)
What’s highly subjective is “assessing the level of wrongness about [landing on a DAB vs. landing on a different use of the search term].” You still haven’t answered why Paris should not be disambiguated per your reasoning, nor why the utility of primary topic in the early days no longer applies today. What has changed that’s relevant to this? -—В²C ☎ 06:42, 25 June 2019 (UTC)
One good question at least! At least two things have changed. We now have experience that should lead us to question the concept of P T. We didn't have that case history before. And, we now have articles on the vast majority of articles (such as Paris) that will ever have the strongest claims to being P T. Not all of them, but enough of them to cover the rest with commonsense and the occasional exception.
But I didn't realise that you were asking for my justification for not moving Paris. Have you read the proposal? Did you think that I was advocating moving Paris? I explicitly say that I am not wishing to move articles such as mathematics and London, and give reasons. Must I repeat them here?
And agree that assessing the level of wrongness about [landing on a DAB vs. landing on a different use of the search term] is highly subjective. But I think that your experience of finding it so inconvenient (ie highly wrong) to land at the DAB at The Americans may well be unusual, and finding that out is the reason I asked for a straw poll. Andrewa (talk) 13:38, 25 June 2019 (UTC)
Interesting. I think we agree that it's not reasonable for "the americans" to take readers directly to The Americans (2013 TV series). Dab The Americans has an easily discovered link to the 2013 series. As it happens, so does The Americans (1961 TV series), but that hatnote is really a mistake per WP:NAMB because "The Americans (1961 TV series)" can't possibly mean the 2013 series. In the general case, there are too many alternative meanings for a hatnote and a good number for a dab. Certes (talk) 08:19, 25 June 2019 (UTC)
I'm going to risk taking that as a win. You would find it OK to arrive at the DAB, as would I. Any other opinions?
And I certainly agree that this is interesting. It helps me and I hope others to understand why some users might want to keep Primary Topic as it is, despite (as I see it anyway) all evidence indicating that it's causing problems, and these problems being so easily solved. But that doesn't mean I can (yet) produce a consensus for solving them.
The only reasonable destination for The Americans is the DAB. Same applies to New York IMO.
But our current policies probably put New York City at the base name, and possibly The Americans (2013 TV series) as well (the RM on that question has arguments by highly experienced users going both ways). Are these occasional exceptions? Or are they part of a pattern (as I believe)? Andrewa (talk) 15:12, 25 June 2019 (UTC)
Andrewa, to be clear, I wouldn't say landing on the DAB is "highly wrong" (I put it at a 5 on a scale of 1-10). I'm just saying it's about as wrong as landing on the article about some other use of the term with which I'm searching (which for me is also about a 5). Speaking of 1-5 and the A,B,C thing, I think you misunderstood the point there. I was not proposing anyone actually using such a system for determining primary topic. It's a thought experiment to help get clarity in this discussion about where we are on the relative wrongness of landing on DAB pages vs landing on articles that match the term with which we're searching, but are not about the topic we're seeking. It's also about getting a better feel for the role likelihood of being sought plays in all this. I'll review your reasons for not moving London, etc. I'll just point out whatever those reasons are, they have not been presented in this discussion. That is, the reasoning you and others have presented here all suggests London, Paris etc. should be disambiguated. So if that's really not what you want, you might consider presenting your reasoning accordingly whenever you're talking about changing primary topic. --В²C ☎ 17:16, 25 June 2019 (UTC)
about as wrong as landing on the article about some other use of the term with which I'm searching... Disagree. The DAB is short and contains a link to the article I want. The wrong article may be long and may not contain any such links. No comparison IMO. Andrewa (talk) 18:38, 25 June 2019 (UTC)
Yes, it's quite clear we disagreed on this point, and I tried account for it by assigning 7 to landing on a dab page, and still 5 to landing on the "wrong" article, above. Maybe you think that be more like 8 and 2? I don't know. I can't speak for you. Just saying we can use this model. You plug in the your preferences for the wrongness values and we go from there. --В²C ☎ 22:11, 25 June 2019 (UTC)
I don't think we can come up with any values that will make this model useful. Getting the right article would be a 10 of course. But getting the wrong one could be anything from a 0 to a 7 or even 8, zero for if they give up, seven or eight for a short article with a clear hatnote to the right article, and all values in between are possible depending on both the article and the user. Similarly, getting to a DAB could be anything from a 0 to a 9.8, the 9.8 if the DAB is short and has the right article clearly identified near the top (such as your experience with The Americans), the 0 if it's not identifiable or not there at all (it happens). Andrewa (talk) 07:11, 26 June 2019 (UTC)
Andrewa, okay, so the main if not only reason you wouldn't favor disambiguating London, Paris, etc. is because it would be disruptive in terms of breaking links, etc. But, if it wasn't for that, you acknowledge your reasoning does suggest all topics with ambiguous names, no matter how well known, how relatively likely to be sought, or how historically significant, should be at a disambiguated title. Only if they happen to be already established at an ambiguous title they should stay? Is that right? --В²C ☎ 17:34, 25 June 2019 (UTC)
No. Andrewa (talk) 18:32, 25 June 2019 (UTC)
@Andrewa: re: "the dab is short". Really? Meet John Smith. Or even Victoria. bd2412 T 04:26, 2 September 2019 (UTC)
See #Relative length of DABs below. Andrewa (talk) 10:12, 3 September 2019 (UTC)

Relative length of DABs

BD2412 (who is rightly renowned for hard work and diligent research) said re: "the dab is short". Really? Meet John Smith. Or even Victoria. [7]

I was taken aback. Here's my subsequent research.

Sorry about the URLs, I know it's a quick and dirty!

All the DABs I looked at, even the John Smith one to which BD2412 had previously referred us twice above, are shorter than all of the articles. As I expected.

John Smith is an outrider as is to be expected, given its unique status as an extremely common name, even to the point of being used in English culture as the prototypical common name. It's almost half the length of the article on the explorer. But I'm guessing that's not typical.

Victoria seems more typical, but even that's long for a DAB. New York is likely even more typical IMO.

And AFAIK that is without taking images into account... only the source bytes of the file name are counted. In practice, the articles will be significantly longer again for the purposes of this discussion when the sizes of the images are considered. I guess other things are missing as well... transclusions for example. But I think you get the picture.

Am I missing something? Andrewa (talk) 10:26, 3 September 2019 (UTC)

And this seems to support my view too. Andrewa (talk) 20:45, 3 September 2019 (UTC)

Why DAB length matters

#Relative length of DABs above may explain something that has long puzzled me.

I have always assumed that landing on a DAB instead of the sought article was a minor inconvenience at worst. A DAB is, or should be, an efficient way of finding the sought article.

But others and notably Born2cycle have taken the opposing view. [8] [9]

In particular, they contend that landing on what they call the wrong page (which includes DABs in their thinking) is just as bad as landing on a wrong article (which to my way of thinking does not include DABs). This has led to some for me frustrating discussions at User talk:Andrewa/wrong article and related user and user talk pages where I have tried and failed to clarify the difference that I see.

If they are correct then much of my thinking needs to change, and perhaps much of Wikipedia's current treatment of ambiguous page names too. And if BD2412 is correct that DABs are not generally speaking shorter than articles they may well be correct. But I cannot see any chance of this view being correct. I would appreciate other comments. Andrewa (talk) 18:41, 3 September 2019 (UTC)

If with my search term I'm searching for a particular article and I'm taken anywhere other than that article, then that's not the article I'm looking for. If it's a DAB we can quibble about it being the "wrong article", since it's technically not an article (though it is a page in article space), but it most definitively is not "the right page" and it is therefore the "wrong page". I also don't see it being better or worse if that "wrong page" is a DAB page with a link to the article I seek, or another article with a hatnote link to the article I seek. --В²C ☎ 18:47, 3 September 2019 (UTC)
I personally really hate long disambiguation pages. When I'm navigating as a reader, I find they eat serious time in looking for the correct page, particularly as it often happens that the usage I want is near the bottom somewhere, buried underneath a shedload of placenames across the various US states, or lists of people etc. As such, I actually myself think (and I may be in a minority of one on this) that even our convention of redirecting partial dabs to the more general dab is wrong. We should have a separate dab page at the partial dab, which lists the same things as are found pertaining to that topic in the main dab. It causes a maintainability hit, but for the considerable benefit of giving users interested in one of the 2-3 topics in question a short dab page to look at.  — Amakuru (talk) 19:00, 3 September 2019 (UTC)
Agree. But I think that's a different issue. At the risk of repeating myself, a DAB should be an efficient way of finding the sought article. And I think they can be. Yes, splitting the long ones might be a good move. Andrewa (talk) 19:15, 3 September 2019 (UTC)
Excluding redirects, pages beginning with "A" have an average size of 6325 bytes. Disambiguation pages beginning with "A" average 659 bytes. Certes (talk) 19:03, 3 September 2019 (UTC)
That seems to back up my quick and dirty figures above.
Are these figures wikitext only, or the whole page load? It's the page load that is more significant IMO, although I now concede that both do matter. Andrewa (talk) 19:33, 3 September 2019 (UTC)
Thanks for joining in. I have no wish to quibble. Use whatever terminology you wish, and I'll try to follow it.
The point is just that landing on a DAB isn't nearly as bad as landing on an article that isn't the sought article (of course I'd like to simply call this the wrong article but that meaning seems to be disputed so let us just avoid the term).
That's even part of the reason we have DABs!
There are two main reasons that it's better to land on a DAB. One is that the DAB is likely to be shorter, and that's what we're discussing here. The other is of course that a DAB is specifically designed for this purpose. The article isn't. (And if some DABs aren't very well designed, as suggested above, then they should be fixed.)
Yes, articles and DABs are both in the same namespace. IMO that's completely irrelevant. Andrewa (talk) 22:34, 3 September 2019 (UTC)
Same article namespace is only relevant to the semantic issue of whether it's reasonable to say landing on a DAB is landing on the wrong article. Whatever you call it, it's not the article the user is looking for; in that sense it's wrong. I don't agree it matters that the DAB is likely to be shorter. Unless you're on a 1990s dialup modem, the load time difference is negligible. There is no need to fully load the article. You only need to see the intro to know that it's not the article you seek, and the hatnote on the article is also at the top and loaded and rendered first, so what does article length matter? It's not like you need to scroll down and fully load the article page. That said, I do agree that landing on long DABs is even more problematic than landing on short DABs, but not because of load time. As noted above, it's because of the pain of finding the article you're seeking is usually more challenging the longer the DAB. Hatnote links at the top of articles are much more efficient and far preferable to wading through an annoying DAB page, IMHO. --В²C ☎ 22:58, 3 September 2019 (UTC)
Hatnote links at the top of articles are much more efficient and far preferable to wading through an annoying DAB page, IMHO.. I completely disagree. Hatnotes are a cost to every reader. —SmokeyJoe (talk) 23:31, 3 September 2019 (UTC)
the load time difference is negligible wrong -- it is quite noticeable on phone or tablet even with good connection. And I very much agree with SmokeyJoe that a dab page is much better than hatnote in most cases other than completely obvious primary topics. older ≠ wiser 23:37, 3 September 2019 (UTC)
Hatnotes are a cost to every reader for who the page is the right page. Hatnotes are a benefit for many readers for whom the page is the wrong page, and that is usually due to the ambiguous title. If titles were PRECISE and RECOGNIZABLE, there would be rarely a need for hatnotes and all readers would be better served.
DAB pages are generally pretty nicely structured. The most likely topics are usually listed at the top. They are usually structured logically, with a view to both logic and likelihood. And if disambiguation pages were to all be titled with the "(disambiguation)" suffix, very few would be landing on them unexpectedly. --SmokeyJoe (talk) 00:28, 4 September 2019 (UTC)
The "cost" of the hatnote at the top is zilch. Here's what's at the top of Paris, for example:
This page is about the capital of France. For other uses, see Paris (disambiguation).
Oh, the horror! What a cost!! --В²C ☎ 00:54, 4 September 2019 (UTC)
В²C completely ignores the cost of a hatnote. If hatnote cost were zilch, why are there output formats that strip them? Wikipedia:Article size is a good starting point for reading about reader considerations.
The Paris hatnote is 14 words. At a typical reading speed for unexpected words of 3 words per second, that's ~5 seconds to read. At 5750 page views per day, assuming a pageview is a reader who reads at least the hatnote, that's 20 hours of wasted reader time per day. --SmokeyJoe (talk) 05:47, 4 September 2019 (UTC)
I’m ignoring the “cost” of reading a oneline hatnote, and you’re ignoring the cost of reading a much larger DAB page, like 14,000 bytes of Victoria. --В²C ☎ 07:10, 4 September 2019 (UTC)
I have never ignored that, and I have no idea how you can feel justified to say that. Of course, downloading, and reading to the point of satisfactions or realization that you didn't want it, is a cost, and is a much higher cost than 5 seconds reading time for a one line hatnote. However, few people go to DAB pages who didn't want to, and that number would be far smaller if the DAB page were PRECISEly titled using "(disambiguation)". I would put Victoria at Victoria (disambiguation), WP:MALPLACED be buggered.
The hatnote cost is relatively small, but it is forced on a very large number of readers. The DAB page, at least it is helping a confused reader or rescuing a reader from bad navigation aids. The hatnote is clutter, in prime real estate, for every reader who has exactly the page they wanted, and it is all negative. Interesting notes about similar topics will be in the text body, wikilinked, but the hatnote is slapped on top as if it is part of a hastily cobbled boilerplate page design in anticipation of mistaken arrivals.
The use of PRECISE titles would reduce reader costs all round. The overuse of PRIMARYTOPIC creates small hatnote costs for the majority, a massive pain for screenreaders, and difficulties for readers who assume that another topic cannot be a the PRIMARYTOPIC for the ambiguous basename. --SmokeyJoe (talk) 08:15, 4 September 2019 (UTC)
The simple fact is (as I've said a few times now) that having ambiguous article titles is asking for trouble. And those are all ways in which we cause trouble for our readers when we do. And there will be others, see below. Andrewa (talk) 11:46, 4 September 2019 (UTC)
  • few people go to DAB pages who didn’t want to. Nobody wants to go to a DAB page. People want to go to an article. Sometimes DABs are unavoidable, but they’re never the intended destination. It’s trivial to ignore a one line hatnote at the top. It’s like the top margin crap of any document (perhaps repeated title on every page and page number); it’s there if you need it, but easily ignored if not. A DAB page is impossible to ignore. Precise titles despite primary topics would send far more to the unwanted DABs. -В²C ☎ 14:00, 4 September 2019 (UTC)
    • Nobody wants to go to a DAB page. Disagree. But some people, such as yourself, do evidently get very annoyed. In your case with The Americans, you almost threw your tablet across the room, and even after several occurrences of landing on the DAB you didn't bother to bookmark the page you wanted to avoid further annoyance. And I guess we need to assume others will do the same.
    • For myself, I regard landing on a short DAB which has a clear, unambiguous link to the page I want as a win. Am I alone in this?
    • On the other hand, I'd regard loading a long unwanted article as a lose.
      • If there was a hatnote to the article I wanted, it's still an unnecessary diversion. I love to learn so it's not an annoyance, but to some I'm sure it would be.
      • If the hatnote pointed to the DAB, I've gained nothing (other than perhaps a pleasant diversion and distraction!) compared to getting straight to the DAB, but that one extra mouse click doesn't worry me. It takes no significant time and requires only very basic skills.
      • If there's no hatnote at all, I think I'd be justified in concluding that Wikipedia had no article covering what I wanted. I wouldn't personally draw that conclusion, because I know how slack our navigation aids are, but I fear others might! And there is no guarantee that there will be a hatnote. Perhaps there should be.
    • Precise titles despite primary topics would send far more to the unwanted DABs. Very common assumption and good question. Would they really? How? I think the opposite! Precise or unambiguous itles should make articles easier to find. And if they don't, there's something wrong with the search techniques, not the titles. Andrewa (talk) 21:08, 4 September 2019 (UTC)
      • The search technique relevant here - the one WP:D attempts to improve - is entering a search term in the "Search Wikipedia" box and clicking on Go. Using precise titles for primary topic articles would send such searchers to an unwanted DAB page. Someone entering "paris", for example, would end up at the DAB page instead of at what is currently at Paris. (yes I know in your proposal you wouldn't move Paris but we're talking in generalities here, not about your proposal). --В²C ☎ 21:32, 4 September 2019 (UTC)
You’re assuming the DAB page is ambiguously titled. Which it is, but need not be. —SmokeyJoe (talk) 22:52, 4 September 2019 (UTC)
I'm warming to the idea of disambiguating all DABs with (disambiguation). It might even be worth adding to any eventual RfC. The two proposals do each enhance the other to some degree, so they're not entirely independent. Fix it all at once.
Discuss at User talk:Andrewa/Primary Topic RfC#DAB titles. Andrewa (talk) 00:30, 5 September 2019 (UTC)
DAB pages are generally pretty nicely structured. Agree. The only problem raised yet (not by me) is that some of them are too long. That can and should be fixed if it's a problem. Andrewa (talk) 05:40, 4 September 2019 (UTC)
Victoria is long. Is it too long? Would it be better split into multiple pages? Would you reduce the Victoria page to essentially the current Table of Contents, and make each level 2 heading section a subpage? Is there a rule for or against this? --SmokeyJoe (talk) 05:50, 4 September 2019 (UTC)
I find it fine. Wikipedia talk:Article size is fairly active, that might be a good place to ask. Andrewa (talk) 06:15, 4 September 2019 (UTC)
Same article namespace is only relevant to the semantic issue of whether it's reasonable to say landing on a DAB is landing on the wrong article. Agree. You raised the fact that they're in the same namespace, and it was a waste of your time and ours to do so. It's irrelevant.
it's not the article the user is looking for; in that sense it's wrong... Well, I said I'd adopt your terminology, so I guess you can call it that if you like. But similarly, is every DAB always wrong however it is accessed? And is the whole index of every paper encyclopedia wrong in that same sense? (In the one I most used as a boy, the index was an entire full-size volume.) In none of these cases are they the article the user is looking for, but just like a DAB, the index is there to help the user to find the article they want. Is it really wrong for them to use it to find the article they want, as you use the word wrong? I'm fascinated.
And I suppose Google Web Search is wrong too, as the page it generates is not the page the user is looking for. We must tell them that. I'm sure they'll be very worried. (;->
But a DAB is not as wrong as the wrong article. Is that an acceptable compromise of terminology? Andrewa (talk) 05:40, 4 September 2019 (UTC)
A dab is not as wrong as the wrong article. That's worth repeating. Until 2017, I could look up the population of New York and find it to be 19 million. Wow, that city is more crowded than I thought. Then I use that factoid and make myself and Wikipedia look foolish, or worse. Now that title leads to the dab rather than the state article and I can pick the area I meant from the handy list of common meanings right at the top. I may even realise that the question was ambiguous and check whether I should be researching the state or the city before I make similar mistakes. Certes (talk) 08:10, 4 September 2019 (UTC)
The simple fact is (as I've said a few times now) that having ambiguous article titles is asking for trouble, and that's another good example of the sort of trouble we cause our readers when we do it. But there will be other problems that we don't guess until we notice them happening. Unambiguous (or PRECISE if you like) article titles are quite simply more logical, and will avoid problems that we cannot imagine and don't need to. Andrewa (talk) 11:46, 4 September 2019 (UTC)
"Having ambiguous article titles when there is no primary topic is asking for trouble." There; I fixed it for you. The problem with having the state at New York is not that "New York" is ambiguous, it's that it's ambiguous and the state is not the primary topic. If "having ambiguous articles titles" for primary topics was truly "asking for trouble", then having the respective primary topics at Paris, Albert Einstein and even the recently-moved The Americans would be asking for trouble. But it's not. Going back to the point I was just making with SmokeyJoe above, had we left the DAB at The Americans then every time anyone searched on WP for the epic 2013 TV series with "the americans", they would get taken to the DAB instead of to the article they sought. With the current configuration, with the 2013 series at the base name, the vast majority of "the americans" searchers for whom the 2013 series is the sought topic has to "endure" the one line "for other uses..." hatnote at the top, but isn't that better than having them needlessly load the DAB page, scan it for the topic they seek, click on that, and then finally present them with the article they seek? And yes, the tiny minority looking for one of the other (relatively unlikely to be sought) topics associated with the "the americans" search term does have to go through that, but it's a tiny minority. Isn't that our goal, though? To get as many WP searchers as reasonably possible to the article they seek as quickly as reasonably possible? Isn't that why we have primary topics? --В²C ☎ 16:21, 4 September 2019 (UTC)
We have Primary Topics because of the foolish MALPLACED practice, and because Larry Sanger thought it a good idea in 2001 for titles to be what people would wikilink mid-sentence. —SmokeyJoe (talk) 22:55, 4 September 2019 (UTC)
And Larry was right. But we've failed to take subsequent experience into account in our implementation of the principle. If the destination of the ambiguous term is a DAB, then the editor gets a warning, and wikignomes can easily find and fix the mislinking if they don't, using the pipe trick of course. But if the mislinking is to an ambiguously titled article, nobody gets any warning at all, and as a result we just publish rubbish in ignorant good faith.
But now inclined to agree about wp:MALPLACED. Andrewa (talk) 07:51, 5 September 2019 (UTC)
There, I fixed it for you. Words fail me. No. Give us a break.
The problem with having the state at New York is not that "New York" is ambiguous, it's that it's ambiguous and the state is not the primary topic. That was the initial problem, yes, many years ago, but it developed into Wikipediagate. In this matter at least we are a laughingstock. People are entitled to ask, if NYRM2016 is the best an impeccable panel containing a senior functionary can do, what is the rest of Wikipedia worth? I still believe in Wikipedia, but it still raises serious questions for me. Not for you?
If "having ambiguous articles titles" for primary topicswas truly "asking for trouble", then having the respective primary topics at Paris, Albert Einstein and even the recently-moved The Americans would be asking for trouble. But it's not. Moving Paris or Albert Einstein would create more problems than it would solve IMO, and there is no perfect and magic solution. The jury is out on The Americans IMO. If we do move it back someday, your successful RM will then have caused a great deal of needless trouble.
isn't that better than having them needlessly load the DAB page, scan it for the topic they seek, click on that, and then finally present them with the article they seek? Not significantly, no. The article is 87,266 bytes of wikitext, plus three images (and transclusions etc). The DAB is 1,046 bytes of wikitext with no images. Andrewa (talk) 05:29, 5 September 2019 (UTC)
NYRM2016 is a manifestation of the reluctance to clearly and simply define primary topic exclusively in terms of the likelihood-to-be-sought criteria. But that kind of ridiculous debate is hardly exclusive to RMs focused on determining primary topic, or to RMs at all. It's certainly not evidence against having primary topics at base names. Regardless of the issues with moving Paris or Albert Einstein, your arguments support the position that if we were starting over with a clean slate, we should not have any ambiguous titles at all, and that those two titles would be disambiguated. If that were the case, then users searching with "paris" and "einstein" (currently a primary redirect to Albert Einstein, appropriately) would be taken to DAB pages rather than to the articles they seek.They would not be an improvement by any reasonable measure, and doing anything similar just to new articles is just as flawed for the same reasons. --В²C ☎ 16:57, 5 September 2019 (UTC)
That seems to have about 0% new material, so those who have been following the discussion can probably skip my replies.
NYRM2016 is a manifestation of the reluctance to clearly and simply define primary topic exclusively in terms of the likelihood-to-be-sought criteria. That might have helped, but if they couldn't see the merits of the move case as it was, I'm skeptical that page views would have convinced them. But I admit I do not understand. The two judgements that found no consensus almost look like text written under duress.
But that kind of ridiculous debate is hardly exclusive to RMs focused on determining primary topic, or to RMs at all. It's certainly not evidence against having primary topics at base names. Complete non sequitar IMO.
Regardless of the issues with moving Paris or Albert Einstein, your arguments support the position that if we were starting over with a clean slate, we should not have any ambiguous titles at all, and that those two titles would be disambiguated. Agree. But when we do consider these issues, it's a different ball game. So can we give that straw man a rest?
More to follow. I need to save regularly as my connection is flakey. Andrewa (talk) 01:29, 6 September 2019 (UTC)
Likelihood-to-be-sought criteria is, intellectually, stupid, when viewed in isolation. It ignores the purpose of encyclopedia as a structurally organised repository of information in favour of a very primitive emulation of a component of a search engine.
New York was difficult because, among other reasons, it is not a pure PrimaryTopic problem. The state and the city are far from independent from each other. Each article presents a manifestation that derived from a colony, each with a good claim of direct descent. What do people mean when they think "New York" is a non-factual question that is best handled by artificial intelligence, and a modern search engine is doing that very well. An encyclopedia present comprehensive coverage with navigation to narrowly defined topics. New York is a good candidate for the Broad Conception Notion that sidesteps the very problematic PRIMARYTOPIC conundrums. --SmokeyJoe (talk) 01:33, 6 September 2019 (UTC)
Agree except it is not a pure PrimaryTopic problem. The complications you list are some of what made it a PrimaryTopic problem. But what we mean by Primary Topic is, whatever belongs at the base name, and deciding that is sometimes problematic. Andrewa (talk) 01:44, 6 September 2019 (UTC)
No more to follow after all. The rest is just follow-up on the straw man argument, or in other words that I'm somehow obliged to let B2C choose what evidence I'll disregard in deciding what I recommend. Andrewa (talk) 03:04, 6 September 2019 (UTC)

Recap

The most important thing abut NYRM (which started this whole section) isn't that a group of editors, by whatever means, succeeded in keeping the New York State article at the base name New York, where it clearly didn't belong, for many years. The most important thing is that they did have a case in terms of reader experience, but not under current policy.

The case they did have was that the New York City article didn't belong there either, even if (as I and many believe) it is the Primary Topic. It's much better to have the DAB at the base name. But that is contrary to current policy.

We now have another case at Talk:The Americans (2013 TV series)#Requested move 18 June 2019, with strong arguments both ways. There has also been some discussion of this case above. I'm not wishing to canvass !votes either way there, and have no intention of !voting there myself. My point is that there are good arguments there both ways, under current policy. Nor have I posted a heads-up at the RM to the relevant discussion above. I can argue that both ways too.

Rather I maintain that in both of these cases and many others, the reader is better served by having the DAB at the base name, even if one of the articles is Primary Topic, and that we should change our policies and procedures to reflect this.

Perhaps (as originally proposed above) we should also change the no consensus rule, but that is not necessary to fix the problems with either of these articles. It should be a separate issue. Andrewa (talk) 16:49, 25 June 2019 (UTC)

These are very different cases. In NYRM there are two topics that can at least vie for PT. Anyone searching with "new york" could easily be looking for either the state or the city. In The Americans, one use decimates the others in terms of likelihood of being sought by anyone searching with "the americans". Today, and for the foreseeable future, it's clearly the 2013 TV series. It's the epitome of no contest. How the reader is better served with the DAB page at the base name is beyond me. We do agree on one general point: current policy is problematic. I believe the problem is it's way too vague - that's what allows so many reasonable opposing arguments. That's policy/guideline failure. --В²C ☎ 19:48, 25 June 2019 (UTC)
That's not a failure at all. Minimal rules are what has allowed WP to evolve towards its best self through true consensus, that is through the millions of edits and thousands of individual discussions among large numbers of editors, not straightjacketing rules decided by the relative handful of regulars who frequent policy talk pages. Too many rules stifle growth and discourage editors. If there are "many reasonable opposing arguments", that means there is no true consensus. Making New York a dab page has hurt more readers than it has helped, but has both hurt and helped only to a small degree, even for such highly viewed articles. In the end, it doesn't really matter that much if some true primary topics (whatever that means) are at their base name and others are not. Usually everyone finds what they're looking for, even if it takes an extra click or two. The far greater danger is a culture of some editors telling all other editors how to do things, at a macro level. Station1 (talk) 20:27, 25 June 2019 (UTC)
Minimal rules? We're way beyond minimal rules, and that's what often leads to all the unnecessary conflicts. We need to minimize the rules and make them more clear; less contradictory. --В²C ☎ 20:33, 25 June 2019 (UTC)
Conflicts are not always unnecessary or 'bad'. They often clarify consensus incrementally. My fear is that when rules are made more clear and less contradictory, they may well be in a direction that neither you nor I nor large numbers of other editors may agree with. Then we're stuck. Station1 (talk) 21:06, 25 June 2019 (UTC)
Well said. Have you seen User:Andrewa/Rules, rules, rules? I'd be very interested in your comments. Andrewa (talk) 06:25, 26 June 2019 (UTC)
Agree that New York and The Americans are very different cases. That's the whole point. In both cases, P T has been a failure. Andrewa (talk) 22:10, 29 June 2019 (UTC)

The Arericans has now been moved with the closing summary The result of the move request was: moved per WP:PRIMARYTOPIC criteria. WP:PRIMARYREDIRECT wouldn't work per WP:PRECISION. If the article enters into the predicted slow decline and later is no longer the primary topic, the articles can be rearranged then. [10]

It will be interesting to see whether the predicted slow decline actually takes place, and what happens if it does. It's possible this will become a classic case of a move that is supported by the current policy despite not being in the best interests of readers. Andrewa (talk) 11:40, 30 June 2019 (UTC)

And it's possible (even likely) that this is a case of a move that is support by the current policy because it is in the best interests of the readership, and that's why the current policy is the current policy. -- JHunterJ (talk) 13:03, 1 July 2019 (UTC)
Agree (except I'm skeptical that it's even likely, but that's not the main point IMO). And your input at User talk:Andrewa/P T examples and scenarios#The Americans, where I will answer this in more detail, would be very welcome. IMO the RM was a good call given the current policy. Consensus can change of course, but I have yet to claim or achieve consensus support for my rather radical proposal on Primary Topic. And the proposal has itself changed radically since it was first formulated, and will probably change further before any RfC is moved by myself. Andrewa (talk) 22:58, 1 July 2019 (UTC)
I'll be happy to participate in the eventual RfC, but there too it will be a repeat of the simple statements I've made here. -- JHunterJ (talk) 11:48, 2 July 2019 (UTC)
Possibly, but hopefully not! (;-> I won't be moving an RfC until and unless I have more evidence, such as the eventual result of the RM discussed at User talk:Andrewa/P T examples and scenarios #The Americans, and analysis of more of the examples on that page. I'm sure you would respond to the evidence presented. Andrewa (talk) 01:04, 3 July 2019 (UTC)
That page is dauntingly long, but the most interesting cases can now be discussed at User talk:Andrewa/P T test cases... the name says it all. There are only four listed currently, but they include The Americans. Discussion there more than welcome... as I've often said, I see changing my mind as a way of building it, not diminishing it, so discussion that helps me to do this is always welcome, and one of the benefits of editing Wikipedia. Andrewa (talk) 15:02, 12 July 2019 (UTC)

Search box algorithm

Where do I find the search box algorithm?

Because I think it could be improved.

In particular, it seems to discriminate against redirects. If I start to type in a title that exists as a redir, it doesn't even appear in the dropdown list until the exact match is complete. See

http://alderspace.pbworks.com/w/page/135472575/Wikipedia%20search%20box

for an example... the upper screenshot has only one letter still to type, but no hits. But when I add the last letter (lower screenshot), bingo, there it is.

Surely, that could be improved upon? Andrewa (talk) 21:42, 8 September 2019 (UTC)

mw:Help:CirrusSearch has an overview and links at the bottom to further details. Lieutenant General of the Armed Forces was only created three hours ago and may not yet have filtered through to the index of titles close to search strings. Certes (talk) 22:00, 8 September 2019 (UTC)
Awesome, thank you!
But the timing doesn't seem to be the problem... try Wikipedia:Alternatives to deletion for example. Same result exactly. Andrewa (talk) 22:08, 8 September 2019 (UTC)
When I type "Wikipedia:Alte", that redirect appears as suggestion 8 of 10. When I type "Wikipedia:Alternatives to d", it is the only suggestion. Certes (talk) 22:13, 8 September 2019 (UTC)
Curiouser and curiouser... I wish I'd taken screenshots! Because now that's what happens for me too.
But it's not what happened when I first looked for that redirect (relevant to another discussion). Nor when I immediately retried it just to be sure. Thanks for the links above, I will investigate further. Andrewa (talk) 22:48, 8 September 2019 (UTC)
Information on your “search” box behaviour can be found at Special:Preferences#mw-prefsection-searchoptions. Note that it is very short of a search engine search, it is title auto-complete allowing for two typos, by default. It is very useful, it is good for what it is, but it is not “search”, but goto. —SmokeyJoe (talk) 23:23, 8 September 2019 (UTC)
I think that to call it not “search” is just confusing... I use it as a search regularly, and the box I see is labelled "Search Wikipedia". I take your point that the type of search is different to say Google, but it's still a search so far as most people are concerned. Andrewa (talk) 08:08, 9 September 2019 (UTC)
I think the labelling of the “Go to title” box as “search” has been and is horrendously confusing. If you enter text and go, it does not invoke the search function, but does precisely as described in your preferences. To access a real search, you have to scroll down below all the suggestions, or you can invoke the GUI with no entered text to get to the real full search function interface. Usually, it is far easier to use google with a “+wikipedia” term. If you do that, google will remember your personal unarticulated preferences. —SmokeyJoe (talk) 22:50, 9 September 2019 (UTC)
You can do a real search by preceding the title with a tilde, e.g. ~wave, but that's not intuitive. Certes (talk) 22:58, 9 September 2019 (UTC)
Where is that documented? Andrewa (talk) 23:38, 9 September 2019 (UTC)
Help:Searching#Search box. Certes (talk) 23:54, 9 September 2019 (UTC)
Thanks, got it! Andrewa (talk) 00:37, 10 September 2019 (UTC)
I think it's only confusing to those with some background in computing science, and not even all of those. We're a general encyclopedia. Most of our readers understand search Wikipedia without confusion or difficulty. Andrewa (talk) 23:38, 9 September 2019 (UTC)
I am skeptical of that. I believe that most readers are familiar with searching by google, which is a powerful AI dynamically learning search engine. The confusion I allude to is with editors, and I believe that most editors have some computing knowledge. I believe in precise titling and labelling. For the Go box, it should be labelled "Go to title". The magnifying glass should be reserved to a real search, pressing it should always invoke a real search engine. At least, if you attempt to go to a non-existing page, the search function is invoked, but if there is a bad title, a bad redirect, or an overzealous PRIMARYTOPIC decision, a Go box user who think they are searching for something will be waylaid. --SmokeyJoe (talk) 00:11, 10 September 2019 (UTC)
That reminds me of dealing with IBM... occasionally, when they couldn't fix a problem for which they'd accepted an APAR, they'd close it as "Documented as a restriction" and eventually the manuals would be updated to say that the bug existed.
I think our users want a search box. I certainly use and value it as such. If we can improve it we should. But it's not just a go to and most users wouldn't even know what that meant, although they would guess close enough maybe. Andrewa (talk) 00:37, 10 September 2019 (UTC)
Au contraire, it is barely more than a GOTO. It is more in that it handles 2 typos and has some redirects functionality when generating the JAVA dropdown lists of suggestion auto-completions, and in the case of zero matches it invokes the search function. For what it is, it is great, but it is mislabelled, it is not a search box. The criticism are: (1) it can suck readers down holes of bad, imprecise titling, if they really wanted to do a search; and (2) it is causing editors to misuse the word wikt:search. --SmokeyJoe (talk) 00:44, 10 September 2019 (UTC)
Very interested in how it can suck readers down holes of bad, imprecise titling. Certainly agree that imprecise titling is bad. We should fix it! Andrewa (talk) 01:39, 10 September 2019 (UTC)
It's also possible that yesterday's denial of service attack is still having effects or is being repeated. That might prevent your partial title from being transmitted to Wikipedia's servers, or the suggestions from being returned. I briefly had timeouts a few minutes ago, though the problem may lie elsewhere and everything works smoothly for me now. Certes (talk) 23:37, 8 September 2019 (UTC)
Curiouser and curiouser... I wasn't even aware of that DoS attack. Andrewa (talk) 08:08, 9 September 2019 (UTC)
A Dutch friend (not an editor) saw a news report on 6 September that Dutch (but not English?) WP was down, possibly an attack. I suspected an attack on English WP at about that time, I couldn't load any WP pages (I tried in those two languages). Narky Blert (talk) 14:53, 9 September 2019 (UTC)
Wikimedia handled the DoS in a low-key way: see WP:VPT#Wikimedia down and the Phabricator report. If the Twitter account (now suspended) is to be believed, those responsible soon moved on to various gaming servers but gave no details of their methods. (I suspect a botnet abusing IoT, which is an accident waiting to happen.) NCSC seem to have a clue and may be able to prevent a recurrence. Certes (talk) 16:19, 9 September 2019 (UTC)
The same minute I suggested above that DoS may be the problem, Phabricator was updated to confirm that there had been more disruption. The problem seems to have gone away and, if it recurs, it will be a symptom of a more widespread problem. Certes (talk) 18:57, 9 September 2019 (UTC)
Now working for me: suggestion 5 of 10 for "Lieutenant General o" and sole suggestion for "Lieutenant General of the A". I think that either new pages take a while to index, or there were connection problems. One thing I have noticed is that redirects are suppressed when their target is shown: typing all or part of "Adipocyte" doesn't suggest Adipocytes. That feature seems sensible. Certes (talk) 13:26, 9 September 2019 (UTC)
I can see the logic, but that feature seems to me to be undesirable in at least some scenarios. It means that if a surfer types in wave wanting the physics article, this is what they see. It would be far better for them if wave (physics) appeared on that list.
And that is probably an example of the problem I see with our treatment of Primary Topic. We have been unimaginative. Someone recently said that the search box is not even an example of a search. To many of our readers (myself included) it is the normal way of searching Wikipedia. More important, we assume that everyone agrees on the default meaning of a term... and we call this meaning Primary Topic of course. But for many if not most ambiguous terms, some of our readers will assume a different default meaning. And that will be true whatever criteria we use for Primary Topic. Andrewa (talk) 21:19, 9 September 2019 (UTC)
I can see the advantage of suggesting Wave (physics). On the other hand, we don't want to fill the list with ten redirects to the same article, to the exclusion of other topics, especially if five of them are misspellings. There are possible solutions but they may require more effort than the benefits merit. For example, we could introduce a concept of "main qualified title" – the title we'd use if the topic were not primary – but someone would have to identify and mark that redirect for millions of articles as well as modifying the software to display them. Showing no redirects is probably less bad than showing them all. Certes (talk) 21:54, 9 September 2019 (UTC)
All good points. Redirects are already categorised. Redirects labelled R from misspelling, R from incorrect name etc should be sorted last in the list. But R from alternative name should be above them. Redirects from article moves should be shown unless specifically excluded, and a template and category should be set up to allow this... R excluded from searches perhaps. And probably an explicit R included in searches for completeness. I don't think any other new concept would be needed. Andrewa (talk) 23:38, 9 September 2019 (UTC)

Missing lines

As well as wave (physics) there's another important omission here, and that's wave (disambiguation). It would be really helpful to let the reader know that this exists... and maybe to tag it in some way so they get a hint that it's a list of the articles they might want. I'm not convinced that every reader, particularly a new reader, will correctly guess the meaning of (disambiguation). Andrewa (talk) 01:22, 10 September 2019 (UTC)

For searching Wikipedia mainspace, I much prefer to google "Kiss the Sky +Wikipedia" for example. I pretty much only use the internal Wikipedia search engine for searching other namespaces. --SmokeyJoe (talk) 05:06, 10 September 2019 (UTC)
I use it often in the main namespace. It almost always works better than Google, in my experience. Suppose I want the chemical element krypton. Why would I type krypton site:en.wikipedia.org (which just took me to Krypton (TV series) with no hatnote, no idea why) or even just krypton wikipedia (which takes me to the element as the second hit, after the TV series) when typing krypton in our search... sorry, non-search labelled search... box takes me straight to the correct article? Andrewa (talk) 06:06, 10 September 2019 (UTC)
You use what, the "Go to title" function, or the Wikipedia internal search engine. The "Go to title" function works great when you know what you are looking for and are confident that a page on it exists, and it has a simple title. It works poorly for something less tangible that might be covered in a section in some other article, or if you have the name wrong. Google is great for when you are searching for something common but you are using the wrong words, google somehow "knows what you mean". The Go to box is useless if your words are on the tip of your tongue and wont come off.
Parenthetically disambiguated articles, like Krypton (TV series), don't get hatnotes. You are supposed to know to remove the parenthetical part from your url to get to the DAB page, or the unexpected primarytopic. This is easy on a desktop, annoying on an iPad, and a right pain in the arse on an iPhone in mobile view. --SmokeyJoe (talk) 06:15, 10 September 2019 (UTC)
I type krypton into the box labelled Search Wikipedia and press the enter key. Almost always works well. Am I the only one who does this? It seemed rather intuitive.
And I'm not for one moment suggesting that Krypton (TV series) should get a hatnote just because Google is wrongly taking readers there. The point is just, Google is not God. We easily can and should do far better with our own internal navigation aids. (And quibbling about whether these are really "search" or not... words fail me. Our own documentation pretty consistently calls it search, and I think that's what our readers understand it as. But if we form consensus to change what we call that box, fine.) Andrewa (talk) 06:40, 10 September 2019 (UTC)
Is it "quibbling", or do words matter? Using the Go Box is not like doing a search. Using the Go Box is like going to the alphabetically listed contents and choosing a quick match. That is not a search by any reasonable definition, but Wikipedia is teaching readers and editors alike that it is. If we are to accept the Wikipedia interface as beyond the influence of mere mortals, than we need to recognize the creation of yet another word with multiple conflicting meanings (Contronyms). They should be avoided. There has been a lot of talk of "what readers will search for", which is firstly a psychoanalysis of a hypothetical person, and secondly uses a word that carries contradictory meanings. Even if that can't be fixed, is should be recognized. --SmokeyJoe (talk) 07:10, 10 September 2019 (UTC)
Words do matter a great deal. I'm very much a fan of Benjamin Whorf et al in that regard. But trying to impose this esoteric distinction is quibbling. We know what we mean by search, and if English is sloppy in this regard this is not the place to fix that.
But just to repeat, if you can establish a consensus to change the text from Search Wikipedia to something else, I'll go along with it. I probably won't support it, depending on the exact proposal, but you can still form a rough consensus. Have a go by all means. Andrewa (talk) 09:39, 10 September 2019 (UTC)

The above is such a mess re stringing that I'm giving up with this particular reply. See MOS:INDENTMIX and User talk:Andrewa#Mixed indents.

But replying to Wave is a failure of WP:PRECISE, and a poor application of PRIMARYTOPIC. It should be moved to Wave (physics). It fails as a PRIMARYTOPIC because it is not the PRIMARYMEANING, but a derivative of that. [11]

Exactly. Except it's not an unusual failure, just a good example. To the point that I have now of course come to the conclusion that we have such a poor idea of what we even mean by Primary Topic that it would now be better deprecated.

Any ambiguous article title fails WP:PRECISE and is just asking for trouble. Articles that are already at ambiguous names (or are at names that later become ambiguous) shouldn't be moved, that's also just asking for trouble. But just deprecating P T for new names is almost as good, and very little trouble indeed. Andrewa (talk) 07:11, 10 September 2019 (UTC)

Primary topic and Incomplete disambiguation conflicts

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.
There is a consensus in favour of Option 2, and that the standard for making disambiguated titles such as Foo (bar) a primary topic among all Foo's that are Bars should be tougher than the standard for titles that don't have any disambiguator. (non-admin closure) Iffy★Chat -- 21:00, 3 September 2019 (UTC)

Hi all. I have recently closed a requested move discussion on Steve Smith (cricketer) (located here) which I determined there was no consensus. In short, those supporting the move to a further disambiguated article title argued that the current title was an incomplete disambiguation. Those opposing the move argued the article subject was the primary topic, and similar to some other cricketer articles (e.g. James Anderson (cricketer)), should be titled without their year of birth. Through reading this guideline, I was not able to determine how to title articles when there is a conflict between WP:INCDAB and WP:PRIMARYTOPIC, so I am opening an RFC to see if we can determine which takes precedence in future discussions. The options forward I see (there may be others) are:

I would appreciate the community's thoughts on this matter. Steven Crossin Help resolve disputes! 07:56, 23 July 2019 (UTC) replaced references to PRIMARY with PRIMARYTOPIC --В²C ☎ 21:23, 23 July 2019 (UTC)

IMO a similar argument applies where there is no, or another, PTOPIC, and a qualifier is needed; as with e.g. Steve Smith and James Anderson. If one topic with a qualifier is overwhelmingly better known than others with the same partial qualifier, then it is PTOPIC to that limited extent. I am more reluctant than many to argue for a PTOPIC: they collect bad incoming links which are difficult to find and fix. That said, I agree with your reasoning and the result in that recent RM. Narky Blert (talk) 08:31, 23 July 2019 (UTC)
There is no PRIMARYTOPIC v. INCDAB conflict to resolve. WP:PRIMARYTOPIC does not apply because there is no PRIMARYTOPIC. The challenge is INCDAB: "it should redirect back to the main disambiguation page". Note that the word "should" is not "must". The INCDAB text could be improved, or it could be accepted that "should" allows for many exceptions. --SmokeyJoe (talk) 23:40, 23 July 2019 (UTC)
What about cases where it isn’t "almost certain", but just "significantly more likely", which is the standard for identifying a primary topic? For example, a couple of possible RMs:
Readers seeking House of Cards (TV series) are significantly more likely to be seeking House of Cards (American TV series) (90%) than House of Cards (British TV series) (10%), so move the American series to the "(TV series)" title.
Readers seeking Battlestar Galactica (TV series) are significantly more likely to be seeking Battlestar Galactica (2004 TV series) (75%) than Battlestar Galactica (1978 TV series) (25%), so move the 2004 series to the "(TV series)" title.
One might question whether those particular RMs would go through, but many ones like them would if we apply the standard PT test to disambiguated titles.--Trystan (talk) 14:45, 1 August 2019 (UTC)
If there is a primary topic (as determined by the usual means) for a given search term, then that search term should lead directly to that article because that is what is best for readers. Only when there is more than one search term an article is the primary topic for should this equate to a redirect. It is entirely irrelevant whether this will lead to 10 or 100,000 instances of titles like Alabama (band)/Alabama (Canadian band). The only thing that matters is that we maximise the chances of a reader finding the article they are looking for with as little hassle as possible. If the US and 2004 TV series are primary topics, they should be moved as you suggest. Thryduulf (talk) 15:17, 1 August 2019 (UTC)
Those moves would be quantitatively, objectively worse for getting readers to the right article. Incdab redirects with no incoming links get very minimal traffic. Yesterday, the 30-day stat for the House of Cards one was 43, and 1 (!) for the BSG one, compared to tens of thousands of views for the related articles. If those moves were made, an extremely small number of readers would get to the article without seeing the dab (and of course not all would be at the right article). But far greater would be the number of readers who would now arrive at the wrong article because they picked the incdab title out of the search autocomplete, having no way to tell it’s not the one they are looking for.--Trystan (talk) 13:53, 2 August 2019 (UTC)
Trystan, realizing this is what prompted me to start the #Article space descriptive shortcut redirects to articles with parenthetically disambiguated titles section below; such redirects are more likely to be actually used as search terms in many cases than the parenthetically disambiguated titles. That said, I don't see why the light direct usage of the PDAB titles would make "getting readers to the right article" "quantitatively, objectively worse". --В²C ☎ 16:58, 2 August 2019 (UTC)
The benefit of the hypothetical moves described above is that most people currently going to the incdab title (eg, BSG (TV Series)) would get to the right article without going through the disambiguation page. The cost is that people looking for the lesser-accessed article, like the 1979 BSG or the UK House of Cards, are more likely to end up at the wrong article, for example by picking “BSG (TV series)” from the search box autocomplete (where as now, the autocomplete offers up fully-disambiguated options). Because the people currently going through the incdab redirect (up to dozens) is so minuscule compared to the views that these lesser-viewed articles get (10-20 thousand), the cost will be much higher than the benefit.--Trystan (talk) 18:45, 2 August 2019 (UTC)

Important questions

There are some really good points made above IMO. I'm not yet ready to !vote on the RfC, but I'd like to discuss. All of the following is said above or at least hinted at.

There's a critical difference between natural disambiguation and parenthetical disambiguation, and the rules may not necessarily be the same, and aren't, and our current policies and guidelines do say this but perhaps not clearly enough.

A natural disambiguation is an English noun phrase that can be expected to appear in reliable secondary sources, and it isn't even a candidate for an article title unless it does appear in them. Our disambiguation conventions (by which I mean any construction with which we disambiguate which is not a natural disambiguation) on the other hand produce titles that will not necessarily appear in sources at all, although comma disambiguation in particular may.

So our disambiguation conventions give titles which we may adopt even if they do not ever appear verbatim in sources. All such conventions should probably follow the same rules... parentheses, commas, are there any others? I guess we should look to see whether the pipe trick recognises any others.

This should be reflected in our article title policy and already is implicitly but perhaps not even consistently. And it seems to me that this RfC is seeking to clarify exactly this. Andrewa (talk) 00:38, 15 August 2019 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Implementation

The recent changes to the WP:INCDAB guideline do not match the consensus of the above discussion. Participants were nearly evenly split between whether parenthetically qualified titles should have primary topics. There was some middle ground. Several contributors who thought that qualified titles should have primary topics supported a significantly higher standard than the normal PT test. Several contributors who thought that qualified titles shouldn’t have primary topics allowed for rare exceptions. An addition to the guideline based on the above discussion should reflect this middle ground.

However, as implemented in the guideline, it’s straight Option 2: apply the standard primary topic test to qualified titles. It states that such instances are rare, but stating they are rare doesn’t make them so. There is no reason to think they will be rare as the guideline is written.--Trystan (talk) 13:39, 11 September 2019 (UTC)

Agreed. The guideline needs to reflect the whole of the (concise) closing statement, not merely one part of it. Certes (talk) 14:17, 11 September 2019 (UTC)
I have re-added a sentence to INCDAB which sets out the consensus that a higher standard applies to qualified titles.--Trystan (talk) 12:48, 12 September 2019 (UTC)

Is Dorotheus a DAB or a name page?

Dorotheus is currently more of a name page than a disambiguation page, though it has a ((hndis)) template, and Dorotheus (disambiguation) redirects to Dorotheus. I'm thinking that the best option would be to move Dorotheus (weevil) (and its blue link) to Dorotheus (disambiguation), add the names that are just 'Dorotheus', and include a link to Dorotheus (of course). What would be the appropriate template to put on Dorotheus given that it includes given names and mononyms? Does ((given name)) take care of it? Leschnei (talk) 13:46, 10 September 2019 (UTC)

Technically the weevil (which doesn't even have an article) makes it a dab but in spirit it's a name page. I would say that ((given name)) covers mononyms. (Even if there were surnames, ((given name|type=both)) would suffice.) A hatnote can deal with the weevil: we don't need a dab just for ONEOTHER meaning. Certes (talk) 14:14, 10 September 2019 (UTC)
@Certes: thanks for the input. I've replaced ((hndis)) with ((given name)). However, I didn't add a hatnote because Dorotheus (weevil) is a red link. I moved it to See also; maybe not the best solution but I'm sure that someone will put me right. Leschnei (talk) 01:04, 11 September 2019 (UTC)
Yuck! what a mess! part given name, part hndis (everyone on the page seems to have a mononym), part dab. Why can't people and things arrange themselves into tidy categories?
I suggest leaving the page as a ((given name)) page with the weevil as an SA. Messy and unstylish, but it has the merit of containing all the relevant information to point readers to what they might be looking for. Narky Blert (talk) 18:58, 13 September 2019 (UTC)
More