Wikipedia talk:Manual of Style/Disambiguation pages

Proposal to resolve nomenclatural confusion between split long lists and parenthentically disambiguated page names

FYI

 – Pointer to relevant discussion elsewhere.

Please see Wikipedia talk:Naming conventions (lists)#Fixing disambiguation confusion.  — SMcCandlish ☏ ¢ 😼  22:49, 25 January 2024 (UTC)

There's a discussion about band names as parenthetical disambiguators

See Wikipedia talk:WikiProject Albums#WP:THEBAND in article titles, where the original poster wrote:

So the possibilities seem to be:

• Option A: Strength (The Alarm album)
• Option B: Strength (the Alarm album)
• Option C: Strength (Alarm album)

What do we do?

The wise users of this page may have some ideas. SchreiberBike | ⌨  03:12, 26 January 2024 (UTC)

IAR restatement does not belong on this page

For some reason this page devotes a section to a restatement of WP:IAR, which I removed yesterday since it has nothing to do with DAB pages. Editor PamD put it back with the explanation that "it's a useful reminder that dab pages are there to help readers".

This is not at all true. Quite the contrary, it wastes the time of every reader who gets that far in the document. A conscientious reader will go through the section, then re-read it trying to figure out what the relevance is to DAB pages, and finally realize that there is none. Just because IAR applies to all Wiki pages, it doesn't belong on the every Help page. Dan Bloch (talk) 22:08, 26 January 2024 (UTC)

I disagree. It's helpful in the specific context of dab pages to remind editors that the detailed rules laid out here will sometimes be overridden by the need to make the page serve it's intended purpose.--Srleffler (talk) 20:59, 27 January 2024 (UTC)

Why is it helpful here, but not helpful in every other Wikipedia: page? Dan Bloch (talk) 21:40, 27 January 2024 (UTC)

Because dab pages are far more rule-bound than most pages are. Dabs have a highly constrained style. Editors sometimes forget that even in such a tightly constrained case, sometimes you have to break the rules to accomplish the goal.--Srleffler (talk) 00:03, 28 January 2024 (UTC)

As IAR applies to any guideline, I would expect this section to have a few representative examples to illustrate its point and justify its existence. —Bagumba (talk) 01:12, 28 January 2024 (UTC)

At bare minimum. And they would have to be essentially incontrovertible examples nearly no one could disagree with. I'm skeptical that's possible, and tend to agree with the OP. Guidelines should never pre-emptively cite IAR; it just invites unjustified "exception" making on an "I like it better" basis.  — SMcCandlish ☏ ¢ 😼  07:04, 3 February 2024 (UTC)

MOS:DABCOMMON and pseudo-primary / almost-primary topics

MOS:DABCOMMON says that if there is a single WP:primary topic, to put that by itself at the top of the dab page. But what about when no single topic is primary but one topic is the most common among the various topics? See, for example, Twister and This version of Jump. I was under the impression we should only format a dab page that way if there is a dominant primary topic (i.e., if the dab page is not at the base name). So I changed the format of Jump as shown in the current version [converted to a permalink for posterity].

Should we change item 1 of the instruction to say something like "The primary or single most common topic, if there is one, should be placed at the top."

Should my change to the format of Jump be reverted? Should the format of Twister be changed?

—⁠ ⁠BarrelProof (talk) 05:13, 21 February 2024 (UTC)

There's a threshold for making a topic primary, putting it at the base name and calling any dab Foo (disambiguation). There's another threshold for hoisting a topic to the top of a dab. Is the second lower than the first? In other words, is there any case where a topic is close enough to primary to appear at the top of the dab but not primary enough to live at the base name? If not then we should either move the topic to the base name or demote its dab entry to the ranks. The two cases here are complicated by the fact that they would be a primary redirect rather than an article title. Does Twister (tornado) or Jump (leap) qualify as meeting the second threshold but not the first?

I don't support the current wording of Jump, where an artificial topic of "several songs" has been created as a secondary entry for WP:DABCOMMON. If we are going to hoist Jumping, better to leave it as the only entry at the top. Certes (talk) 11:24, 21 February 2024 (UTC)

I'm not sure I quite understand the base question, when it's covered by the existing phrasing of MOS:DABCOMMON, just below the mention of the primary topic: In cases where a small number of main topics are significantly more likely to be the reader's target, several of the most common meanings may be placed at the top, with other meanings below. See Mojave or Mercury for examples of this.

The exact details of the formatting of the top part is actually an area where we don't know enough about. I don't think I've ever seen any research done, such as trying to figure out what's the most common way our average reader parses a disambiguation list. The guideline text has mostly grown organically, with bold edits by editors, not necessarily based on hard science.

I recall seeing at least one case - echoes - where we had the single common link formated as a link inside the top sentence. This didn't seem to have the desired effect, so I reformatted it to be part of a top list with a longer caption. The next month the clickstreams showed more readers noticed it - though still not very many.

At Charlotte, I remember we had added a top list of common meanings, and then someone quickly added an explicit section heading. Stats showed a lot of interest in the top items of that common meaning section, though not as much for all of them (do note that we don't have too much of a sample here so as time passes this may change).

In the specific case of "several songs", that seems unlikely to be useful because it's not a link - we'd be forcing the reader to keep reading / scrolling / searching instead of just clicking, which seems suboptimal. I recently formatted such items with links to their sections at Major, you might want to have a look at that. --Joy (talk) 14:41, 21 February 2024 (UTC)

One interpretation of the base question is: New York usually means one of two topics, and they are listed prominently at the top rather than being buried in #Places. Is it ever good to do this with a single topic, or would such an important topic necessarily be primary? Certes (talk) 15:02, 21 February 2024 (UTC)

Yes, I think so, because we have numerous examples of a single topic being more popular than others, but not to the level of being a primary topic, because there's some sort of level of ambiguity that keeps it in a plurality as opposed to a supermajority, so to speak. In the RM for the aforementioned echoes, a lot of people thought echo was like that with regard to long-term significance. Something similar happened in/after the RMs for rumours, hurts, erasure. There's probably many more examples, these are just some of the recent ones that I could think of off the top of my head. --Joy (talk) 17:10, 21 February 2024 (UTC)

I agree with the OP that we should say that there can be a single popular topic at the top, even though it isn't the primary topic. This currently isn't clear. The MOS:DABCOMMON instructions called out, In cases where a small number of main topics are significantly more likely to be the reader's target, several of the most common meanings may be placed at the top... implies that the rule is only for cases with more than one such topic.

But the devil is in the details. What would it look like? Adding The primary or single most common topic, if there is one,..., as the OP suggests, implies that the new text would look the same as the primary topic case ("A school is an institution for learning"), which fails on two grounds: the wikilink (school) only succeeds because it's the primary topic; and additionally we would typically want the wording to be "most commonly refers to", not "is".

I think the new text should be added in the MOS:DABCOMMON section. Revised wording (new last sentence) might be something like:

In cases where a small number of main topics are significantly more likely to be the reader's target, several of the most common meanings may be placed at the top, with other meanings below. See Mojave or Mercury for examples of this. In the uncommon case where there is only one such topic and it is not the primary topic, it can be likewise be added at the top, e.g., "Twister most commonly refers to a tornado."

Dan Bloch (talk) 23:51, 21 February 2024 (UTC)

Jumping is almost certainly primary for the word Jump, so the base title should redirect to it, and the dab should be at Jump (disambiguation). We should definitely not add "or single most common" to WP:DABORDER #1, as primary topics and common meanings are two different things that are handled differently: primary topics are linked in a non-bulleted, complete sentence at the very top, followed by "X may also refer to:", whereas common meanings, even if there is only one, are regular, bulleted, sentence-fragment entries, with some sort of introductory line or header. I don't really think it's necessary to clarify that WP:DABORDER #2 applies to a single common meaning; after all, one is a "small number". But if such clarification is necessary, it should just be a short parenthetical: "In cases where a small number of main topics (or just one)"; the whole extra sentence proposed by Dan Bloch is really excessive IMO. —swpb^{T • beyond • mutual} 19:15, 22 February 2024 (UTC)

An RM discussion at Talk:Jump#Requested move 18 April 2017 did not reach a consensus on the primary topic question, so Jump is a disambiguation page at the base name. —⁠ ⁠BarrelProof (talk) 19:24, 22 February 2024 (UTC)

Well, if consensus is that Jumping is really a common meaning and not a primary topic, then it should follow the common meaning formatting. To me, it seems like the primary topic, but I'm only one editor. The important point is that it should be one thing or the other; there are no "pseudo-primary" topics. —swpb^{T • beyond • mutual} 19:28, 22 February 2024 (UTC)

I suspect there are a lot of cases where there is only one "almost primary" / "highest priority" topic, i.e., one topic that has much higher long-term notability and gets much more readership interest than the others, but not enough to meet WP:PRIMARYTOPIC criteria. For example, consider a hypothetical case where one topic has 35% of page views and the rest of the interest is scattered amongst 5+ other topics, and the dab page has 50 topics on it, and no other special considerations apply. In such a case, we cannot create a reasonable short list at the top that has more than one topic in it, because it is too difficult to figure out which and how many entries should be in that short list. I like the idea of using "most commonly" rather than "is" for such a situation. I agree that we should avoid letting the instructions get too complicated, but I think it is probably pretty common for the short list to only have one good candidate. —⁠ ⁠BarrelProof (talk) 19:46, 22 February 2024 (UTC)

This is an interesting situation. Wikinav for Jump shows 41% of outgoing views in Jan-2024 go to jumping and 32% to Jump (Van Halen song); in Dec-2023 the percentages were nearly equal (a slight edge to the Van Halen song). Two other topics had >5% as well. However, given that there are multiple songs by the name, I think it would be disservice to readers to pull out the Val Halen song into a common uses section. older ≠ wiser 20:14, 22 February 2024 (UTC)

I'm not committed to that sentence, but without it we still need an example. So how about:

In cases where a small number of main topics (or just one) are significantly more likely to be the reader's target, several of the most common meanings may be placed at the top, with other meanings below. See Mojave, Mercury, or Twister for examples of this.

Dan Bloch (talk) 19:47, 22 February 2024 (UTC)

Y —⁠ ⁠BarrelProof (talk) 19:59, 22 February 2024 (UTC)

I've made the change discussed above ("a small number of main topics (or just one)..."). Dan Bloch (talk) 19:40, 24 February 2024 (UTC)

The words "several of" don't make sense with that change, so I deleted those words. —⁠ ⁠BarrelProof (talk) 01:20, 25 February 2024 (UTC)

This issue also affects Foo (disambiguation) pages. For example, Apple (disambiguation) sensibly has a short list at the top because, when the reader types "Apple", we feed them the fruit which is good for them rather than Apple Inc. which most of them were actually looking for. Although three common topics are listed in this case, it's quite feasible that there might be just one common topic, even when a different topic is primary. Certes (talk) 20:03, 22 February 2024 (UTC)

Interesting case. For that one we have not just a short list, but also a regular primary topic identifier at the top saying "An apple is an edible fruit." That is followed by "may also refer to" and the short list. That seems sensible to me. —⁠ ⁠BarrelProof (talk) 20:41, 22 February 2024 (UTC)

There are plenty of similar cases. In the tech field we have Alphabet (disambiguation) – mainly for Alphabet Inc. – and Telegram (disambiguation), where Telegram (software) has 66 times more views than anything else including the primary topic (though Telegraphy gets almost as many). Elsewhere, we have Model (disambiguation) with Model (person) being five times as popular as the primary topic, and Baler (disambiguation) listing Baler, Aurora with nearly twice the following of the farm machine. Sports are a popular place for alternative meanings, with Arsenal (disambiguation) being dominated by Arsenal F.C. rather than the ammo dump. Certes (talk) 22:20, 22 February 2024 (UTC)

Let's note for the record that these aren't all identical, though - I don't think there's been serious proposals in recent memory to change most of these, because the long-term significance arguments have been inherently clear, but the discussions at Talk:Telegram (disambiguation) did show we do have a bit of a navigation problem there. --Joy (talk) 21:07, 24 February 2024 (UTC)

Switching to an -ing form in the middle of a disambiguation list

So, some dispute with an revert-warring anon has broken out, at the DAB page Hustling Hustler. There's a line item there that has long read:

• Hustler, someone who deceives others by hustling, usually in sports

following our normal practice of using the term that is the disambiguation page as the line-opening word. The anon absolutely insists on changing this to:

• Hustling, deceiving people into betting on a game at their disadvantage

though this does not have clear pertinence to the disambiguation Hustling (and if kept in this form would be moved to the "See also" section as a related word). The anon is badly misinterpreting MOS:DAB, but will not take no for an answer, at least from one party (and acuses of "vandalism" if challenged), so it will probably require multiple parties' involvement to settle this silliness. (PS: Their version also has the problem that it changes the definition to refer to games when the term is primarily used with regard to the related but distinct sphere of sports).  — SMcCandlish ☏ ¢ 😼  21:23, 21 February 2024 (UTC)

Seems like a basic rule of list-construction to use parallel forms. I'm not sure where in the MOS that might be mentioned (if it is at all). older ≠ wiser 21:40, 21 February 2024 (UTC)

@SMcCandlish:, what disambiguation page are you referring to? You linked four times to the article Hustling, but zero times to disambiguation pages. --R'n'B (call me Russ) 21:42, 21 February 2024 (UTC)

Hustler. older ≠ wiser 21:43, 21 February 2024 (UTC)

D'oh! Yes, I meant Hustler, which just goes to show that switching to "hustling" in mid-list is confusing.  — SMcCandlish ☏ ¢ 😼  23:48, 21 February 2024 (UTC)

But to be honest, aside from the form, I think the updated description is actually a bit clearer. The former presumes a reader knows what "hustling" is. older ≠ wiser 21:43, 21 February 2024 (UTC)

I tend to agree with you. The insistence on starting every line with the disambiguated term can lead to awkward formulations when carried to an extreme. It's a good style practice where it flows naturally, but style shouldn't trump comprehension. --R'n'B (call me Russ) 21:47, 21 February 2024 (UTC)

There is no comprehension problem of any kind with "Hustler, someone who deceives others by hustling, usually in sports" appearing in the disambiguation list of Hustler.  — SMcCandlish ☏ ¢ 😼  23:48, 21 February 2024 (UTC)

Really? Seems a pretty pedestrian error to use a variation of the same term to explain that term. older ≠ wiser 23:56, 21 February 2024 (UTC)

They could be combined:

• Hustler, someone who engages in hustling (deceiving people into betting at a disadvantage), usually in sports

That's just a matter of copyediting. The issue here is the attempt to force the -ing form of the word into a list in which is does not belong, and abuse of guideline wording that does not mean what the anon things it means, to try to make that happen.  — SMcCandlish ☏ ¢ 😼  23:51, 21 February 2024 (UTC)

Yes, this is better. older ≠ wiser 23:57, 21 February 2024 (UTC)

Isn't this a clear case for a one-off exception to MOS:DABPIPE? "Hustler, someone who deceives people into betting at a disadvantage, usually in sports". Link and term are at the left where expected, and there's no circular definition. —swpb^{T • beyond • mutual} 19:46, 22 February 2024 (UTC)

That's the right approach. The piped link makes it clearer while preserving the preferred form. --Srleffler (talk) 19:10, 24 February 2024 (UTC)

Sorting entries by clickstream

At IPA, User:Joy is insisting on sorting the "commonly refers to" list according to the clickstream data rather than alphabetical. Is there a guideline that supports doing this? Has there been discussion on sorting order on DAB pages? Nardog (talk) 02:03, 10 March 2024 (UTC)

The order for the commonly-refers-to list isn't documented specifically, but the spirit of the text suggests that MOS:DABORDER #4, which allows either way of sorting, should apply to it.

Also FWIW, of the two commonly-refers-to examples, Mercury is in frequency order and Mojave is near-random but closer to frequency than alphabetic. Dan Bloch (talk) 02:55, 10 March 2024 (UTC)

I've elaborated further at Talk:IPA#clickstreams data.

On a more general note, we know that alphabetical lists work well as a safe option when there's no other way for the reader to predict their traversal of a long list, but in short lists it's unlikely that readers spend a lot of time figuring out sorting.

The key question here would be what constitutes a short list. In the case of top 4 MOS:DABCOMMON items right at the top of the page, it seems easy enough, but we also have various different cases. For example, the WP:NAMELIST guideline has completely decimated lists of people, where practically none get listed up front in short lists, and practically all get listed in long lists. This has had a negative effect on navigation that I've complained about before - wherever we happen to have a vaguely popular item in the short list, it gets way more traffic than a otherwise very popular items in the long list.

Random Brazilian footballer lists, or for example various entertainers like singers or rappers, or in turn minor toponyms and whatnot, effectively spam the top-level disambiguation lists just because they are single-word matches, while biographies with vastly different long-term significance are relegated to long lists, where they're actually much harder to find. A recent egregious example of something like this is Talk:Boyle#post-move.

We'd really have to do some research about what actually makes the most sense for readers (Wikipedia:Readers first). From other web user research it's become generally know that the contents of the first page of text is the most relevant (i.e. scrolling is not great), and that people read typical websites in an L-shaped fashion, from the top left towards the top right and then from the top left to the bottom left (i.e. there's a huge difference in where something is). Obviously another question is if and how the average web user changes their behavior as they become the average Wikipedia reader. The difference between the observed Wikipedia reader clickstreams from the list at IPA and the general Google Search Trends for IPA are an interesting case for this - apparently people at the encyclopedia act differently in this regard than people at the general search engine.

In an earlier thread, I also noted differences in how the existence of a section heading and the length of caption may apparently affect the outcomes of our MOS:DABCOMMON lists. --Joy (talk) 10:52, 10 March 2024 (UTC)

Generic issue about dab page items

Please express an expert opinion in Talk:Qasr#Agreed: ONLY places called solely Qasr to be listed. The issue is whether two-part names, such as "proper_name + object_type", e.g., [[<something> Prison]] and the likes, belong to [[<something> (disambiguation)]]. - Altenmann >talk 21:00, 15 March 2024 (UTC)

The guidance for this is at WP:PTM. --Joy (talk) 22:30, 15 March 2024 (UTC)