On the whole, I think our policies and guidelines are great, but this is a space to collect my thoughts on a few particular places where I think our policy (or the de facto interpretation of that policy) leads us to poor outcomes. Maybe some day I'll polish one or two of these up into village pump threads or something, but for now they have Old Man Yells at Cloud status. If you have opinions on any of these, feel free to hit me up on my talk page. I'm always open to having my mind changed!
More aggressive splitting of articles helps us realize the promise of hypertext to improve the experience of navigation for readers. It also helps us exploit the power of Wikidata and the category system.
For example, Pedro Fernandes de Queirós is an instance of a common pattern where our main coverage of a notable expedition is in the article for that expedition's leader. Even though this arrangement doesn't necessitate a WP:SIZESPLIT, a split would give us a more granular link target for when editors want to refer to Queirós' search for Terra Australis rather than to Queirós himself. It would also play nicely with our categorization scheme, in that readers would be able to find information about the expedition when browsing categories like Category:Pacific expeditions or Category:Expeditions from Portugal. Similar arguments apply to covering a creative work in the article about its author.
Another example is productions of musicals. A revival or international production of a particular musical will often be the subject of significant RS coverage and easily pass WP:GNG, but we never give them dedicated articles. Each production has a separate set of information: they have their own cast, their own production history and dates, their own critical reception, and their own awards. It makes perfect sense to encapsulate that information in a separate article, where it can have its own dedicated infobox, cast list, categories, wikidata item, etc. But instead we dump it all in the article for the work, with the result that the article either becomes unmanageably big, or omits noteworthy verifiable details about individual productions, or both. This huge table of cast members across 9 different casts is very typical of articles on musicals, and would not be necessary if we followed summary style.
WP:SONGCOVER says: Only cover versions/renditions important enough to have gained attention in their own right should be added to song articles. When a song has been recorded or performed by more than one artist, a particular artist's rendition should be included in the song's article (never in a separate article)
(emphasis added).
I just don't see any rational reason for this. Whitney Houston's cover of I Will Always Love You obviously passes GNG by miles. There's plenty to say about that cover specifically (its chart performance, its awards and critical reception, its music video, etc.), but for some reason it needs to live as an article-within-an-article at the song article. How is this better for readers?
A somewhat analogous situation is video game remakes/remasters, which we do allow to occupy separate articles (e.g. Final Fantasy X/X-2 HD Remaster).
Related discussion at WT:Notability (music) from 2021-11: Really, a cover can never have an independent article? (scroll down to the bottom of the section for links to earlier relevant discussions)
Even a tiny, single-paragraph permastub can serve a useful purpose, especially if it's on a subject that is liable to be linked from other articles. I think it's important to keep in mind that the concept of notability is a direct consequence of our policy on verifiability. See the second sentence of WP:N: Information on Wikipedia must be verifiable; if no reliable, independent sources can be found on a topic, then it should not have a separate article.
If there is useful information about the subject which is verifiable (via reliable secondary sources), and it's not excluded by WP:NOT, we should have an article about it. To the extent that this risks opening the floodgates to promotional articles about marginal extant companies, people, software, awards, etc. I think it should be countered by vigorous enforcement of WP:NOTPROMO and similar. But for historical subjects that are very unlikely to attract any kind of COI editing, I think we should be very liberal about stubbing.
A common feature of borderline-notable/promotional articles about people or organizations will be an "awards" section that lists a bunch of non-notable awards without providing any context to the reader or indication of why they're noteworthy. Example - is this an encyclopedia article or a résumé? I wonder whether there should be some policy guidance about this situation?
MOS:BIRTHNAME advises to give the full birth name of an individual in the first sentence of the article. e.g. Joey Fatone starts off:
Joseph Anthony Fatone Jr. (born January 28, 1977) is an American singer, dancer, actor, and television personality.
This is silly. His full birth name is WP:TRIVIA, and arguably shouldn't even be in the article. If it were to go in there, it should be in the "Early life" section, certainly not the *very first thing that readers see*. cf. WP:ASTONISH. (To get an idea of the weight which Fatone's birth name is given in RS coverage of the singer, a Google News search for "Joseph Anthony Fatone Jr." brings up exactly two results, both of which are ultimately quoting from Wikipedia itself.)
MOS:DABMENTION: If a topic does not have an article of its own, but is mentioned within another article, then a link to that article may be included if it would provide value to the reader.
In practice, this is interpreted very liberally, and we see a lot of dab entries like "Foo, a planet from the Star Trek episode Foobar", where the planet gets just a passing mention in the plot summary of the target article. A memorable case was discussed at Talk:Split: should that disambiguation page include an entry for "Splits Magazine, a fictional magazine in the Glee episode "Home""? (Consensus was ultimately against the inclusion of such an entry, but the fact that a case like this that seems like it should be an obvious no-go was the subject of considerable debate suggests to me that the bar has been set way too low.)
Suggested additional criterion. Only add a dab entry if it's plausible that a reader would type ambiguous term X when looking for information about this topic.
The same test could also be fruitfully applied to borderline WP:PARTIAL matches. e.g. is it plausible someone searching for information about DVD players would search for "player"? No? Then let's remove that from that dab page.
Song titles are perhaps deserving of special mention here. These often take up a considerable amount of space on dab pages. Is it likely that readers are going to be searching for the names of deep album cuts and expecting to find articles about them? Dab entries do have a cost associated with them. I'm dubious as to whether every notable album ever should give rise to a dab entry (or redirect, if unambiguous) for each of its song titles.
A few thoughts on stub templates:
It occurs to me that, despite the WP:NOTDIC policy, whether we'll have a standalone article on a certain subject depends to a great (probably too great) degree on whether that subject has a name. For example:
It's also interesting to consider the flip side: are there topics which we probably would have articles for if only they had a name? For example, we have articles on smile, frown, smirk, glaring, etc. but no article on the open-mouth facial expression characteristic of surprise.
A lot of requested moves are opened with rationales that are patently contrary to WP:AT policy. By far the most common is "it's the official name". A discussion founded on a non-argument like this isn't really worth pursuing further. I'd love to see a protocol where these RMs can be speedily closed, with no prejudice against a new RM being reopened at any time with a valid rationale.
In theory, there are two sorts of category: topic categories and set categories. For a given topic area, sometimes there is a linguistically natural way to name these. For example, Category:Film contains articles and subcategories related to film (e.g. Film industry). Category:Film contains only things which are films. This is a very, very useful distinction to preserve.
But in a lot of cases we conflate these concepts, especially when there is no linguistically natural way to delineate them by name. If Category:Books is a set category containing only books, then what should we call the topic category for articles on topics relating to books (such as Book restoration or History of books)? WP:CATNAME suggests that the names of topic categories should be singular rather than plural, but we rarely do this for count nouns because a name like Category:Book looks silly. So we settle on Category:Books serving both roles simultaneously.
But the navigational and organizational benefits of keeping these concepts separate is worth the cost of a slightly awkward name. If there's no natural way to name the topic counterpart to a set category, I would suggest we just adopt some mechanical naming convention like Category:Topics related to books, or Category:Books (topic), or Category:Book-related. Yes, these are all aesthetically kind of awful, but I think we would get used to them.