Wikipedia talk:Article size#rfc 25B7179

How can this guideline be applied to lists? It does not seem to have instructions about when to split a list (which would indicate the inverse of how big a list can be). WP:SIZERULE seems focused on word count. I think this guideline had kB-size ranges before, but if it is gone, is kB size irrelevant? I'm asking all this because there is some interest in recombining list of cult films after a big overhaul last year (it has around 2,700 films from at most 20 book references). See discussion here: {{sectionlink|Talk:List of cult films#Combine pages}}. Erik (talk | contrib) 14:47, 28 January 2025 (UTC)

I did try the Prosesize tool, and here are the results for the T page:

• HTML document size: 180 kB
• Prose size (including all HTML code): 224 B
• References (including all HTML code): 37 kB
• Wiki text: 23 kB
• Prose size (text only): 102 B (19 words) "readable prose size"
• References (text only): 3138 B

The prose size does not seem to consider text in the table, so how can the B/kB values be used to determine an ideal list size? Erik (talk | contrib) 14:58, 28 January 2025 (UTC)

I [https://en.wikipedia.org/w/index.php?title=Wikipedia%3AArticle_size&diff=1277073253&oldid=1275017686 added an 'update' inline tag] to this sentence: {{tqb|A page of about 10,000 words takes between 30 and 40 minutes to read at average speed, which is close to the attention span of most readers.{{cite book |author1=John V. Chelsom |author2=Andrew C. Payne |author3=Lawrence R. P. Reavill |year=2005 |title=Management for Engineers, Scientists and Technologists |url=https://books.google.com/books?id=-aP30hEFB6sC&pg=PA231#v=twopage |edition=2nd |location=Chichester, West Sussex, England; Hoboken, NJ |publisher=John Wiley & Sons |isbn=9780470021279 |oclc=59822571 |page=231 |accessdate=20 February 2013}}}}

The source is outdated as it's from 20 years ago, and since then, attention spans have shrunk (quite significantly?) with the rise of smartphones and social media [https://www.apa.org/news/podcasts/speaking-of-psychology/attention-spans]. If anyone has a more recent source (preferably from 2023 or later) to update this sentence with, that would be great. Some1 (talk) 01:42, 21 February 2025 (UTC)

:(Disclaimer: I always favour large articles) That sentence also seems to presume that people ought to read the entire article, start to finish. That seems like a big assumption to base a policy or guideline on. And I don't think it's a good idea to make scientific arguments in a policy/guideline. Policies and guidelines are instructions and oughts, they don't need to justify themselves in detail. Jo-Jo Eumerus (talk) 08:28, 23 February 2025 (UTC)

::{{blue|That sentence also seems to presume that people ought to read the entire article, start to finish.}} I agree. I assume most readers probably just skim the lead and then head straight to the sections of the article that interest them. For example, I recently answered a question [https://en.wikipedia.org/w/index.php?title=Talk:Lyle_and_Erik_Menendez&diff=prev&oldid=1277170657#Marriage_witnesses on a talk page] where it seemed pretty clear that the IP had dived straight to the #Marriages section (or at least skipped the preceding sections of the article). And it's not just this Article size guideline, but some other guidelines/MOS also seem to assume that most people read Wikipedia articles from top to bottom or in a sequential manner. Some1 (talk) 14:46, 23 February 2025 (UTC)

{{reflist}}

@SandyGeorgia my recent edit [https://en.wikipedia.org/w/index.php?title=Wikipedia:Article_size&diff=prev&oldid=1292504067] was reverted.

I propose changing “almost certainly” to “very likely” in the >15,000 word guideline.

“Very likely” still communicates the need for trimming or splitting but better reflects editorial discretion and aligns with Wikipedia’s typical tone.

Almost certainly can sound overly definitive. In statistical or probabilistic language, “almost certainly” implies a near-100% probability, while editorial decisions on Wikipedia are ultimately subjective and context-dependent. Using “very likely” better reflects that editors should use judgment rather than follow an absolute rule.

Wikipedia typically favors language that guides rather than dictates, allowing room for editorial discretion. “Very likely” aligns better with Wikipedia’s tone, which avoids prescriptive language unless a policy is non-negotiable.

“Almost certainly” could be misinterpreted as a hard rule rather than a strong recommendation. ( I have seen this multiple times ) Cinaroot (talk) 05:11, 28 May 2025 (UTC)

• Support. I like the reasoning which I find compelling and complete. I cannot add additional reasons, as Cinaroot covers all I would say. Mburrell (talk) 16:29, 28 May 2025 (UTC)
• Support as complete commmon sense. EEng 16:33, 28 May 2025 (UTC)

• The present wording seems more appropriate and still allows for discretion. "Very likely" is better aligned with the intentions of the next step down. Nikkimaria (talk) 00:21, 29 May 2025 (UTC)
• Support "Almost certainly" means "always" (100%). That is not the case here. Even "very likely" (90%) is far too high. I would go with just "likely" (66%). Hawkeye7 (discuss) 03:43, 29 May 2025 (UTC)
• :Where are you getting those numbers from? Nikkimaria (talk) 04:11, 29 May 2025 (UTC)
• ::https://chatgpt.com/share/6837de9b-9c70-8001-b4a6-1b3aa06b87c6 Cinaroot (talk) 04:12, 29 May 2025 (UTC)
• :::What real source are you getting those numbers from? They don't appear to be universally shared interpretations. And as noted in the original revert, we should not set this level as at or lower than the next step down. Nikkimaria (talk) 04:20, 29 May 2025 (UTC)
• ::::I don't have a specific source. But like chatgpt says - its a commonly accepted numbers Cinaroot (talk) 05:22, 29 May 2025 (UTC)

• :::::Real citation needed. Nikkimaria (talk) 05:28, 29 May 2025 (UTC)
• ::::::https://www.theguardian.com/news/datablog/2015/aug/14/how-probable-is-probable?utm_source=chatgpt.com
• ::::::https://en.wikipedia.org/wiki/IPCC_Fifth_Assessment_Report?utm_source=chatgpt.com#Summary_for_Policymakers
• ::::::Chatgpt is not making it up. Cinaroot (talk) 05:36, 29 May 2025 (UTC)
• :::::::Your Guardian source assigns 'almost certain' a value of 93±6%, which is the range of your proposed new wording per your AI. So why would we change the wording at all? Nikkimaria (talk) 05:47, 29 May 2025 (UTC)
• ::::::::For a casual reader - almost certainly is like a near-100% probability ( id say 95-100) Cinaroot (talk) 05:50, 29 May 2025 (UTC)
• Oppose, all our policies are negotiable. Anyway, if this was a "hard rule", it'd say so rather than couching it. (And let's not base our policies on ChatGPT.) CMD (talk) 05:42, 29 May 2025 (UTC)
• Oppose {{tq|"Almost certainly" means "always" (100%)}} umm, no it doesn't. The word "almost" is there for a reason. Some1 (talk) 11:54, 29 May 2025 (UTC)
• Oppose for two reasons: that {{xt|almost certainly}} just isn't synonymous with {{xt|certainly}} as suggested above, and that {{xt|very likely}} allows for too much ambiguity for articles that in reality will require a compelling argument for why their exceptional length isn't a serious problem. Remsense ‥ 论 23:40, 29 May 2025 (UTC)
• :Their size is almost never a serious problem and seldom is something done about it. Hawkeye7 (discuss) 23:47, 29 May 2025 (UTC)
• ::The only compelling argument is that this guideline needs to be abolished. Hawkeye7 (discuss) 23:53, 29 May 2025 (UTC)
• :::I don't have the same sense that the impetus to trim articles below novella-length rarely happens and rarely helps when it does, but maybe we can agree to disagree there. I do think there are clear concrete problems with almost all articles when they get to this size, though. They are inherently less trivial to read and navigate, but the real problem is the size almost always results from severe focus problems, and since editors are less likely to be familiar with an article's structure the longer it gets, the more likely it is that these problems snowball. There's much less wrong with a 20k word article in a traditional encyclopedia, in other words. Remsense ‥ 论 00:09, 30 May 2025 (UTC)
• ::::Here is a conversation from my recent memory where an editor advocated for a split just because the article crossed 9,000 words. People are thinking - 15k words - we need to divide/trim.
• ::::Talk:Al Jazeera controversies#Split?
• ::::We did some trimming to satisfy that editor because of this policy. In my view, no trimming or split was needed. You can check the history and view a revision before the trimming began.
• ::::You can see how this is becoming a problem. I have encountered this in multiple instances. Cinaroot (talk) 00:33, 30 May 2025 (UTC)
• :::::I only see two editors in that discussion, so by "we", you mean you and FortunateSons. Some1 (talk) 00:43, 30 May 2025 (UTC)
• ::::I am always willing to agree to disagree. {{smiley}} There was a major disagreement over the guideline at Wikipedia:Good article reassessment/John von Neumann/1. The article was in the area of interest of multiple projects. Hawkeye7 (discuss) 01:21, 30 May 2025 (UTC)
• Support. What is the rationale for "almost certainly"? According to whom, or what? Are we talking about 2005 bit transfer rates? Browser limits? Server load? None of those apply now, and though I agree with opposers that "almost certainly" is not "certainly", there is no reasonable rationale for it. If I am a voracious reader and I want to read 33,000 words about Tartan, please raise your hand if you believe I must not be allowed to do so, and explain to me why I must be frustrated in my desire by some artificial limit imposed by others who wish to stop at 15,000 words, and declare that I must stop there, too? Who are they to tell me I cannot go on reading further? As a reminder: studies have shown that the majority of users never read past the lead. If that majority shows up here and votes to reduce the article split threshold to current max lead size, will you support that decision? If not, then why cut short the articles that I want to read past your comfort level? Leave me alone, let me read, and go find something else to do. Leave the long articles alone, *if* they are well written, and deserve to be that length. Apply yardsticks of quality, not quantity. A download of all 7M articles of Wikipedia fits on a thumb drive in my pocket smaller than my car key. Quit worrying about size; this is not the 20th century. Mathglot (talk) 08:51, 31 May 2025 (UTC)
• :Right, we do want to ultimately analyze {{xt|quality, not quantity}}. This guideline is a pragmatic one discussing quantity as a guiding heuristic of quality, given the realities of an internet encyclopedia where articles are assembled piecemeal and often with little supervision. The numbers are a bit arbitrary, but the point is to catalyze awareness and editorial attention concerning what information is most important and how it can be structured to be as accessible as possible. On that note, I strongly object to what seems to be a false dichotomy about readers: just because the lead is by far the most vital doesn't mean we should write only for the exact opposite diehard demographic. It seems quite unfair to spurn readers who may not know much about tartan (FWIW, this includes me, and I'm not yet willing to make a firm judgement on that article that I presume you linked as an example of a well-written, but very long article—I gave it a good-faith go, and it certainly is not obviously structurally flawed but naturally I don't know whether it could be better since I'm not an expert!) but may come away with little additional working knowledge plucked from the plentiful detail of that article. It may even ensure they join the majority of readers who don't get much farther than the lead.
• :In structure, these are fundamentally articles, not books! Books have pages, indices, and well-defined authorship, each of which ensure additional levels of coherency and intent in their composition. We do not have the luxury of confidence as if we had those things.
• :Ultimately, people who want to cut articles down have to take the initiative and demonstrate the cutting makes a better article. Maybe I haven't been party to enough of these discussions, but it really doesn't seem difficult to demonstrate the value of retaining content if the only opposing argument is handwringing about word count. This can be accomplished through a surface-level survey of what secondary and tertiary sources feel the need to include, which does not seem as much a burden as more detail-oriented discussions over NPOV often can be. Deletionist fanatics (some would label me as such) obviously have to do better than that. Remsense ‥ 论 09:30, 31 May 2025 (UTC)
• ::You're making it hard to disagree. {{smiley|sad}} Both the logic and the benefit of size limits have been challenged. (WP:Size considered harmful) {{tq|Sometimes an article simply needs to be big to give the subject adequate coverage.}} (WP:SIZE) Hawkeye7 (discuss) 10:39, 31 May 2025 (UTC)
• :The best way to deal with extremely long articles would be to spin sections off into separate articles and then use WP:SUMMARYSTYLE for the parent article. That way, information is still retained while the main article remains navigable. {{pb}} Take Department of Government Efficiency, for example. The article is currently at 8180 words. If it hadn't been split into multiple child articles, such as:
• :*Network of the Department of Government Efficiency (1818 words),
• :*US federal agencies targeted by DOGE (6798 words),
• :*Response to the Department of Government Efficiency (6854 words),
• :*Lawsuits involving the Department of Government Efficiency (2350 words),
• :the main article would be an unwieldy (and frankly unreadable) ~26,000-word mess. So if we truly want to focus on "{{xt|quality, not quantity}}", we should make these extremely long articles actually readable and navigable for our readers. Some1 (talk) 15:35, 31 May 2025 (UTC)
• Oppose watering down. A numerical target is generally useful to aim for. Too many articles suffer from excessive detail. There may be a handful of articles where such a massive length is not bloat, but not more than one in a thousand if that many. —Kusma (talk) 10:34, 31 May 2025 (UTC)
• : Our policy is that there is no such thing as excessive detail. Material can only be removed if it is unsourced, undue, untrue, libelous, patent nonsense or vandalism. (WP:DON'T PRESERVE) The recommended solution is to spin off subarticles. (WP:SPLITOUT) Hawkeye7 (discuss) 10:44, 31 May 2025 (UTC)
• ::I think procedural friction crops up when spinning off one article would seem to knock the parent article out of balance as regards the yet-unsplit material. There's a kneejerk impulse editors have (yours truly at least, but I have to imagine I'm not alone) where they can't endorse solving one problem in a way that creates another, even temporarily, even if that's just a part of the process sometimes. So these discussions can often assume a more binary, existential character than maybe they need to? Remsense ‥ 论 10:58, 31 May 2025 (UTC)
• :::They often do, and weirdly once an article is spun out, reversing that can become existential in the opposite direction. Another issue with spinout discussions occurs when they don't state what actual content can be spun off and how. Spinoff discussions about conceptual ideas don't really work. CMD (talk) 13:30, 31 May 2025 (UTC)