Wikipedia talk:No original research

A big fish in a small pond

I've been thinking about the problem of describing quantitative facts in non-numeric terms. I'm thinking of something like this:

• Editors should use common sense when describing quantities in non-numeric terms; for example, a three-inch (8 cm) fish can be described as a "small" fish, and 3 m (10 ft) fish can be described a "large" fish, even though a three-inch coin would be considered large and 3 m telephone pole would be considered small. So long as editors agree that the description is fair and appropriate to the article, it is not a violation of this policy to use such descriptions.

What do you think? WhatamIdoing (talk) 21:10, 22 May 2026 (UTC)

Contextualisation is needed for everything subjective. Small and large are subjective opinions needing context. Small compared to what.

Facts, like 3cm or 3m, don’t require context to know what they mean. SmokeyJoe (talk) 00:30, 23 May 2026 (UTC)

I don’t think it is about editors agreeing. Most editing is one editor editing alone. What matters is whether the meaning is clear. SmokeyJoe (talk) 00:31, 23 May 2026 (UTC)

I was thinking about the opening line of WP:CALC: "Routine calculations do not count as original research, provided there is consensus among editors that the results of the calculations are correct, and a meaningful reflection of the sources." If editors agree that the description is a fair and meaningful reflection of the sources, then it's not OR. Compare, e.g., the difference in a sentence like this:

• "The bully pushed down the small child."
• "The bully, who was estimated to weigh 65 kilograms (143 lb) and be 160 centimetres (63 in), pushed down the child, who was estimated to weigh 35 kilograms (77 lb) and be 140 centimetres (55 in)."

Even if we have the exact numbers, we should instead provide an encyclopedic summary. The point is that the bully is bigger than the victim; the point is not exactly how tall or heavy they each are.

'Small' and 'large' aren't really opinions, in the same sense as "It's my opinion that coffee tastes bad" or "It's Barbie's opinion that math is hard". If someone posted online that three-inch fish are "large", or that three-meter-long fish are "small", you wouldn't think that this is an example of the wide diversity of human opinions on subjective and aesthetic matters. You'd probably think they were an idiot or (more likely online) a troll. WhatamIdoing (talk) 02:05, 23 May 2026 (UTC)

I don't see why WP:OR needs this, let alone with such examples - in context, a three-inch fish can be the large one. The two examples about a bully present a false dichotomy; as OP says The point is that the bully is bigger than the victim and there are all sorts of ways to express that, including "larger" or "smaller", and it depends on the context - the event we're describing and the way we're describing it - which ways work well. Do we really need WP:OR to state that qualifiers too must satisfy WP:BLUESKY or WP:CALC, or be sourced? NebY (talk) 15:36, 23 May 2026 (UTC)

A three-inch fish isn't a large fish compared to fish in general. It could be larger than usual for that species, or larger than expected.

I'm seeing editors put in inappropriately precise statistics instead of translating those words into a relevant description. For example, and editor proposed the other day that an article say that a decades-long gentrification effect happened in a place that had a 26.8% poverty rate. The place did not have a 26.8% at every moment during that time.

Why not just say it had a "high" poverty rate? (It's about double the regional rate, and of course the whole point of gentrification is that it happens in areas with high poverty rates.) Well, it occurred to me that one of the barriers to writing in an appropriate encyclopedic manner is rigid-minded editors who insist that if the exact word isn't in the source, then it's OR, unless the OR policy explicitly exempts it (in which case, they want to see book, chapter, and verse to prove that it really is exempted). This would provide good editors with a defense against the folks who think that if it's not plagiarism, then it's an OR violation. WhatamIdoing (talk) 17:45, 23 May 2026 (UTC)

A “small fish” is a pejorative. A “big fish” is subtle mocking. The addition of editor-chosen adjectives is usually acceptable OR, comparable to the spirit of WP:CALC. There is a possibility of embedding subtle POV into text. This should be recognised, and is to be expected, it is more than acceptable.

Avoiding adjectives, and using pointless measurements, is a bad answer to a bad question.

Editors should be presumed capable of adding description to quantitative facts, in their role in editing the text for easy reading of the source-based intended meaning. SmokeyJoe (talk) 05:41, 25 May 2026 (UTC)

I agree that Editors should be presumed capable of adding description to quantitative facts, in their role in editing the text for easy reading of the source-based intended meaning. What I'm concerned about is that some editors believe that this is actually banned by NOR. The obvious way to handle those beliefs is to have NOR say, in effect, "BTW, it's okay to describe quantitative facts by using words like 'small' or 'large' instead of copy/pasting exact numbers".

(If you don't like fish-based examples, then please suggest some others.) WhatamIdoing (talk) 02:20, 27 May 2026 (UTC)

I haven’t met editors who believe that NOR bans things. I don’t read NOR as being written in banning language. I read it as being full of things to think about. I agree that “it's okay to describe quantitative facts by using words like 'small' or 'large' instead of copy/pasting exact numbers", but would go further. It’s more than okay to describe quantitative facts by using words like 'small' or 'large'. In good prose, descriptive words are better used than exact numbers. Exact numbers are important, but unless the exact number is the point of comment, exact numbers are usually better contained in a table.

Fish-based examples are fine. SmokeyJoe (talk) 22:29, 27 May 2026 (UTC)

The policy includes the exact words "The prohibition against original research" twice. Obviously it is banning original research.

I would like editors, including editors for whom common-sense generalization is difficult, to be able to read this policy and agree that describing quantitative facts by using words like 'small' or 'large' instead of copy/pasting exact numbers is permitted. I think that the CALC standard:

• "Routine calculations do not count as original research, provided there is consensus among editors that the results of the calculations are correct, and a meaningful reflection of the sources."

is the appropriate standard. The equivalent statement might be something like

• "Describing quantitative facts by using words like 'small' or 'large' do not count as original research, provided there is consensus among the editors that the description is fair and a meaningful reflection of the source."

but I think that my suggestion above:

• "Editors should use common sense when describing quantities in non-numeric terms; for example, a three-inch (8 cm) fish can be described as a "small" fish, and 3 m (10 ft) fish can be described a "large" fish, even though a three-inch coin would be considered large and 3 m telephone pole would be considered small. So long as editors agree that the description is fair and appropriate to the article, it is not a violation of this policy to use such descriptions."

might help editors understand what we're really aiming for. With the CALC wording, someone might believe, e.g., that the fish's size is merely mentioned in passing, and therefore no description of its size (whether 'small' or 'three inches long") could ever be "a meaningful reflection of the source", even if they agreed that the description was fair. WhatamIdoing (talk) 21:21, 10 June 2026 (UTC)

That suggested wording rather demonstrates how your "common sense" may not match a reader's. I don't know what your grounds are for calling three-inch fishes small – among my sister's tropical fish, or my dentists', a three-inch fish is large. It's quite possible that the worldwide median size of a fish is less than three inches and even that the mean size is too.

You mention poverty above and justify your adjective "high" by comparison to the regional rate though that rate is not to be stated or sourced, and by claiming that if gentrification takes place, then the area must be poor. The latter reasoning is backwards; the former implies that the editor really does know what the regional rate is but isn't telling us. We shouldn't encourage either approach. NebY (talk) 22:00, 10 June 2026 (UTC)

Yes, among popular aquarium fish, a three-inch-long fish would be larger than is typical – for the tiny subset of fish that is aquarium fish. But among all the fish in the world, a fish that is smaller than your hand is obviously not a large fish. You would not see a three-inch fish caught by a sports fisherman and say "Wow, what a big fish!" (though you might say that for a preschool-age child). You would not sit down to a fish dinner, see a filet from a three-inch fish and say "What a big piece of fish I'm eating!" You would recognize that as small.

For the poverty example, the source says that the poverty rate is 26.8%. The source presents it as if that's high, without using that word. The source also provides a link to the US Census (which says that the poverty rate is no longer 26.8%, so saying "The poverty rate is 26.8%" would be a factually incorrect statement). It takes two clicks to get the equivalent US Census page for the state, which has a poverty rate of 13.4%, and the US as a whole, which has a poverty rate of 10.6%.

I think that saying 26.8% is higher than 10.6% (national) or 13.4% (state) is within the range of WP:CALC. I assume that you agree.

I also think – and I think this should be clearer in the page – that saying 26.8% is "high" is a fair way to present this, and that depending on the point that's being explained in the article, that it will sometimes make more sense to say "high poverty rate" than to say "poverty rate of 26.8%, compared to 10.6% for the whole nation and 13.4% for the whole state". WhatamIdoing (talk) 22:34, 10 June 2026 (UTC)

Source-based intended meaning? The editor must be confident in understanding the entirety of the source, including its reliability and POV. The smallest unit of meaning is at the level of a paragraph. Looking for meaning by fact or clause is going too deep. SmokeyJoe (talk) 05:44, 25 May 2026 (UTC)

I suspect that you're using a definition of meaning that is deeper than the community cares about. It doesn't take a whole paragraph for editors to grasp the meaning of "His birth date is 32 Octember". WhatamIdoing (talk) 02:23, 27 May 2026 (UTC)

Maybe, but I don’t think so. In the context of WP:NOR, I think it is very important to immediately distinguish between data and meaning. “The meaning of "His birth date is 32 Octember" is indeterminable. Is that typos, cryptic, or deliberate obscure expression? A birth date is a a fact, data. The meaning of a birthdate requires a paragraph to express what meaning an author gives it. A birthdate alone doesn’t have meaning. SmokeyJoe (talk) 22:37, 27 May 2026 (UTC)

Then infobox content is meaningless? WhatamIdoing (talk) 21:22, 10 June 2026 (UTC)

Often. Blueboar (talk) 23:28, 11 June 2026 (UTC)

• I am not sure we have a problem of writers stripping adjectives so, this seems like CREEP. Also, I would be wary of encouraging editors to analyze primary data -- it's different when "small" is a summary of what a secondary source says, as a secondary source can use context and other words to essentially indicate, small. Alanscottwalker (talk) 12:51, 11 June 2026 (UTC)

My thought would be to usually leave out subjective characterization type terms. They are usually a blend of information and opinion rather then just information. North8000 (talk) 14:01, 11 June 2026 (UTC)

So when the only thing the article needs to communicate is that this fish isn't very big, or that this poverty rate is high, what would you write? "The poverty rate is 26.8%", and hope the reader knows how that compares to the average? I doubt that's a fair assumption, especially for readers who don't live in that area/country. I've got no idea whatsoever whether 26.8% would be high, medium, or low poverty rate for any country at all in Africa or Asia, and I assume the converse is also true. WhatamIdoing (talk) 19:32, 11 June 2026 (UTC)

Routine calculations vs. reliably sourced data

I'm not sure if this has been discussed before, but I think the 'Routine calculations' section should explicitly state that reliably referenced values should take precedence over calculated results. That is, if a calculated result is listed in an article and then a reliably sourced value is found, the latter should normally replace the former. Yes, I've seen the opposite happen from an otherwise well-intentioned editor. Praemonitus (talk) 17:59, 6 June 2026 (UTC)

• I'm intrigued. How has it happened that a reliable source has come to a different total than an editor? Was the calculation based on disputed or approximate data, or on numbers that weren't comparable? Or are we just dealing with a volunteer editor who got their sums wrong? Context would be nice.—S Marshall T/C 18:57, 6 June 2026 (UTC)

For what it's worth, sometimes even reliable sources do make errors, and nothing in our policies should compel us to repeat such errors from sources if there is consensus against it. Sławomir Biały (talk) 19:14, 6 June 2026 (UTC)

Professional with training may have an understanding of the material that Wikipedia editors do not, simple calculation could fail to contextualise the data involved. Though it's very difficult to say anything definitive without some context. -- LCU ActivelyDisinterested «@» °∆t° 22:18, 6 June 2026 (UTC)

Another possibility is that the routine calculation could be based on newer sources than a source that contains the value directly. Jc3s5h (talk) 22:27, 6 June 2026 (UTC)

The particular context was the ellipticity of Jupiter's orbit around the Sun, which the editor computed (I believe) from the perihelion/aphelion, and came up with a slightly different value than the reference we use from NASA. I'm not clear why, and NASA didn't provide details on how theirs was derived. Perhaps it's a rounding error somewhere? Praemonitus (talk) 23:39, 6 June 2026 (UTC)

Fascinating. Of course NASA is definitely a reliable source for the orbital eccentricity of Jupiter. We may be dealing with a rounding error or a problem with observational frame (e.g. Jupiter doesn't orbit in quite the same plane as Earth so measurements taken from Earth might need compensating for). I'd certainly prefer NASA's figure in that situation.—S Marshall T/C 00:57, 7 June 2026 (UTC)

I would take NASA word for it when it comes to orbital mechanics. -- LCU ActivelyDisinterested «@» °∆t° 11:55, 7 June 2026 (UTC)

I'd imagine the ellipticity of one planet's orbit among the many interacting orbits of the solar system varies over time and that NASA's calculation would be appropriate to some moment or specific period. The values used in an independent calculation might or might not be similarly appropriate, we don't know and so we have to use the reliable source.

This isn't a case in which we can see the workings and at least understand an argument that the reliable sources are out of date and that a better value can be determined (e.g. Talk: Tesla (unit)#Conversion to gauss is incorrect). The possibility of such cases makes me doubt that we want to completely close the door by making an explicit statement as proposed. NebY (talk) 12:22, 7 June 2026 (UTC)

Well In that case I'd definitely expect to see a note of explanation, so that future editors can understand why it was calculated in lieu of using a referenced value. Just using calculation to replace a sourced value without providing a solid reason seems inexcusable. Praemonitus (talk) 13:19, 7 June 2026 (UTC)

Indeed. NebY (talk) 13:37, 7 June 2026 (UTC)

IMO it is editor's judgement which would be based mostly on the actual reliability of the source. This is different that wp:RS criteria. One common example is published articles about electrical power quite commonly have mistakes, generally with mixing up amount of rate of energy usage (e.g KW) with amount of energy (e.g. KWH). The sources have the trappings to meet wp:rs criteria but have so little knowledge that they they make mistakes that are obvious to WP editors. I would define actual reliability as expertise and objectivity with respect to the specific item which cited it. North8000 (talk) 14:12, 11 June 2026 (UTC)

Larry Sanger proposing WikiProject Intellectual Diversity

 You are invited to join the discussion at Wikipedia talk:WikiProject Council § Proposing a new WikiProject Intellectual Diversity. TarnishedPath^talk 06:40, 19 June 2026 (UTC)

"What is not original research > Original images" has 1 paragraph on topic and rest off-topic and duplicate

In the Original images" subsection of "What is not original research" the first paragraph is on point, but the remaining paragraphs seem to me a distracting noise as they don't have much to do with "What is not original research" (or the "WP:No original research", for that matter):

Suggestions:

Remove "Images of living persons..." all covered by Wikipedia:Biographies of living persons § Images.

Remove "Most images wholly generated by AI..." all covered by Wikipedia:Image use policy § AI-generated images.

Remove "It is not acceptable for an editor..." all covered by Wikipedia:Image use policy § Image content and selection (after my copying there) Zlamma (talk) 00:21, 20 June 2026 (UTC)

Please respond agree/disagree to give feedback. Zlamma (talk) 17:57, 27 June 2026 (UTC)

Generally, editors seem to prefer some duplication, so that if they look in one or two places, they'll find it. The downsides are that each individual page is longer, and that pages can drift into having conflicting rules.

As for your specific suggestions:

• "Images of living persons must not present the subject in a false or disparaging light" – OR is one of the ways to present a subject in a false light.
• "Most images wholly generated by AI should not be used in mainspace" – AI-generated images are a way to engage in OR.
• "It is not acceptable for an editor to use photo manipulation to distort the facts or position illustrated by an image" – Distorting information in an image frequently constitutes OR.

Therefore, all three sentences are all relevant to OR in user-created images. WhatamIdoing (talk) 06:04, 28 June 2026 (UTC)

Thanks for your thoughts. I now see something of what you say. But, while they are relevant to OR (top level), having them take over more than half of 'What Is Not OR' (sub level) subverts the purpose of the sub-level (loops the reader's flow to be thinking back on the top level). All other sections of WINOR seem to be happy with simple 'as long as', 'provided', 'while' but images seem to be trying to use this section for a list of image content policies. Maybe this because no single section in IUP addresses this, and policies are scattered around IUP, "Biographies of living persons" and MoS/Images, but I don't think WINOR is a good place to try to capture them all.

I have now attempted to boldly group (without changing meaning) all the 'what IS OR' statements to show what they are (this also allowed to spare any paragraph-entry clause cruft) and make each link to their authoritative source. Maybe this will further inspire some more conciseness in this section (e.g. moving this list to a single place, perhaps IUP, and linking there as a simple elaboration of "so long as they do not illustrate or introduce unpublished ideas or arguments"). Zlamma (talk) 10:21, 28 June 2026 (UTC)

Transliteration

There is currently a section on translation and transcription, and I believe it would be worth expanding this to include transliteration as well. As far as I am aware, transliteration is currently only mentioned in language-specific guidelines (e.g. MOS:ZH, MOS:AR, MOS:UKR, etc.), but it is usually more straightforward than either translation or transcription. Alongside "translating sourced material into English" and "transcribing spoken words from audio or video sources", I propose including "transliterating between writing systems" (or a similar wording), which I believe would codify existing practice. Freelance Intellectual (talk) 13:25, 8 July 2026 (UTC)

Are you seeing disputes around this? I'm not looking for examples that are just someone disagreeing about which transliteration system to use, or whether to bother transliterating a name, but someone saying that transliteration is the equivalent of making up something new. WhatamIdoing (talk) 23:31, 10 July 2026 (UTC)

Yes. Earlier this year, I was involved in this discussion about romanisations (which became extremely long). WP:TRANSCRIPTION was brought up by one editor, but since it didn't explicitly refer to transliteration, another editor disputed its relevance. The same editor cited support from an older discussion (that I was not involved in) where arguments were made such as "the MRT station name itself has no published Min Nan romanisation and therefore should not have one added to the infobox". I agree with you that the important question should be which transliterations to use (if any), and I believe it would be easier to focus on that question if there's clear policy that transliteration itself is not OR. In the quoted example, the question should have been whether a Min Nan romanisation is appropriate, not whether it is has been published. Freelance Intellectual (talk) 08:37, 13 July 2026 (UTC)

Thanks for the example. I've added this:

"Transliteration is not considered original research. See Wikipedia:Romanization for links to language-specific advice."

This will not help editors with the dispute over Mandarin vs other Chinese languages/dialects. WhatamIdoing (talk) 17:03, 13 July 2026 (UTC)

Thanks, that looks like very clear wording. I agree it wouldn't resolve that particular dispute, but it would avoid one way of getting sidetracked. Freelance Intellectual (talk) 20:43, 13 July 2026 (UTC)

Routine calculations > editors *may* show the work, and missing when they *should* do it

Regarding § Routine calculations' "In some cases, editors may show their work in a footnote."

This being a policy, informing that users may show their work, seems to not be adding valuable instruction, given that the paragraph already talks of permitting original calculations. (Or maybe I just haven't observed that the majority of disputes around routing calculations are protests to adding explanatory footnotes notes, but not the outcomes of calculations - please advise if this is the case.)

What seems to be missing though, is providing at least some criteria as to when they ought to do it.

In the interest of time, let me produce a concrete alternative:

Numeracy may be necessary to follow a "routine" calculation, particularly for articles on mathematics or in the hard sciences. In cases where the calculation is complex, error-prone, or likely to be contested, editors should accompany the result with information that will allow others to check for their accuracy (e.g. as an explanatory note).

... but treat it as just 'food for thought'. Generally, I did my best to preserve the spirit of the clause - not all calculations need be explained - but common-sense rules apply similar to those around WP:V: consensus-seeking WP:ONUS (when contested, or likely to be contested due a derivation being complex, thus risking a human error), with 'burden on introducer' WP:BURDEN.

If this version doesn't seem good but my point has merit, do propose a wording, which I suggest to at least contain one basis for users wanting to verify a result of calculations.

Context: The way I found this apparent omission in the policy is by trying to find a basis for the idea by a number of users to put some guidelines for uploaders of image and audio demonstrations of mathematical concepts and physical phenomena, to accompany these with material allowing to check for human error. (I beg you though 🙏, even if you oppose ideas in that past topic related to files, let's keep this topic small, and the files issue separate from this one. So, keep this about the calculations only, which the NOR readily identifies.) Zlamma (talk) 15:38, 10 July 2026 (UTC)

• Explain to me how this would work in practice, please. Take, for example, footnote #1 of 4D N = 1 global supersymmetry. What information could I give you that would enable you to check the accuracy of that?—S Marshall T/C 17:21, 10 July 2026 (UTC)

  Thanks for looking at this!

  • Re: Explain to me how this would work in practice. Note that I am not yet asking to accept my specific wording for the content of the proposed "should" prescription, other than just introducing one. I'll do my best to answer, but I will refrain from specifics because I am flexible about what the specific protocol should be (e.g. how much would just rely on author's awareness of error-proneness, and how much on consensus).
  • Re: What information could I give you that would enable you to check the accuracy of that? That's a fair challenge — I don't have a ready answer for that specific footnote, and that may be exactly the kind of case where describing an exception to the 'should' is also appropriate.

  Trying to use your example still, I'll need to adjust it slightly: I don't have time to study the particular subject (nor an interest in studying it), and hence I don't have any specific doubts about the accuracy. I also don't aim for this instruction to be the means for getting free lessons, but merely for providing a guideline consistent with WP:V for the cases where WP:V doesn't apply (because the problem isn't the lack of citations). Such a guideline seems missing. That's why, to use your example, I need to assume that "there are grounds for the accuracy to be doubted" (e.g. the work was error-prone) and that "the footnote showing the work is missing".

  With the above two assumptions, the way things could play out in this case is that a competent user requesting the work to be shown would now be able to point out a policy similar to WP:BURDEN, which contains a requirement or at least an encouragement of the author to provide evidence of computations. This could help resolve the issue quickly. Otherwise, in an extreme case, it could result in the author pointing at the *may* as evidence that the discretion is theirs, refusing to provide evidence, leading the keen verifier to disengage from Wikipedia, and thus cementing the number of accuracy checks performed on Wikipedia content at a rather low level.

  BTW, I tried looking at it from the other side, and while I appreciate that some freeloader-learners and cranks could abuse this rule, it doesn't strike me as a bigger problem than the one I'm describing. It doesn't seem to me harder to provide mathematical evidence for something that is true or to point at the necessary educational material, than to dispute someone who subjectively insists on not making error-checking on Wikipedia easier. At any rate, a little math proof job done once is reusable for debating any new doubts - once found in at least a Talk page, it can be pointed at, ending the discussion.

  With that said, I'd genuinely like to hear where you land on this. Are we aligned with the usefulness of introducing a *should*? If not, I'd be curious about why. Is it because of edge cases like the 'cranks' and 'freeloaders' (e.g. the problem is bigger than I imagine) or is there anything else I missed? Zlamma (talk) 10:39, 11 July 2026 (UTC)

  All right. Here's my concern with WP:CALC.

  • Mathematics is best explored in sequence. Any good maths teacher or lecturer will introduce you to concepts in a structured way, so some complexities are avoided until you have the tools to understand them.
  • Anyone could find our article on 4D N = 1 global supersymmetry. But you've got to digest and comprehend quite a lot of previous content before you can understand it.
  • [This next point may well be controversial.] My position is that if you do have the mathematics to understand that footnote, then it is, in fact, self-verifying. It's a fully self-contained and complete, logical, and inevitable proof of the fact or claim it supports. It doesn't need a citation.
  • And it's not just proofs, it's also worked examples. Mathematics articles really do benefit enormously from worked examples. Following through a worked example is a key part of grasping a complex mathematical concept.
  • Worked examples above a certain complexity will be copyright. Therefore it's policy that you can't copy/paste them from a textbook.
  • This means that according to S Marshall policy has to allow Wikipedians to generate worked examples that can't be cited.

  We really don't want to say that out loud on the policy page, though. A Wikipedian generating a novel calculation so as to prove a recondite and obscure point in a rarefied article on applied mathematics is one thing. A Wikipedian generating a novel calculation so as to prove their recondite and obscure point about how climate change is fake is quite another.

  In other words, while it would be possible to add a specific exemption to WP:NOR allowing formal proofs and worked examples for articles of postgraduate complexity that are within the scope of WikiProject Mathematics, I don't like the game theory of that at all.

  Therefore I prefer to keep the wording brief and vague.—S Marshall T/C 11:40, 11 July 2026 (UTC)

  

  Where's the citation verifying that these are red flowers?

  I think that the case of "policy has to allow Wikipedians to generate worked examples that can't be cited" is more obvious if you look outside of higher mathematics. For example: "red flowers" is a straightforward description of this image, but we have no reliable source for it. More complexly, years ago, I wrote a large illustration of using a grading rubric for standardized vs non-standardized grading in Standardized test#Use of rubrics for fairness. I doubt that we will ever find a reliable source to support it, but I hope that anyone who understands the sourced content of the Wikipedia article will be able to determine that the example given is accurately illustrating several of the main points. WhatamIdoing (talk) 18:37, 11 July 2026 (UTC)

  Although that's true, it might be less pertinent to Zlamma's question? He wants to think about when calculations should be cited.—S Marshall T/C 19:40, 11 July 2026 (UTC)

  Thanks, both. Oooh, I need to clarify quickly!

  This is not about requiring citations. It's purely about the "work" which the editor did to arrive at the result, and it's about [requiring/encouraging] it in cases where the editor has not provided it (my assumption that "the footnote showing the work is missing"), and its absence is making it difficult to check the calculation for accuracy (my assumption that "there are grounds for the accuracy to be doubted").

  To hopefully make it even clearer, in my example (modified version of S Marshall's) an acceptable solution to the contrived case of the editor not showing their work would be having them provide the footnote as it stands (no citations).

  Frankly, I am actually with you on making exceptions for the kind of information that S Marshall describes, but it pains me that there seems to be little in the policy that proves to the skeptics (who would rather insist on citations in every case), that the exemptions to requiring citations are controlled, and have proper boundaries that would prevent abuse by making it easier to reveal inaccuracy/dishonesty than to hide it (similar to WP:BURDEN, but for 'calc work', not citations). Case in point is the *may* in the sentence. Not only is it not even an encouragement, but some abusers could even say: "it's a 'may', not a 'should', so I don't need to show you the work behind my result". Zlamma (talk) 01:58, 12 July 2026 (UTC)

  I have a bit of a campaign against the word may in policies and guidelines. It's not clear to me whether this sentence is meant to communicate "In some cases, editors might show their work in a footnote" (it might happen; it might not) or "In some cases, editors are allowed to show their work in a footnote" (this behavior is not banned, even if you are putting something in the article that you can't cite a source for it).

  To give a non-mainspace example, the original version of this page views table didn't cite the data source (and was more or less wrong, or at least misleading, because it included redirects rather than just articles). Having a footnote is IMO better. I'm not sure I'd require it, but I'd definitely want to allow it. WhatamIdoing (talk) 02:47, 12 July 2026 (UTC)

  Re: It's not clear to me whether this sentence is meant to communicate "...might..." or "...are allowed..." Yes, that's precisely my point, though to me, "might" still feels almost as ambiguous as "may" (an impersonal "might happen" seems redundant in a policy). Currently I wonder whether it's

  • "the adding editor is allowed, but not obliged to show it" (so, if they are asked for it, they don't need to oblige)
  • or merely "the adding editor is allowed to show it in a footnote", full stop (so if they are asked to remove it, they don't need to oblige, BTW this feels mildly redundant because it is allowed under Wikipedia:Explanatory notes which even deals in more detail about how they too should not be overused).

  Seem's like we have just the man to clarify, since it was @S Marshall who added the sentence in his (overall a certain improvement) edit here (the talk topic referred to by the summary is likely this one).

  If my proposal to also add at least positive encouragements (but maybe even requirements, while trying to describe conditions that will keep the rule from being a nuisance) is not successful, then I feel there should at least be a change of verbiage that makes it clear which of the above two is meant.

  If my proposal is successful though, then I feel there's no need for any words that mean "the adding editor is allowed" because if there are cases when they are encouraged or even required, then that readily implies that they are allowed. Zlamma (talk) 11:15, 12 July 2026 (UTC)

  Revisiting that discussion, I still regret and deplore that we don't have an essay called WP:VERIFIABLEBYWHOM. I wrote "may", and I meant, "should consider whether to". In some cases it would be appropriate to show your working, and in others, not.

  For an example of where I feel that someone should show their working, consider the infobox of Pluto. Exactly as Zlamma wishes, it quotes the surface area of Pluto as 1.774443×10⁷ km² and then shows its working in a footnote. Thus, someone who wants to quibble the given figure could pop over to Talk:Pluto and moan that the calculation assumes that Pluto is a sphere when it's not: it's an oblate spheroid. (The resulting error is minuscule and it would be foolish to quibble the figure given. Naturally, only a few months ago on 26 March 2026, someone did.)

  For an example of where I feel that someone should not show their working, consider Cohen's kappa#Examples.

  Q: Can you see why not?—S Marshall T/C 13:10, 12 July 2026 (UTC)

  Re: I wrote "may", and I meant, "should consider whether to"

  Excellent, then we seem to be readily in an agreement on one essential part of this topic.

  In the interest of a more focused discussion about anything remaining (which I'm happy to continue), could we just make the minimal change to follow S Marshalls original intent:

  editors mayshould consider whether to show their work in a footnote.

  @S Marshall, @WhatamIdoing?

  (Feel free to propose different wordings, while trying to keep the scope equally small)
  Zlamma (talk) 16:34, 12 July 2026 (UTC)

• I don't love that. The extra words don't do much to clarify.—S Marshall T/C 17:42, 12 July 2026 (UTC)

  No worries. Do you agree in principle that *may* could use some clarification, given it didn't make us guess your intent? If so, perhaps you can propose the word(s)? Zlamma (talk) 18:45, 12 July 2026 (UTC)

  How about changing it to a statement of fact, e.g., "In some cases, it may be helpful to show your work in a footnote"? WhatamIdoing (talk) 19:10, 12 July 2026 (UTC)

  Not sure why, but now it's me who wishes that the sentence was at least clear about

  • 'Helpful to whom?' (I would hope for something that means 'other editors', or 'readers', and not just the editor introducing the calc result)
  • 'Helpful for what?' (I would love to see something like 'for checking accuracy', so that the editor doesn't only think 'for finding out more interesting details')

  I guess I worry that the reader won't guess these. Am I scope-creeping what I wanted to have a small scope now? Zlamma (talk) 19:35, 12 July 2026 (UTC)

  Zlamma and WAID will doubtless have seen this, but for record purposes, I note that this is when I edited the policy to avoid "may".—S Marshall T/C 20:04, 12 July 2026 (UTC)

  Thanks @S Marshall for the edit. I like "Consider showing your work." better than "It may be helpful.", in that it asks the editor for some effort.

  Now whether to add more changes can be discussed on each change's merit. Zlamma (talk) 21:12, 12 July 2026 (UTC)

Continuing - on Examples at Cohen's kappa

Good to have closed off part of the subjects.

Continuing one parked question at User:S Marshall's earlier response the other subjects:

Re:

For an example of where I feel that someone should not show their working, consider Cohen's kappa#Examples. Q: Can you see why not?

My guess is either of:

• The remaining computations are trivial. The formulas are given and each term can be discovered by just looking at an earlier definition. Providing them in a note would bloat the article's notes section.
• The reader benefits from going through the formulas, and finding the right place to take the numbers from. I found it especially helpful that I had to look back at the table to see the values in their locations, which would be lost if the lazy part of me followed the bait of a note that contains the numbers extracted from the table already.

Is that correct?

Overall, I agree that any extra footnotes there would be detrimental.

Zlamma (talk) 13:26, 13 July 2026 (UTC)

• I think those examples aren't trivial to work through and copy/pasting a calculation above a certain length and complexity is (a) plagiarism and (b) potentially a copyright issue. I'd want that calculation to have been done by a Wikipedian, as I hope it was.—S Marshall T/C 14:33, 13 July 2026 (UTC)

  I see. I didn't check what example values were in the source. Nevertheless, plagiarism is definitely the least subjective of the reasons, compared to the ones I gave. Zlamma (talk) 14:41, 13 July 2026 (UTC)

Continuing - Remaining changes to add a bit of specifics

So now, in § Routine calculations, after identifying one point of agreement above and changing "editors may show" to a language that asks editors for some effort, we have:

Consider showing your work in a footnote.

Now, what's left is to quiz our sentiments (other Wikipedians also welcome to chip in) about any more changes, adding specifics about circumstances when the user should do the 'consideration' or what the consideration should be about.

Continuing where we parked these, i.e. User:S Marshall's earlier response: For an example of where I feel that someone should show their working, consider the infobox of Pluto. ... the surface area of Pluto as 1.774443×10⁷ km² ... then shows its working in a footnote:

On that example, yes, we likely all have agreement that the notes are useful here. Earlier in the Cohen's Kappa example we also agreed that it's a cases when explanatory notes are detrimental.

So next, do we want to add a bit more specifics of "what" and "when" to consider, to make the policy side with the editors which are looking to achieve correct usage of the footnotes (have them supplied where useful, have them removed where not)?

In my mind we should add at least minimum specifics, and if we absolutely must minimize the policy's volume, then out of the two problems, and this may be controversial, the danger of the lack of helpful footnotes is the more detrimental problem to Wikipedia's quality (stifles the extent of accuracy checks over its content) in contrast to the dangers of the overzealous footnotes. The former problem is also harder to deal with (footnotes are harder to add then remove than to add). So, it seems most necessary to try to make editors feel responsible for adding the footnotes where helpful. This is also consistent with how WP:V does include WP:BURDEN, and generally devotes more space to encourage adding source information than it does to prevent overzealotry in the practice. Though, if the constraint on policy wording is not a factor here, I would not oppose if we managed to find a terse text that covers both sides.

A version of my old "food for thought only wording, changes welcome" adopted to the current is:

Consider showing your work in a footnote, especially in cases where the calculation is complex, error-prone, or likely to be contested.

Again, I did my best to preserve the spirit of the clause - not all calculations need be explained - but common-sense rules apply similar to those around WP:V and providing citations: consensus-seeking WP:ONUS (encouraged when contested, or likely to be contested due a derivation being complex, thus risking a human error), with 'burden on introducer' WP:BURDEN.

Again, if this version doesn't seem good but my point has merit, do propose a wording.

• I definitely don't want to add anything to the policy there. I've said "brief and vague". I would not object to a supplemental essay that explains this.—S Marshall T/C 14:35, 13 July 2026 (UTC)

@S Marshall

Re: I would not object to a supplemental essay that explains this

I also don't object to it, but can you clarify: if an explanatory essay was created (which would attempt to answer the whens and whats about the 'consideration', e.g. in the spirit of your message giving both sides to consider, and to which you would be welcome to contribute), would you even concede to linking from around the WP:CALC section? Zlamma (talk) 15:25, 15 July 2026 (UTC)

Depends on the essay, but I would certainly commit to considering it.—S Marshall T/C 15:54, 15 July 2026 (UTC)

Re: "I definitely don't want to add anything to the policy there. I've said "brief and vague"."

Thanks for stating. It will be good to hear other opinions.

BTW, the more concrete story behind my support for adding something here, is actually your idea to qualify into WP:CALC the audio files generated algorithmically to aural phenomena like beating or harmony.
In that thread, I think there was general consensus about writing a guideline that gives encouragement to adding information to the files, helping to check their accuracy. Sadly, the thread spiraled out and died despite my attempts at refocusing it.
But I really do like your idea to just look at WP:CALC to resolve such cases. We could treat it as the 'constitutional' basis for guidelines on algorithmically generated files (applying to any form, not just audio). But it lacks a WP:BURDEN-equivalent clause for results of more complex, error prone calculations which need more info than just the result, to check for its accuracy. So it's insufficient for reaching the full goal. As such, the files are still at a danger of being fought with WP:V (1 2, 3) on the basis of not qualifying for WP:CALC. Zlamma (talk) 15:33, 13 July 2026 (UTC)

I specifically do not want "checking for accuracy" to be listed as a good reason for providing formulas and other details from a calculation. I'm happy if it does provide a way to check a result, but if we say that's a reason, some editors will think that listing their formulas and calculations is a substitute for citing a proper source, and then the old Usenet personality may come back with his "How I Disproved Einstein" pseudoscience "because I can show all my calculations in footnotes, and you all can check them for accuracy". (This person was one of the main historical reasons that this OR policy was created). WhatamIdoing (talk) 17:10, 13 July 2026 (UTC)

@WhatamIdoing, thanks for clarifying. Just so I fully understand, are you also looking to preserve the lack of any clarification about what the user should be 'considering' (the 'vagueness', mentioned by S Marshall, is also intentional to you), so I should stop bothering you in a search for a compromise?

If not, then I observe that the paragraph readily starts with provisions both against the said personalities ("there is consensus that the results are ... a meaningful reflection of the sources") and in support of, perhaps more aptly named, 'correctness' (there is consensus that the results of the calculations are correct).

Maybe the way to address your concern would be to just connect to these provisions which, thanks to involving 'consensus', will be both a pseudoscience-deterrent and a help for those who try to positively influence the level of peer-review of Wikipedia:

• "Consider showing your work in a footnote in keeping with the consensus requirement that routine calculations be correct and reflect the sources."
• "Consider showing your work in a footnote to help other editors in providing the required consensuses."
• "Consider showing your work in a footnote to help achieve the consensus about the correctness and reflection of the sources."
• "Consider showing your work in a footnote to help achieve the consensus about the result being correct and a reflection of the sources."

(any better wordings welcome)

I am, after all, looking to introduce a restriction into a clause that is otherwise more permissive, so if this is done well, it can deter abuse, not attract it. Zlamma (talk) 09:54, 14 July 2026 (UTC)

"To increase reader understanding, or avoid disputes about the calculation, consider showing your work in a footnote"? This is how some policy guidance works in the pedia, 'do this to avoid issues'. Alanscottwalker (talk) 11:20, 14 July 2026 (UTC)

Footnotes do not help achieve consensus. If the goal is to help achieve consensus, then the calculation can be posted on the talk page. WhatamIdoing (talk) 16:27, 14 July 2026 (UTC)

Well, that's rather too categorical: both substantively and literally, what text is written and presented in the article (including footnotes) matters greatly in achieving consensus. Alanscottwalker (talk) 16:44, 14 July 2026 (UTC)

I'm with @Alanscottwalker on this ^^. Zlamma (talk) 20:33, 14 July 2026 (UTC)

@WhatamIdoing, while suspending my opinion on the assertion, can I ask you to also respond to my request to clarify your stance, i.e. whether you envision that adding some specifics about when an what to 'consider' could be an improvement, if a different wording was used (which you are welcome to propose), or whether you, like S Marshall, prefer to not add any specifics to the clause (which, by S Marshall's admission, is vague)? Zlamma (talk) 16:57, 14 July 2026 (UTC)

I do not think that adding specifics about what/when/why to include a footnote would be an improvement.

I think that adding a statement that claims a footnote will help editors achieve consensus would be either nonsensical (footnotes do not achieve consensus; humans do) or occasionally harmful (by misleading an editor into thinking that adding a footnote will make unwanted content by acceptable).

If this is not clear enough, then try this: Stop trying to add anything else to this sentence. Go read Wikipedia:Avoid instruction creep if you need more explanations. WhatamIdoing (talk) 19:29, 14 July 2026 (UTC)

Re: I do not think that adding specifics about what/when/why to include a footnote would be an improvement.

Thank you. This makes your opinion clear enough. Definitely was looking for this statement, as I understood that S Marshall is not open for finding a compromise wording, but wasn't sure about you, as you only raised wording objections.

Re: If this is not clear enough, then try this: ... Go read Wikipedia:Avoid instruction creep if you need more explanations.

No need for raising the voice. I certainly have already read that explanatory essay that isn't policy (including its § Not every instruction is creep"). Likewise, I also hope you are aware of the policy document Policies and guidelines and first rule under § Content:

• Be clear. Avoid esoteric or quasi-legal terms or dumbed-down language. Be plain, direct, unambiguous, and specific. Avoid platitudes and generalities.

I'm unable to square the above official Policy with a viewpoint that this admittedly vague clause should stay vague, especially in the light that an equivalent WP:BURDEN policy exists for citations.

Principally, I find the idea to deliberately create vagueness incongruent with a transparent, multi-generational, volunteer-run community project. Zlamma (talk) 20:32, 14 July 2026 (UTC)

Further reading and External links sections not exempt?

Seriously!? I really thought "Further reading" and "External links" sections would be safe from this policy. From what I learned, that may not be the case. How does simply adding it into a "Further reading" or an "External links" section exemplify "using" a source? I thought these sections are supposed to prevent sources from being used for original research within an article. George Ho (talk) 19:45, 13 July 2026 (UTC)

Context. The issue is that George wants to add links to articles that have nothing to do with the article subject but happen to use the two words "intellectual diversity" together. That's the OR here. It is not helpful to our readers to link to random stuff that has nothing to do with the article in the further reading section. voorts (talk/contributions) 20:14, 13 July 2026 (UTC)

Maybe we shouldn't use "OR" when we mean "off topic". Per the lead of Wikipedia:External links, one category of "acceptable external links" is "those that contain further research that is accurate and on-topic".

We probably need a page somewhere (more than Wikipedia:Scope first, then article title) that explains that just because a page is called "X" doesn't mean that everything that is called "X" belongs on that page. WhatamIdoing (talk) 21:37, 13 July 2026 (UTC)

Thanks. OR was the wrong term to use. I am not super familiar with EL, but now I know that's in there. voorts (talk/contributions) 21:44, 13 July 2026 (UTC)

Seconded, some editors really do not seem to understand that Wikipedia discusses a subject, not a shared title. Traumnovelle (talk) 22:00, 13 July 2026 (UTC)

We could put something about this in Wikipedia:Neutral point of view, as that's really where we talk about what belongs in/out of a given article. WhatamIdoing (talk) 19:30, 14 July 2026 (UTC)

Add "This policy is non-negotiable" to WP:OR for parity with equal importance WP:NPOV

Wikipedia:Neutral point of view says:

This policy is non-negotiable, and the principles upon which it is based cannot be superseded by other policies or guidelines, nor by editor consensus.

We should simply add that same text here to the lede. That is all. Any content in the enclycodpia must comply with both without exemption so there is no valid reason to omit it surely. — Very Polite Person (talk/contribs) 14:57, 27 July 2026 (UTC)

Oppose. WP:CGTW#8 – This seems like a weird roundabout way to weaken Wikipedia's requirement for the primacy of NPOV following on from discussions at WP:FT/N. In general a WP:POLICY is a "standard all users should normally follow" not a hard-and-fast rule, which is why WP:IAR exists for exceptional cases. If this proposal is actually an attempt to fix a real problem then could we have some actual examples of the problem that would be fixed, with some before/after examples of how this change would apply in practice? Bon courage (talk) 15:30, 27 July 2026 (UTC)

• Oppose as I did on WP:V. My opposition is of course non-negotiable on these issues that effectively weaken the primacy of WP:NPOV. Yesterday, all my dreams... (talk) 16:11, 27 July 2026 (UTC)