Wikipedia talk:Manual of Style/Accessibility

"WT:ACCESS" redirects here. For the related wikiproject talk page, see Wikipedia talk:WikiProject Accessibility.

"WT:COLOR" and "WT:COLOUR" redirect here. For WikiProject Color's talk page, see Wikipedia talk:WikiProject Color.

Discussion about accessible alternatives to font styling

A discussion at Template talk:Alphabet didn't began only incidentally as an accessibility discussion but that's become its focus, if anyone cares to participate. Largoplazo (talk) 23:27, 6 February 2026 (UTC)

Denoting a “winner” from an unordered list

In many cases where we have multiple lists on a page, each with multiple unordered candidates and one winner, is it enough to put the winner at the top of the list (and tell readers that we have done so)? For example:

Winners are listed first.

Category 1

• Winning candidate
  • Non-winning candidate 1
  • Non-winning candidate 2
  • Non-winning candidate 3
  • Non-winning candidate 4

Or should we be writing the word “winner” (or equivalent) to denote the winner inline?:

Category 1

• Winner: Winning candidate (winner)
  • Non-winning candidate 1
  • Non-winning candidate 2
  • Non-winning candidate 3
  • Non-winning candidate 4

(Two different options there; obviously we wouldn’t write winner twice.)

I am also aware many screen readers (I think the vast majority by market share?) will also read out the nested list. — HTGS (talk) 20:10, 7 February 2026 (UTC)

Yes, the first option is enough, but both work. Graham87 (talk) 01:27, 8 February 2026 (UTC)

MOS:LANG for names?

Does MOS:LANG imply that the names of people born in non-English-speaking countries be rendered in {{langr}} tags, e.g. {{langr|it|Enrico Fermi}}? Common sense tells me "no", since presumably there are literal millions of non-English names that are not currently wrapped in {{langr}}, and featured articles like Enrico Fermi don't seem to do this. But a plain reading of MOS:LANG seems to imply that we should. Any help is appreciated! Suriname0 (talk) 21:29, 6 March 2026 (UTC)

WP:WIP. Some aspects of Wikipedia lag more than others, MOS:LANG among them. Paradoctor (talk) 23:04, 6 March 2026 (UTC)

Yeah we probably shouldn't be doing that across the board. Whn reading aloud, a name like that wouldn't be spoken in a differen accent; Albert Einstein and Ludwig van Beethoven would be even more ridiculous examples. Graham87 (talk) 03:48, 7 March 2026 (UTC)

We should hear people's names pronounced in English, so I would say no. That would be like your pretentious friend who says "Darling, I spent a week in pah-ʁEE and another in mah-DɾEEDH." In the case where we're presenting the name as spoken in that other language, of course, it should be tagged. Largoplazo (talk) 04:38, 7 March 2026 (UTC)

In what cases are we "presenting the name as spoken in that other language"? I guess in non-English quotations that include the name? Regardless, this question was motivated by a dispute at Augusto Genina. It sounds like there's genuine lack of clarity here around the intended style – I'll leave it to other editors to sort out where and when {{langr}} should be applied. I have no idea if this is a wide-spread confusion or if it merits inserting a footnote in the MOS about lang templates for proper nouns. Suriname0 (talk) 05:37, 7 March 2026 (UTC)

Looks like MOS:DIACRITICS contains the closest thing to explicit guidance on this: "When non-English text should not be italicized, it can still be properly tagged by using the {{lang}} template with the |italic=unset parameter: {{lang|de|Nürnberg|italic=unset}}." But can is not should, which leaves it vague both when non-English proper nouns should not be italicized and whether {{langr}} should be used. Suriname0 (talk) 05:43, 7 March 2026 (UTC)

When we're writing about someone by name in English, we're writing in English, and it should be tagged as in English. If the name were written in another script than Roman then it would need tagging but we shouldn't do that outside the lead sentence anyway. —David Eppstein (talk) 06:34, 7 March 2026 (UTC)

MOS:LANG is about accessibility. An existing tag can be ignored, but a missing tag cannot be inferred by the user agent. Also, IETF language tags serve purposes other than screen readers. Automated translation comes to mind, for one.

Albert Einstein is not a valid counterexample, because he was an American. Same goes for Enrico Fermi. More generally, some foreign language terms have been swallowed whole by the English language, and are therefore English language terms. Loanwords are discussed in the guideline. Check the paragraph before last.

hear people's names pronounced in English That's handled by WP:PRONOUNCE.

should be tagged as in English Nope. English is the default. We tag non-English text, as it is stated right at the top: By default, English Wikipedia articles state explicitly to the browser that they are written wholly in English. Non-English text should be tagged as such Paradoctor (talk) 11:14, 7 March 2026 (UTC)

Einstein (1879–1955) didn't become American until 1940, which is to say, for only the last 20% of his life. He was Swiss for far longer than that. Similarly, Fermi (1901–1954), initially Italian, didn't become American until 1944. --Redrose64 🌹 (talk) 11:41, 7 March 2026 (UTC)

Missing the point. I mean, if you wish to make a case their names weren't borrowed in English, fine, just do that. But on the evidence available right now, their names are loanwords. That they had other citizenships is not in conflict with that.

And as I said, the guideline already covers this, regardless of which determination is made in the individual case. Paradoctor (talk) 12:00, 7 March 2026 (UTC)

Suggested clarification

Regarding alt text, this guideline now has "Images and icons that are not purely decorative should include an alt attribute that acts as a substitute for the image for blind readers, search-spiders, and other non-visual users. If additional alt text is added, it should be succinct or refer the reader to the caption or adjacent text." [bold emphasis added]. I'm very familiar with alt text, and have added it to many scores of images; but I was not able to understand what "additional alt text" in the second sentence meant ... it sounds like it is a second "alt" that supplements the "alt attribute" of the first sentence. I searched for that phrase in the Talk page archives and finally discovered a conversation from 2010 that explains it: the guidance is saying that the alt text should either be (a) a reference to the caption; (b) a reference to some body text; or (c) a succinct text ("additional" to the body text & caption). Does anyone object to clarifying the wording in the 2nd sentence to be: "Images and icons that are not purely decorative should include an alt attribute that acts as a substitute for the image for blind readers, search-spiders, and other non-visual users. If the alt attribute is used it should be succinct, or refer the reader to the caption or adjacent text." [changed words are in bold] Noleander (talk) 13:55, 5 April 2026 (UTC)

I boldly made the change in the guideline. Feel free to revert if it is wrong. Noleander (talk) 13:59, 5 April 2026 (UTC)

Should a sentence be added to "Images" section regarding Alt text?

Should a sentence about Alt text be added to the Images section of the Accessibility MOS? I ask because in a recent FA Talk page discussion several very experienced editors maintained that Alt text was not required for articles because the MOS did not give clear guidance on how to write Alt text. Comments from that discussion include (some comments referred to an RfC from 2019 which covered the same topic):

• "I have given up making use of the [Alt text] guideline, and simply add whatever text I think might be helpful. I'd much prefer to have a good guideline in place but a couple of attempts to improve it have failed, so now I just ignore it."
• " ... [in 2019] there was a fatal lack of clarity as to what ALT text should say."
• " I would appreciate guidance on whether what I've written is actually useful."
• "Have we figured out yet [since 2019] whether acceptable ALT text is that propounded by [names two somewhat conflicting Alt text guidelines from US and BBC]"
• "I recall also that [in 2019 RfC] there were contradictions between MOS:IMAGES, MOS:ACCESS#Images, and MOS:ALT; have those been resolved? If yes to all of them, then I'd support explicitly adding accessibility to the criteria."

In 2026, some editors still find the guidance confusing, possibly because it is spread across several locations: MOS:ACCIM in the Accessibility MOS; the Icons MOS; the Alt text essay; and Help:Alt text. Concrete guidance on how to write Alt text is in the essay and help page.

Editors looking in the "Images" section of MOS:ACCIM for Alt text guidance may come up empty-handed, because the two key statements are found in other places:

• Nutshell header: "Wikipedia pages should have ... alternative text so that content is easy to navigate and accessible for all readers, including those with disabilities." [no mention of "images"]
• Icons subsection: "Images and icons that are not purely decorative should include an alt attribute that acts as a substitute for the image for blind readers, search-spiders, and other non-visual users" [within "Icons" subsection, not in "Images" section]

Since many veteran editors are working under the impression that the MOS Alt text guidance is vague to the point of being worthless, perhaps we should add a sentence into the "Images" section of MOS:ACCIM that (a) plainly states Alt text is generally required; and (b) directs users (via body text, not a hatnote) to the essay and Help page. Thoughts? Noleander (talk) 23:08, 15 April 2026 (UTC)

I am conflicted. On the one hand, I absolutely believe we need to be supporting (pardon me if this is not the politically correct term d'jour) blind people, and that means we need to be writing ALT text. What makes me conflicted is that I honestly don't know what makes a good ALT text. I wrote some time ago I should note that there is disagreement about what makes a good alt text. Many people would look at the ones I write, point to various official recommendations, and say that I'm being way too verbose. That's fair, and you can make up your own mind what style makes sense to you, but don't allow "there's no clear consensus on the best way to do it" to become your excuse to not do it at all. But, a few weeks ago, I asked for guidance at WT:WPACCESS. We didn't get any. So it's difficult for me to continue to feel committed to this. RoySmith (talk) 23:48, 15 April 2026 (UTC)

I believe that the essay and the Help page have excellent guidance, with concrete examples. Especially the "Examples" section of the essay and the external links to Alt text guidelines in the Help page. My proposal is to prominently link to those concrete examples from the top-level MOS:ACCIM section. Noleander (talk) 23:52, 15 April 2026 (UTC)

Yeah, just adding the links would work. Graham87 (talk) 05:36, 16 April 2026 (UTC)

My experience is that the essay and help page are exactly like all other online sources I have found in being totally useless at telling me what to include in the alt text for mathematical diagrams (the primary form of image that I create and use here). Does it just say that it's a diagram and say something about the general topic, so that someone using a screenreader can tell what it is and move on? List the exact coordinates of every feature in the image, so that its mathematical content is accurately conveyed (as someone has done for the 120 points in the third image of No-three-in-line problem, fortunately omitting coordinates for the 7140 lines also shown in the image) but making it impossible for a human to get through it at all using a screenreader? Something in between?

The essay starts out telling me that alt text "conveys the same essential information as the image" and "ensures that no information or functionality is lost", often a total impossibility. The help page states that alt text "should convey meaning, rather than a literal description of the image itself", and goes on to give an example of a logo where the alt text is supposed to say that it is a logo rather than detailing its precise placement of features, a complete contradiction to the previous "same essential information" guidance.

Even for conventional artworks, it is very unclear and very context-dependent how much detail is appropriate. The essay shows a painting of Napoleon, recommending alt text saying only that it is a painting of Napoleon. The recommended alt text does not say that he is standing in a study next to a writing desk rather than a heroic figure on horseback, but those are very different images conveying very different things about Napoleon that surely screenreader users might want to know. Or might not want to know, if they are more interested in skipping past the "it's an image" information to the more-informative text. How are we to know and meet the needs of all of them simultaneously? I do believe in adding alt text to the diagrams in my lecture notes, or the photos of my cats in my social media, but in those cases I have a much more focused idea about who the audience is and what they might want to know.

Adding to all this is the issue that the vast majority of our images already have informative captions and it can be very difficult to find a description of the image that is not merely a repetition or abridgement of the caption, or a litany of the minutiae in the image. Much of the online guidance about alt text states that alt text can be optional when a caption provides the same information that would be in alt text (often true for our captions), but the proposed addition here does not allow any exceptions: it demands that all images have alt text even when others would say that some images do not need them.

Which is to say that I strongly believe that the 2019 "fatal lack of clarity as to what ALT text should say" continues to be as valid today. —David Eppstein (talk) 06:08, 16 April 2026 (UTC)

FWIW, I tried asking Claude to describe one of the images in the no-three-in-a-row article. This seemed like it would be an excellent tool for this job, but sadly the results were sub-optimal. Still, I could see LLMs having a place here. With an ALT text, the reader is stuck with whatever the author wrote. With an LLM, there's the ability to have some interaction, asking for clarification of something that's unclear. RoySmith (talk) 13:10, 16 April 2026 (UTC)

The first response from Claude is actually an accurate but overdetailed description of a different image from the article than the one you asked about, more or less copying the caption of the image but adding to it irrelevant details.

LLMs have a place in making alt text obsolete, because if LLM description were good enough we could count on screen readers using that instead of our own often-not-very-good prepackaged alt text. They would have the advantage of having a better idea where the reader comes from.

I also have some suspicion that some part of the push to alt text all of the things comes from a corporate push to grab conveniently labeled learning material for LLMs rather than from actually making content more accessible. —David Eppstein (talk) 16:35, 16 April 2026 (UTC)

I think the best course of action to resolve discrepancies is to work on MOS:ALT so that it gains consensus as a standalone MoS page, and standardise all other pages based on that. ~~ AirshipJungleman29 (talk) 11:24, 16 April 2026 (UTC)

I agree completely. Mike Christie (talk - contribs - library) 12:29, 16 April 2026 (UTC)

@AirshipJungleman29, @Mike Christie - The MOS:ALT essay may never gain strong consensus and become a full guideline. I realize that some editors feel the Alt text examples in the essay and help page are confusing or contradictory, but that is no reason to hide the examples from editors. Many editors may find the examples to be helpful for the specific images the editor is working with. Do you have any objections to adding links to MOS:ACCIM to help curious readers find concrete Alt text examples from notable, reliable institutions? Noleander (talk) 12:53, 16 April 2026 (UTC)

Noleander, MOS:ALT is the first link in MOS:ACCIM. I cannot see how the examples are being hidden. If MOS:ALT might not be able to become a full guideline, that is a good sign that it needs to be worked on. ~~ AirshipJungleman29 (talk) 12:57, 16 April 2026 (UTC)

The examples are hidden because the only two places alt text for images are mentioned in the guideline are (a) in the Nutshell way at the top of the guideline; and (b) in the "Icons" section, of all places. The guideline does not yet have a sentence in the Images section that plainly says:

"Alt text is generally required for most images. For guidance on how to write alt text, see this essay and this help page."

Simply because some (not all) editors feel the essay & help page are confusing is no reason to obfuscate the guideline. Our failure to provide that plain instruction is a real disservice to the blind users of WP. Noleander (talk) 13:10, 16 April 2026 (UTC)

Alright, I'll add a subsection. We'll see if it lasts. ~~ AirshipJungleman29 (talk) 13:33, 16 April 2026 (UTC)

Thanks for doing that. My intention here is not to solve the confusion surrounding Alt text. I'm simply trying to help readers of the guideline locate the examples (which are in the essay and help page). For example, the examples in the Harvard external link are very instructive. Editors that want to roll-up their sleeves and create Alt text will want to find those examples. Noleander (talk) 13:46, 16 April 2026 (UTC)

Ah, my mistake. I'll trim it down. ~~ AirshipJungleman29 (talk) 15:11, 16 April 2026 (UTC)

Thanks again for the improvements. Following-up on that, I just made two tweaks:

1. Moved your new sentence on Alt text to be adjacent to the existing two sentences on Alt text (all within the Images section).
2. Renamed the "Icons" subsection to "Alt text" because (a) That section contained only two sentences, and both were about Alt text; and (b) That subsection covered images and icons, not just icons. I'm not sure how it came to have the title "Icons".

No changes were made to any guidance or body text. Noleander (talk) 23:55, 16 April 2026 (UTC)

I don't really know what "good" Alt text is, but I think there is a problem with encouraging Alt text regardless of whether or not it is good. Editors should not just be adding any Alt text for the sake of having it. There are 122 Featured articles with the alt text "see caption", which is a completely useless alt, and potentially insensitive to visually impaired people (due to the word "see"). Plantdrew (talk) 16:53, 16 April 2026 (UTC)

Is it useless, though? If the caption adequately describes the image, to the point that another such description would be redundant or repetitive, shouldn't we just point to it somehow (using non-ablist language)? —David Eppstein (talk) 17:34, 16 April 2026 (UTC)

Screen reader apps will typically read both the caption and the alt text, if available. So in the case where the caption alone meets the needs of blind users – which is fairly common – the alt text can be omitted. Regarding the "see caption" phrase, that is not found in the Alt text guidelines. Rather, the guideline MOS:ALTINCAPTION recommends "Refer to caption". It also offers the option "refer to adjacent text" when the body text describes the image. Noleander (talk) 18:17, 16 April 2026 (UTC)

That seems like good advice that should be made more prominent. It could apply to many, maybe even most, of our images. —David Eppstein (talk) 20:40, 16 April 2026 (UTC)

Images and zooming

Hello!

Due to an eye issue, I use the zoom function on my browser to read and edit Wikipedia. Sometimes, articles do not respond well to the zoom. The most frequent problem is when images are placed next to tables, like in this article. The images "jump" above the table, which in this case creates a long stack of images on the right side of the page, with a vast area of empty white space on the left. It looks quite a mess.

In my view, this problem is reason enough to remove those images if they are unnecessary for the page, or to simply move them if they are necessary. However, I have encountered resistance to this view (most recently here) and I was hoping to seek input from others. Please let me know your thoughts. OrdinaryOtter (talk) 07:08, 25 April 2026 (UTC)

I see the same poor display (stacked images first, with whitespace on the left, followed by the table) happens on narrow screens. Therefore, I think it's general image-placement MOS issue, not an accessibility issue. I removed two of the images as purely decorative (at best) and outside of any reasonable value-added to the content. DMacks (talk) 07:33, 25 April 2026 (UTC)

I will make a talk page thread about it. What is the relevant part of the MOS? OrdinaryOtter (talk) 07:46, 25 April 2026 (UTC)

For a start, I've removed the width: style declarations from the table at Pichichi Trophy#Men, which seemed to be unnecessary; the table is now significantly narrower. Besides which, none of the other tables had column widths set. Because the images are in portrait format, I've also added the |upright option to them, this reduces their widths by 25%. Consequently, there is now space available between the table and those images for an increased zoom level before there is a collision. But please be aware that we can't code to suit everybody's situations all at once. --Redrose64 🌹 (talk) 10:52, 25 April 2026 (UTC)

Thanks for the reply, and for your help. The situation is better than it was. 😊 OrdinaryOtter (talk) 17:05, 25 April 2026 (UTC)

One follow-up question: Shouldn't we try to code in a way that accommodates all readers, even if we fall short some (or most) of the time? OrdinaryOtter (talk) 17:28, 25 April 2026 (UTC)

Discussion at WP:AINB §  Wendy at AgenticCommons

 You are invited to join the discussion at WP:AINB §  Wendy at AgenticCommons. Kowal2701 (talk, contribs) 18:06, 26 May 2026 (UTC)

Wikipedia talk:WikiProject Astronomy has an RfC

Wikipedia talk:WikiProject Astronomy has an RfC for possible consensus. A discussion is taking place. If you would like to participate in the discussion, you are invited to add your comments on the discussion page. Thank you. It involves colors proposed for assignment to an infobox. –LaundryPizza03 (dc̄) 01:57, 6 June 2026 (UTC)