Wikipedia talk:Manual of Style/Accessibility/Alternative text for images: Difference between revisions

 

Line 65: Line 65:

For what I consider near-perfect alt= text, see the adjacent humorous image. The alt= text for the image reads:{{tq2|A monochrome photo of poor quality. A woman is inserting a card in a time clock machine while a queue of women, about eight in frame, await their turns behind her. The woman’s other hand is on the lever that causes the machine to stamp the current time of day on the card. The women are dressed for factory work, wearing smocks and hair coverings. The machine’s clock reads 12:58.}} Please note that ”the alt= text contains no information except visual information.”

For what I consider near-perfect alt= text, see the adjacent humorous image. The alt= text for the image reads:{{tq2|A monochrome photo of poor quality. A woman is inserting a card in a time clock machine while a queue of women, about eight in frame, await their turns behind her. The woman’s other hand is on the lever that causes the machine to stamp the current time of day on the card. The women are dressed for factory work, wearing smocks and hair coverings. The machine’s clock reads 12:58.}} Please note that ”the alt= text contains no information except visual information.”

This paints a much more vivid mental picture than what Basics currently suggests. And the mental picture is the whole point, right? Alt= text speaks to the vision processing part of the brain, not the cognitive thinking part. Or something like that. 😀

This paints a much more vivid mental picture than what Basics currently suggests. And the mental picture is the whole point, right? Alt= text speaks to the vision processing part of the brain, not the cognitive thinking part. Or something like that. 😀

Autism spectrum probably makes it easier to write good alt= text, but I don’t think it’s a ”requirement”.

Autism spectrum probably makes it easier to write good alt= text, but I don’t think it’s a ”requirement”.

This page is within the scope of WikiProject Accessibility, a group of editors promoting better access for disabled or otherwise disadvantaged users. For more information, such as what you can do to help, see the main project page.AccessibilityWikipedia:WikiProject AccessibilityTemplate:WikiProject AccessibilityAccessibility

Please see Wikipedia talk:Linter#Wikipedia Mobile App: Image Recommendations and New Lint Error. This is just a cross-post after the message at Wikipedia talk:WikiProject Accessibility#Proposed tracking for images without alt text. Graham87 (talk) 15:53, 27 February 2024 (UTC)[reply]

The page should be rewritten; the Napoleon example at the start is contradicting the “avoid having the same details in both” advice in the “Basic” section. It would also help to collect best practices and guides from other websites. —Kusma (talk) 09:05, 22 May 2024 (UTC)[reply]

I agree, but a previous attempt to change this was reverted. Mike Christie (talk – contribs – library) 09:27, 22 May 2024 (UTC)[reply]

100% in agreement. The alt is not the same as the caption and yet throughout this article many of the examples are either the same or the alt merely says see caption. The Napoleon example at the beginning sets the tone for the contrary information throughout. I use this as a guide instead. It is much more clear. The alt is supposed to explain what is in the image to someone who cannot see it for whatever reason and why it is being used, i.e. function/context, for the article. SusunW (talk) 12:17, 22 May 2024 (UTC)[reply]

The redundancy was not there when the Napoleon image was originally introduced as an example, it was introduced in a “copyedit” that was never reverted. —Kusma (talk) 12:31, 22 May 2024 (UTC)[reply]

I have removed the “in his study in the Tuileries” from the ALT text, but the text still requires improvements beyond what I can do. —Kusma (talk) 12:35, 22 May 2024 (UTC)[reply]

In particular, should the ALT just be “painting”? —Kusma (talk) 13:09, 22 May 2024 (UTC)[reply]

I think that totally depends on the context of the article and our guideline should say that. To avoid redundancies, if it is simply to provide an image of Napoleon, painting would be fine. But, if it were being used to illustrate an 18th century/early 19th century French military uniform, one would explain what it looks like. Like you, I am probably not the person to rewrite the guide, but am happy to help. SusunW (talk) 13:18, 22 May 2024 (UTC)[reply]

Agree with you all this needs to be changed. Did the recommendation |alt=photograph , |alt=painting, or |alt=sculpture come from a particular (outside and reliable) guide? If not, I personally think we should remove it; I have never seen a guide on alt text mention this sort of solution, and don’t think we should be recommending it here unless it’s at least recommended as a standard somewhere reputable. Also, the example given with Napoleon at the top doesn’t follow any of the examples for using an image of Napolean lower on this page, under “Importance of context”. Maybe we can choose what context it should be under—lets say, the article on Napoleon himself—and pull from one of those alt texts instead? As WebAIM says in their alt guide, “Context is Everything”. Harvard’s recs on alt text emphasize context as well. From the previous discussion, WHATWG‘s HTML Living Standard specification on alt was linked, which I think also nicely sums guidance on alt text. – Whisperjanes (talk) 17:53, 22 February 2025 (UTC)[reply]

Just to emphasize further, I don’t think a single alt word of “painting” or “sculpture” would meet WHATWG’s recommendation of being an “appropriate replacement for the image”. – Whisperjanes (talk) 17:56, 22 February 2025 (UTC)[reply]

That is probably correct, but I think these words are better than “refer to caption”, which is an extremely poor alt text for three visually different toothbrushes. —Kusma (talk) 18:04, 22 February 2025 (UTC)[reply]

I agree. I think both of those areas of the page should be rewritten.

WebAIM says that if an image functions as a link, but already has a descriptive caption, then repeating info is the best option. Their exact words: The redundancy is necessary to adequately describe the function of the linked image, especially if it were to be accessed in isolation from the adjacent text, such as if a screen reader user were navigating by links. I think this is what we should be following.

How do you all feel about replacing the suggestion of the “Refer to caption” with two solutions:

1. For the toothbrush example, we suggest the editor describes the visual elements in more depth, since the current caption doesn’t adequately describe visual differences.
2. For other types of linked images with captions, adequetly but succinctly describe the image, even if info is repeated. This is for scenarios where the linked image is read in isolation from nearby text, such as by a screen reader.

And for linked decorative images, what are your thoughts on:

1. In the lead, replacing the examples of |alt=photograph , |alt=painting, or |alt=sculpture with instead briefly describing the image (e.g. “Squiggly line”, or a similarly-realistic scenario for when editors would use a purely decorative image). And mention linked images are functional images and therefore need descriptive alt text.
2. In the “Decorative images” section’s second bullet point, doing something similar as the above.
3. Add an example (somewhere) of how a screen reader would likely read a linked image with non-descriptive, single-word text. Like if we used “photograph”, a screenreader may read “Link, Image, Photograph”, which does not help a user know what they would be clicking on.

Obviously the above are not written-out fully, but wanted to get those thoughts down. – Whisperjanes (talk) 15:59, 6 April 2025 (UTC)[reply]

It says The blue (i) has alt text “About this image”., but the blue (i) was removed. I tried to restore it in [1] but was reverted as Miscalculation – no longer valid. @Moxy: What should be done to fix this? 174.138.212.166 (talk) 02:33, 3 August 2025 (UTC)[reply]

not sure what your asking …..you restored an edit from 2016 that was removed because the default image size has changed 2x since then. Am i missing something? Moxy🍁 20:42, 3 August 2025 (UTC)[reply]

I didn’t mean to change it, but I’m facing wikitext difficulties with the image size.

Currently, the bottom right corner is missing https://en.wikipedia.org/w/extensions/ImageMap/resources/desc-20.png?15600 . It needs to be there because it is mentioned by The blue (i) has alt text “About this image”. In my edit, I tried to fix that by removing |thumb.

But doing so causes the image to be enormous as |upright=0.75 is apparently ignored without |thumb. I tried to maintain the same image size using |160px. 174.138.212.166 (talk) 22:51, 3 August 2025 (UTC)[reply]

Discussion at Wikipedia:Village_pump_(technical)#How_to_make_|upright=0.75_work_without_|thumb? 174.138.218.72 (talk) 00:45, 5 August 2025 (UTC)[reply]

I went with Redrose64’s solution, which used |frameless to support the existing |upright=0.75. 174.138.218.72 (talk) 01:02, 6 August 2025 (UTC)[reply]

the section “Links and attribution” (MOS:EMPTYALT) should contain some guidance to editors on how to handle alt text with decorative images (for example, as used in header templates, where they are complementary to the title and body text). looking at the templates index, the guidelines are sometimes ignored (by adding alt="" to files not in the public domain, or no alt text being passed). this would be difficult to update myself as I lack template editor permissions, others should take over this step. Juwan (talk) 15:41, 21 September 2025 (UTC)[reply]

I somewhat strongly disagree with Basics. If I can get a consensus-in-principle here, then I can work on the details.

Alt= text is a substitute for vision, nothing more. If it’s not in the image, it doesn’t belong in the alt= text. In a photo of Donald Trump, there is nothing showing the man’s name. So the man’s name should not be in the alt= text. The name is in the caption.

Even if something did show Trump’s name, the alt= text would simply describe it. It still wouldn’t call the man “Donald Trump”. This is similar to the Wikipedia:No original research concept.

For what I consider near-perfect alt= text, see the adjacent humorous image. The alt= text for the image reads:

A monochrome photo of poor quality. A woman is inserting a card in a time clock machine while a queue of women, about eight in frame, await their turns behind her. The woman’s other hand is on the lever that causes the machine to stamp the current time of day on the card. The women are dressed for factory work, wearing smocks and hair coverings. The machine’s clock reads 12:58.

Please note that the alt= text contains no information except visual information.

This paints a much more vivid mental picture than what Basics currently suggests—starting with “Alternative text should be short […].” And the mental picture is the whole point, right? Alt= text speaks to the vision processing part of the brain, not the cognitive thinking part. Or something like that. 😀

Autism spectrum probably makes it easier to write good alt= text, but I don’t think it’s a requirement.

This would make my 2025. Comments eagerly anticipated! ―Mandruss ☎ IMO. 02:38, 7 October 2025 (UTC)[reply]

Agree …….what should change ? Moxy🍁 15:06, 9 October 2025 (UTC)[reply]

I don’t know, beyond a rewrite of Basics. Still waiting for that consensus-in-principle. ―Mandruss ☎ IMO. 16:37, 9 October 2025 (UTC)[reply]