screen readers are something else tbh. a friend was asking what full-width characters (Ｔｈｉｓ) sounds like in a screen reader, and i was curious myself

so I proceeded to struggle with finding a screen reader on my linux system and eventually settled on just using my phone, which has a built-in screen reader

and it at least didn't say anything weird like "fullwidth latin capital letter t fullwidth latin capital letter h" or anything, but it does spell it out, which isn't ideal either

one anecdote.

required alt text on images is a pretty big want with disability rights advocates on bluesky. a significant part of that want is the huge amount of textual information that is frequently embedded in images, but having some text to contextualize any image is desired.

there is a willingness in the community to use tagging and volunteers to reduce the friction of generating text, to provide threads that annotate images, and to work around a frequently unwilling platform provider. some refuse to interact with unannotated images.

still images are the only thing you can attach to bluesky posts, so that project could, in theory, address the totality of their alt text needs atm.

normative conflicts like this on social media are fascinating to me. I'm currently studying this exact thing for my phd

there's an inherent chilling effect to being told not to do something, which gets amplified by the way that message is presented. the tone, the phrasing, the number of people saying it, the proposed path to remediation, all impact the presence and degree of any chilling effects. and they're really important! even if that person goes to a space with different norms that allow their previous behavior, or the norms of their current space change, they still hesitate to engage in the same behavior that got their hand smacked the first time. once bitten, twice shy

this is really important if "the behavior" is something like "posting a fundraising link so I can make rent" or "calling out bigotry online". but even in lower stakes discussions like "engaging with this website that needs engagement to survive" or "putting lots of effort into posts that aren't accessible", if that behavior is subject to chilling effects then the website suffers

I agree with 95% of this, but I'm still going to describe my images instead of just punting on it for two reasons:

1. I do think vision-impaired people benefit from context in the cases where an image supplements text. If I was listening to a long-form post and the text said something vague, and then I heard "[image]", I would assume the image is important to understanding the text. The description of the image could still supplement the text even if the experience isn't the same as seeing the image. Images are used for reasons other than to showcase art and photography.

2. Writing image descriptions is actually kinda fun for me. Case in point. -- and, I think more to the point -- it becomes easier to do, and creates the habit of doing it, if you approach it as something whimsical or a writing challenge, rather than a chore. But then, I do enjoy writing challenges and not everyone does.

i do almost 100% of my cohost browsing - my primary social media place - in my screenreader. i think your point about yardsticks is good. i think it'd be better if we had better conversations, and if requests for accessibility and pushback against secondhand requests were both more civil. the app i use for everything on my phone does fine with the double struck text and the mathematical formula. formulae? functions. I like cohost compared to twitter because it works with my screen reader at all rather than basically not working at all. I like that there's a slow march towards accessibility, and that people care, even if they are clumsy. I think there's probably a curb-cut effect when it comes to making things clear in text, and it's easy to forget that ten years ago people would respond entirely in gifs. it's nice that that that's not the primary means of communicating, and I don't know if the trend back towards text is one driven by aesthetics or (accessibility) ethics, probably the former. I like thinking that people are trying their best, and I also think we could rethink our attitudes or approaches to alt-text.

So, I understand where you're coming from - most people can't be expected to learn and remember every quirk of screen readers. It's not a moral failing to not understand details of web development if you're just a user of a website. (I do think it is extremely unfortunate that the majority of web developers never learn about or implement accessibility. It gives my partner lots of job security, but... Dang.)

And you're right, "photo of my cat" isn't very useful as alt text. That's why any good guide for alt text talks about the context of an image.

For example, say someone posts a picture of their cat on top of a fridge, and in the body of a post says "he's up so high!" No alt text at all makes the post meaningless (who is up high? why is this posted?). "My cat" is somewhat better, but still kind of lacking. Where is the cat? But the alt text "A black cat peers down from on top of a fridge" fully describes the context as well as what makes it relatable and funny to post. It's enough info to include a person using a screen reader in the conversation.

And it's true many people talking about screen readers (at least on Cohost) don't have life experience with this. But there are blind folks who do talk about this stuff, as well as UX and accessibility professionals who work directly with disabled users, and it's worth searching them out (in that order).

I can't find the Mastodon post right now, but a screen reader user posted there that the alt text culture on Mastodon compared to Twitter made them cry because they felt so included. That's why I and other are so passionate about this. No, a user isn't a "bad person" if they don't use alt text. But the Internet would be a lot better if more people learned to write good alt text and used it, and if web developers implemented accessibility and made it easier for users to do their part.

(Also, on the mathematical symbols thing - using them as they were not designed on social media was annoying for a lot of screen reader users. But I don't think it's on screen reader software for being bad, because screen readers do handle math when given enough information to actually describe it in a human understandable fashion- mathml (video example included) with a screen reader sounds very different to the example you wrote out. There's a lot of work that goes into making stuff like educational materials accessible, and I don't think it's really fair to compare that to a screen reader encountering isolated characters on a social media website.)

I am genuinely trying to stay on an even keel about this, since I don't use a screen reader habitually and it's not in my lane, but it seems like at least every 2 weeks I have someone on some microblogging platform or other chiding me for not providing alt text, regardless of the following conditions:

1. that scrolling back on my feed would reveal that I do provide alt text on almost all images,
2. that the post itself functions as an image description and adding "alt: a photo of a black cat" to "check it out, my cat reached out and touched my phone screen and took a selfie [image]" would be redundant,
3. the person in question does not follow me, or appear to use screen readers or other assistive tools themselves, and 80% of their replies are finding people who didn't use alt text so they can scold them.

this specific combination of things is starting to get so far up my nose it's sticking out my ear.

From the perspective of building websites and features, I think an underrated issue is that the web makes it kind of easy to fuck up accessibility. There was a phrase I remember hearing a while ago that was something like “the web is accessible by default, we just broke it by trying to make applications instead of documents,” but like, people are going to make applications on the web, and I don’t think it’s desirable to not make them in 2023.

There’s a lot of features like comboboxes and other UI stuff that are pretty common, but not really in HTML by default, and when they do, they aren’t always customizable in the way people would like. The select element is a good example of this. ARIA exists to help bridge that gap, but it’s really low level and easy to mess up!

I think you talk about that feeling of “being afraid of being criticized for making something inaccessible” and while I think there’s a community component to that, I think a lot of it ultimately stems from the fact that the web makes it really easy to make something inaccessible and won’t tell you about it.

I dunno, I do a lot of accessibility work in my day job, and it’s rewarding. But it’s also hard, and not everyone has a full day and the resources to think through these issues and the tradeoffs they involve, especially if they’re just making a shitpost.

"drawing of a fox" - is this a deliberately bad example? like yeah, alt text that takes less than 3 seconds to think of and type out is not going to be of much use to anyone.

as far as i can tell there is no objective benchmark for "good alt text", because it has too many potential goals (blind accessibility, transcribing text better than OCR, being a placeholder for an image that's failing to load, providing context or attribution), some of which will conflict at some points. it is always a judgment call. i have always approached it like that - a "best effort" on the part of the poster to describe the visual parts of something that are describable, and also make sure it's easier to find from text processing operations like search. no two people are going to make the same choices there just as no two people are going to express a thought in words in the same way. because of that inherent subjectivity, i think people need to be patient with each other and with the overall practice, especially as it's still being mainstreamed. and yeah imo that involves not being so hardline you're just shitty to people and blaming them for stuff they don't understand but are likely very persuadable about, and also not acting (as we saw during some of mastodon's userbase growth spurts) like best effort is some egregious woke imposition.

i think comparing people requesting accomodations for their disabilities to someone who is mad you're not speaking the same language they do is kind of disrespectful, to be honest... that seems tasteless.

talking about the rest of the post, "not everything is for everyone" is a fair and true sentiment, but much of the time when people are asking for alt text it's for shit where people just... take a screenshot of some text, or post a response in an image that is easily summarized. sure, describing a cat doesn't enable someone to actually see the cat, but it's not as black and white as that. often times providing context with alt text is actually useful.

It'd be cool if I could just link a screen reader to people who don't get how all the alt text shit works, or spend a lot of time yelling about how it must hypothetically work, but lmao JAWS costs as much as industrial CAD software, which is why nobody who isn't legally required to is designing for it and the ones who are generally just do it by nixing images and any new feature that arrived on the internet after 1995.

anyway, from my experience using readers to assist people who weren't totally no-visual-perception blind, just too blind to read anything or navigate quickly, more or less you hover or tab over elements and it reads out whatever text it can find in the block, about as accurately as any other TTS program does, which means some will handle a given input just fine and others will suddenly start speaking in tongues. I never encountered whatever unskippable alt tag issue is at the center of the cohost furor and nobody will name what they're actually using so I can't speak to that, maybe in my version they hadn't patched that in yet, maybe this is a specific quirk of some terrible shovelware product one guy has ever actually used, it certainly didn't seem to be some kind of universal standard. There barely is a standard, there's just a collection of poorly designed products everything is mediated through. It's like using the internet back when every browser rendered pages completely differently, except IE costs $100/yr and instead of fucking up the page layout because it's two versions out of date from the release the page was tested on it just gives up and doesn't show that element.

I’m not powerful enough to attack the corporations I’m mad at. But this rando poster? Ooh, they’re deliciously vulnerable,

A thing that strikes me in the comments here is that people are having two totally different conversations based on what they've seen - some people who have only seen kinder, more genuine alt text requests who don't get your framing, and some people who have mostly seen the really aggressive demands.¹ Feels like the decentralized nature of this kind of site makes it hard for everyone to even be on the same page for a conversation.

1. As someone who's seen someone on this very site accuse people of being Nazi-adjacent and not thinking disabled people are human for not using alt text right... well, yeah.

the bit that gets me is that often the images i post are screenshots of text from a thing i'm linking to, and that could easily be OCRed with the device i am posting them with. but no social media platform i know of has OCR built in in order to fill in an alt text box.

You *made fun of the person you're subposting about *thoroughly enough, they've left the site. Idk if that was your intent but that was the impact of your difference of opinion about advocacy approach. Seems relevant to the conversation, I thought.

*Edits bc saying this isn't bullying is fair critique. Making fun of someone in five posts in a matter of hours, but never tagging or mentioning them, is not bullying, but is making fun of them. I'm typing this out sincerely, not sarcastically.

I have been writing up a reply about accessibility in game development that is now going to be a repost, cause it's about three paragraphs long and only increasing. Essentially creating any app with multiple systems, but especially games and social media, requires you to anticipate all use cases in the beginning, or you will spend magnitudes of time later on trying to duct tape solutions in later, to the point where it actually creates more problems for more users.

yeah i'm hoping to get some more direction eventually on what actually is helpful-- it's also an accessibility issue for me (person with chronic pain/frequent executive dysfunction) to be expected-or-else-i-am-designated-Bad to write alt text, video descriptions, audio transcriptions, etc. when they may or may not even be of use. sometimes i feel like i nail it with a description that would be useful and interesting to someone who can't see the thing; other times i have no idea how to describe, say, my pages of scribbly OC sketches of a bunch of characters nobody knows shit about and A Lot is going on and i'd have to write paragraphs to describe them aaaaaaaaaaaaaa.

i also don't want to assume what other disabled people are or are not interested in doing. i kinda hate when people assume for me what i can and can't engage with re: disability so i don't wanna just be like "well there's no way they would be interested in what i'm drawing, they can't see it". but i have a limited understanding of what i should be doing. the closest experience i have to using screen readers is maybe, like, using voice controls when my hands are super busted (which sucks, navigating anything without your hands is garbo). i think some of the energy directed at, say, yelling people for using special characters could probably be directed at screen reader software developers apparently not solving long-standing problems i've been hearing about for ages.

i hope i eventually hear from more actually blind people on this. there are some interesting replies here.

The funny thing is the macOS screen reader currently just skips anything that it doesn't see as being plain English text. So for example, in your doublestruck example, it says

sometimes i think about how people on twitter liked to have a good time, until there was a big uprising about it

and the equation just says "(something) equals two plus plus." I have no idea what the "(something)" is meant to be but I sure can't understand it. The "𝑓(𝑥)" on its own generates no sound, and "𝑓(𝑥) =" does the "(something) equals."

So yeah maybe screen readers aren't the accessibility win everyone thinks they are.

I feel you. I want to help folks but I also don’t feel like my alt text would help in any way? If you can’t view my image then would the alt text help? does “two people being romantic at each other” explain anything? And often trying to describe it better takes a lot of time and I’m unclear if it leaves an impact? How much description do folks need for it?

Even the most intense alt text I’ve done I’m unclear if I did it right, and in general I don’t think I could give this intricate a description for all my pieces. (The text was: “a digital portrait mimicking traditional oil portraits with intricate lace and fabric detail, of someone holding sword and flowers with nice lighting.”)

Givin' you the benefit of the doubt that you aren't bein' intentionally obtuse here...

but i still don't even know what good alt text is.

https://duckduckgo.com/?q=how+to+write+good+alt+text+for+images&ia=web

unless the alt text is some breathtaking prose (or the image is a comic with a joke that's funny on its own or something), it might as well say "art you can't see".

Would you tell that to a blind person who asked you what that thing they're passing by is?

instead the clear demand was for everyone — everyone — everyone — everyone — the entire human race to

provide ALT and non-formatted text in a convenient manner. For example: use a damn collapsible. This site makes it so much easier than tumblr.

and some guy complained because he couldn't read my tweets.

Being monolingual and blind are two very different things, even in this context. So you're either 1) saying ALT-less images are intended to be sighter-exclusive or 2) you believe not knowing a language is the same as not being able to read/see.

but not everything is for everyone. not everything can be for everyone.

so people who rely on screen readers can get fucked then? or are they just gonna "have to work harder"?

writing "photo of my cat" doesn't mean they can suddenly see the photo of my cat.

yes, and?
You showed the photo on your phone to the person next to you, but left out the person further away. So they asked "what is it?" and you answered "a photo of my cat" and now they're less excluded.
Did you seriously not understand this when you made your chost?

Is it safe to assume you no longer agree with yourself from 6 months ago? 'Cause damn.

i havent drawn a comic in a good while, but it is among the various projects i have constantly churning so ive thought about it a lot. the only conclusion ive come to is that its probably just best to make it clear with the alt text that its a comic, and then there will be a transcription if i can make it a similar quality to the comic itself. i dont worry about that too much, though i heavily sympathize with the brand of perfectionism youre referring to.

at the end of the day, i dont think comics are something that can be totally accessible? so my main concern is making it clear that its a comic so that folks who arent expecting something like that to be perfectly transcribed can just move on. in any case i know id rather put unfinished stuff out there than freeze and never put anything out at all, but thats just me and i suck at following through on that, i freeze a lot.

though i also. do like to try when im in a good mood. probably makes my posts inconsistent sometimes