shel
@shel

Sorry to use gravis here as an example but everyone needs to stop writing so much in the alt text field. It makes it impossible to see the full photo on mobile1. My phone doesn't even zoom in that far.

The best practice for image alt-text is to be as concise as possible. If you need to say more than one sentence about a photo then just do that in the body of the post and not the image alt text. That's been the web accessibility for screen readers best practice since the invention of the alt text HTML field.

Edit: to be more clear, the rule of thumb is "listening to an image description shouldn't take longer than looking at the image" so you're supposed to leave out conversational language, flourishes, etc. and be concise. "My white cat, one ear is black" instead of "This is my cat Cleo. She is white all over except for just one ear which is black."


  1. Which is the primary device of lots of people. As a non-techie I'm never using social media from a desktop PC. I only touch those at work.


shel
@shel

Please everyone just read this article from WCAG. I'm tired of arguing about this. Concise is good. It's been the standard for a reason. If you don't believe me and you don't believe WCAG, then please turn on the screen reader on your phone, close your eyes, blindfold yourself even, really really fucking keep those eyes completely closed no cheating, and try to use the web like this not just for a few minutes but for two days. Or do what I had to do for a month of every day at around 4pm switching to screen readers and closed eyes to prevent severe pain. For the rest of your day. You'll understand why concise is good.

I'm not trying to attack Gravis he's just doing what everyone else does. He's a great guy. This just happened to be a good example.

The image description should've been

Swarms of computer mice swim towards the Coollink logo like sperm to an egg. Text reads "The Best Connection Known to Man (OK, Second best)"


You must log in to comment.

in reply to @shel's post:

These are good image descriptions, including for sighted people. It's great when people put effort into writing them. We should not have to discourage great captions because of the temporary situation that Cohost's image viewer doesn't deal with them well.

I wouldn't want to promote the idea that accessible image descriptions are a tradeoff, that you have to strike a careful balance when using them. Let's just encourage all helpful captions. Cohost is an unfinished interface and its image viewer will get better at some point.

As someone who had to use a screen reader for a bit due to an acute condition I went through some years ago:

These are good descriptions of the image. They are not Good Image Descriptions when you are using a screen reader to navigate the web. You have to understand that when you're using a screen reader you have to sit and wait for a text to speech engine to read every single thing to you. You can't just quickly skim with your eyes. So everything that's repetitive or lengthy becomes extremely tedious. Even just a regularly recurring menu item being called "Make a post" instead of "post" becomes tedious especially if there's another button called "Make a message" or something so you have to pause and wait to hear which one it is. Most websites are extremely frustrating because nobody thinks about this. Everyone thinks people want an audiobook experience where a human being gently describes the website to you. That becomes tedious and frustrating quickly. This happened when audio books were first invented too actually. Sighted people made these long beautiful radioplay-like experiences and people would just turn the vinyl player up to 2.5× speed so they could read the book as quickly as a sighted person. This is why modern audiobooks don't sound like radioplays anymore.

When skimming through social media, I did not want every single image to start reading me an audiobook. I wanted to just be told "a white cat" keeping it just as brief as a sighted person experiences it.

The rule of thumb is this: Hearing an image description through your screen reader shouldn't take longer than it takes a sighted person to look at the image. Especially on social media where you want to know quickly if you're going to dismiss it and move on to the next post.

If it's a very complicated or dense image, then write the post body in a way that doesn't require staring at the image. Describe the important parts in the commentary. Honestly listening to someone describe the angle of the cat's head and stuff isn't actually that interesting at least it wasn't to me during my brief episode of needing to use a screen reader.

I was so grateful the opthalmologist treated the issue after only about a month of this because wow I was miserable.

What you’re describing here is the exact opposite of what Mastodon usually insists on, which… I haven’t actually read up on best-practices standards. I should see what they recommend now that I think about it.

Mastodon is often wrong :) Lots of people who aren't advocating for themselves or people they know just repeating things they heard.

I actually have a very storied personal medical history with Sight Issues I won't get into all the details of, but there was one incident of my Sight Issues manifesting that made me very passionate about this topic.

In my senior year of college I developed this new fun vision problem where by the late afternoon I just couldn't use my eyes anymore due to extreme pain, and I had a blind friend who was like "use a screen reader! here is how I do things" and basically taught me how to be a part-time blind person until I could see an ophthalmologist, and we bonded over how awful web accessibility is and how hard it is to use screen readers and how like, at the time, the android screen reader was literally unusable because the mandatory tutorial was broken. This lasted about a month before I got treatment.

It led me to incorporate web accessibility into my senior project, and doing a lot of research into accessible web design, and my passion about this is why I originally got involved with Mastodon development, I was making accessibility tweaks to make it better for people with sight disabilities and to work better with screen readers. This led to more web accessibility research and some involvement with WCAG.

And in library school we had a mandatory technology class where we had an entire module on WCAG.

And yeah it drives me crazy seeing people who don't use screen readers and have never used one demanding long fun eloquent passages of novels when all I fucking wanted when I was part-time blind was things to just say "cat dot jpeg" so I could move on.

Yeah, that's what I always figured - sorry, I should probably just have looked up a WCAG post on it before commenting, but my tone was meant to be more in agreement that concise seemed more likely to be best practice.

in reply to @shel's post:

I personally know someone that refuses to attempt alt text because they think it's too hard and won't try, and it annoys me greatly because I actually have used my phone eyes closed before and I get it now. It wasn't fun to scroll through social media and not have a clue what's going on in all these photo posts. (It was also surprising how many websites and apps have shit screen reader support...)

It feels like getting people to write good alt text is an unwinnable battle. Most likely they didn't try at all, or else it's way too long, or they misused the field to link to the image's source or something.

It sucks too because if they knew they're just supposed to be concise and brief it would be much less work for them and they wouldn't feel like it's too hard to do but people who don't understand what good alt text is keep telling people to be long with it which is bad for actually getting it

Because I haven't seen guidance on this anywhere: in your experience, when I post an image with lots of text is it useful for the alt to give both the short description "excerpt from Weber's Protestant Ethic", say, and "transcription in body"? Or just leave it at the excerpt part? I want people to know I'm not just going to leave them hanging, but if it's not useful I'll happily skip it.

When quoting a book or something, the best practice is simply don't use an image at all. Just use the ">" markdown indicator and put the text in the post body as text.

If for some reason you must have a picture, then identify only what is unique information in the picture. Something like "Illustrated page from Winnie the Pooh featuring Pooh and Piglet under a tree. Text transcription below"

I mean you can still do it it's just slightly annoying. To a screen reader user the experience would be the following:

"Image. Page from Weber's Protestant Ethic. Transcription below. Text. [the text from the image]"

If you put the text in the alt text, it poses compatibility issues and is processed by the technology as an image. It also creates accessibility issues for other people with sight issues that aren't using a screen reader (like myself 🙂) If you do both full text in alt text and transcript below, then a screen reader user will get the full entire text twice.

Which can be confusing! Your brain's like internal schematic of the document and where you are in it is thrown off a little like "wait didn't I just hear this did I swipe on the screen wrong?" And you can get lost which sucks.

Pinned Tags