Please stop giving people shit for not writing image descriptions. Typing can be hard enough for some of us as it is, having to type twice as much or even more than the toot itself can be a hellish prospect and we don't/can't always use accessibility devices.

One thing that really sucks about this is there's an expectation that you disclose you have a disability in order for not writing descriptions to be considered 'valid'. I shouldn't have to tell anyone my hands are screwed to be left to post images without threat of being scolded.

Show thread

Lol I am a fan of assessableness features but expecting end users to always use them is stupid. I also think it's a little bit of trolling going on.

@sophia Agree, I already consider stupidly excessive when people write "selfie" and "eye contact".

@aurochs alt text is the little description when you hover over an image. CW is a content warning, it's a feature built into the platform.

People tag for eye contact because some users have conditions that make seeing even eye contact in images range from uncomfortable to panic inducing. Nobody has to tag for it but it's basically a thing to do just to be nice.

@revolverocelot @aurochs content warning tags also help people mute terms they don't wish to see

@revolverocelot @large_dad sorry, should have figured out the CW at least 😅
About the eye contact... I still think it's a bit ridiculous, even more considering that so many people have avatar pics WITH eye contact.

@aurochs @revolverocelot a "hide this avatar" feature would be a really good an useful thing to have

Project idea 

This sounds like a problem that can be solved with software.

Maybe someone with a bit more tensorflow/or similar experience than me, would like to write integration with mastodon.

Project idea 

@slaveriq Sounds like an interesting idea, I'd certainly like to see what could be done

Project idea 

@slaveriq @sophia This is getting dangerously close to full image recognition and description which is basically what google is still pumping all their money into while still having to combine it with captchas?

Project idea 

@fortunos @sophia

I think you can get pretty far. But yes in order to have to perfect descriptions you would have to do lots of research.
But if your goal is to make it easier to write descriptions by filling in some already, then you should be able to get results.

Project idea 

@slaveriq @sophia you could get item descriptions from the google API, possibly train some things like eye contact etc.

But situational descriptions and interaction between elements like that adds a galaxy of complexity. I'd be very interested in seeing someone attempt it and train it automatically on images that get uploaded right now but wouldn't bet on the end results

Project idea 

@slaveriq @sophia related: take a look at picdescbot which takes random images from wikimedia commons and lets microsoft cognitive services describe them. I have no idea if the source code is available

@sophia I like to do gentle reminders to add them occasionally but for people who are able. I hate when people are like IF YOU DONT U R BAD. Um no some people really can't

@sophia I have a friend who does speech to text and like, he can't do alt text at all. But it hurts too much to navigate to the alt text bar and type it in.

People just don't consider that maybe even some of the people screen readers help might also not be able to write alt text either.

@revolverocelot I don't know anyone with a screen reader who would begrudge someone who is unable to do it or it causes them pain, either. A lot of the aggressive stance seems to be on behalf of them, not them personally

@sophia ppl who don't use screen readers get hostile and say they are trying to help those who do, but end up bussing a lot of people who do (along with a whole bunch of other people) and it frustrates me.

I remember a thread some months back that basically claimed if you didn't do alt text you didn't care about accessibility and that's such a toxic stance to take

@revolverocelot I have spent years fighting for accessibility in a number of areas, and I'd never expect anyone to help me at the behest of their own well-being. I don't know anyone who would, it's better to look for pro active alternatives where they're needed

@sophia @aral
I'm asking myself if the description function of the pictures is cool for
Blind people however ?

@sophia that's one of the reasons i made OCRbot actually. the main reason was to remind people that captions are a thing and to hopefully encourage people to use them, but another reason was for people who can't provide the captions themselves to be able to automatically add them

@amsomniac OCRbot was actually done in about ~4 hours of total work, it's really not much. it's built on ideas other people have had and uses code written by others, too. all i did was throw it all together and host it on an instance with a large character limit

@lynnesbian it's done a hell of a lot of good on that front. I'd definitely use it for anything with clear text and really appreciate it being there, you've done an amazing thing with it

Think it would be cool if there was a way one could let another trusted individual get into the post and help in those regards...but I could see why that would be awkward

@sophia "I'm disabled and can't type. Let me type 2,000 words to explain why."

@mdhughes Because all disabilities are consistent in how they affect a person, always, across the board.

@sophia Ask yourself if you're doing unto others as you'd have done to you. If you're crippled, do you want ramps? If you're blind, do you want reasonable steps taken to explain image posts?

@mdhughes I'm not saying don't caption images? I'm saying don't be an ass to people who may struggle with it. If you're going to expend the energy to give them shit over it without knowing their circumstances, you'd have greater impact just adding a description in the comments and help instead.

I don't know anyone who uses screen readers who'd want people to feel they couldn't engage or were pained because of it.

@sophia i find CWs a usability concern for the same reason, and I wish they were at least opt-in rather than opt-out

@sophia I'm sorry you're getting shit for this, especially since you put up described images more consistently than the majority of posters.

@sophia I'm personally quite confused by the people who go around complaining. As you point out in your thread there are all sorts of valid reasons and there should be no pressure or need to disclose them. Generally talking about it and setting an example – if you can is so much more positive.
My approach is: I personally can and don't mind writing captions. I appreciate captions as sometimes they help me understand picture content better. Full. Stop.

@Lumb @sophia and also Lynne created the OCR bot which can be helpful if it’s a mostly text image.

@sophia this is one of my biggest problems with some of these communities of marginalized individuals. There's so much infighting it's easy to lose sight of the real goal

My hands are similarly affected. I usually use a voice typer, it works great on image descriptions as well.
Just as a tip - not to say you need to.

@sophia as always, things are complex and nuanced. those who say "caption yer dam images" are thinking they're doing it from a place of fighting for a good cause, to help the less able, BUT, like you've just said it's also hard for some people to type SO you're also making things harder for the less able. as always, it's best to listen to people who have these problems and boost their voices. I try to caption images but I actually get really stressed out by it sometimes due to my dyslexia and anxiety, so I just don't post an image rather than feel very anxious about my issue with not being able to write properly. a little empathy and thinking about why people don't caption images can go a long way

@CornishRepublicanArmy Yeah, there is a big balance with it. It's really about not making assumptions and throwing bad faith into the mix. If your you'd go to the effort to argue about it you may as well add it underneath and help instead

@sophia tbh it felt like from my perspective

Disabled ppl saying ‘ please try to write image descriptions’

And abled ppl with just a surface knowlged of ablism, acfesablity and completing asses needs decided that harassment/pressure is a good and just thing!

Rather hen *listebing* or thinking abt asexaboity beyond a chance to yell st ppl

@sophia I've been curious about so many images in the Fediverse having descriptions. Is it for accessibility?

oh yea this suck, a lot 

@sophia yea, i had a turn around on one mutual of not even 5 minutes because they screamed and hissed at me to add image descript

i asked them not to scream like this, i usually do put image descripts most of the time and didnt have energy this time

afaik they fuckin blocked me after like

i fucking do!! I FUCKIN DO!! put descripts on them, and i looked like a reply guy asshole who lies

i HATE IT!!!

anymore yellin in my mentions about it and ill probably stop doing it

@sophia I had not considered this, even though I'm one of the people whose hands don't work some days =X. Thank you for pointing it out

Yeah I don't scold people, I just add an #ImageDescription in the reply (when I have the time & energy, like you say, it isn't cost-free to do it)

@sophia I don't mind people asking, but I wouldn't want to be *ordered* to unless, obviously, it's a mod reminding me of their instance's particular rules.

@sophia I feel this. There are days when my vision is fucked and I need descriptions myself so I try hard to include them but other days my mental state is fucked and I can't form the words to describe things.

In my view, adding image descriptions is a courtesy and should be treated as such: A friendly gesture, not an obligation.

I try to add captions if I think they can get the point of the image across, but on top of taking time to type, for some funny pictures they'd end up explaining the joke, thus spoiling it, while also not being funny because visual jokes don't usually work as text.

The answer to that should not be to make everyone take twice as long to toot (but props to all who do!), but to find ways so they don't have to.

* There are OCR-bots out there: could something like that be integrated into a client, so I get the text when inserting an image?
* actually: speech recognition is quite okay these days, even running locally on a phone. Couldn't that also alleviate Sophia's primary problem quite a bit (in situations where it can be used)?

Sign in to participate in the conversation

sparkle sparkle, bitches