Not trying to call out anyone in particular, I see a lot of people doing this, but I'm really curious and would like to understand this better.
Folks who use vague alt text with your images ("a funny book cover", "photo of me outside"), what is your thought process? Is the image description intended for sighted people? Or maybe your admin made alt text mandatory? Or maybe that really does seem sufficient.
@stefan I thought about this a bit during a power outage yesterday. Posts trickled through but images didn’t load. The variability of alt text was interesting.
I think there’s a kind of captioning that assumes the reader can also see the image and just needs explanation, like a family album: “Mum and Dad in Brighton, 1974.”
@stefan Not sure if others have the same issue but I really struggle with some description levels. My mind knows what it wants to describe, but my writing sometimes may not rise up for the occasion. Description of the stuff was always boring for me until I discovered Terry Pratchett and TTRPGs.
I also found myself sometimes not sending a message because I didn't have enough spoons to create the proper ALT text it would deserve. Sometimes I prefer not to send at all than to write a crappy one.
@berniethewordsmith@stefan When I first came to Mastodon, the whole alt-text thing was new to me. I struggled with it, too, but saw the worth and tried my best. Now it’s second nature.
Practice makes perfect - The more you do it, the easier it becomes. And you gain a skill in the bargain 😊
@botvolution I've definitely noticed that on some automated accounts I follow.
I haven't really thought to check, but now I wonder if in those cases the people who set these up actually bother checking for it in the original post.
@Regezi @Baethyn @Stefan Bohacek @botvolution Another interesting question is: What if the image description is in English, but it has to transcribe bits of text in the image that are not in English?
Last year, I was in a situation in which I had to transcribe a whole lot of signs in an image. Some were in English. Some were in two or three of these languages: English, Germany, rather broken German, French. Others were in Latin.
When a sign was in English, broken German and French, I started with a 100% verbatim transcript of the English part. Then I wrote a 100% verbatim transcript of the German part, spelling mistakes, grammar errors and all. Then I added an English translation below that's as close to literal as possible. Then I wrote a 100% verbatim transcript of the French part. Then, again, I added an English translation below that's as close to literal as possible.
Of course, screen readers won't necessarily be able to switch languages in the middle of a description. But if 100% verbatim transcripts of everything are required, I give them, no matter what. And if people don't understand them, I give a near-literal translation in addition.
Sorry for not linking to the example, but the image description is hopelessly outdated and a bad example, and it'd probably be at least three times as long today.
@jupiter_rowland
Yes, this is a challenge and some participants are blind from birth, others have an increasing visual impairment (but can still fall back on memories)...Criticism is perhaps not expressed so much because Alt-Text is a delicate plant that should be constantly taken to heart by even more users...
Perhaps outlines (first the shorter ones with keywords, later the more detailed descriptive formulation) are useful? Do you have people with impairments in the Hubzilla network? @stefan
@Regezi Hubzilla doesn't have a culture of accessibility at all. It isn't even known which pages or "apps" are accessible in the first place. Besides, the huge majority of my readers are on Mastodon.
What I do is:
I write a fairly short, purely visual description for the alt-text. I do so to satisfy those who absolutely demand there be an image description in the alt-text, no matter if there's already one elsewhere.
Still in the alt-text, this short visual description is followed by a note that says that the actual, fully detailed image description with explanations, text transcripts etc. can be found in the post text. And if you're on Mastodon or a similar Twitter-mimicking microblogging project, the description is in the post text, hidden behind a long post content warning, whereas if you're on Friendica, Hubzilla or (streams), the image description will follow right below the image itself. Or sometimes, when I post multiple images, below the block of embedded images so as not to have a wall of text separate the images.
The alt-text practically always exceeds 800 characters regardless which, I guess, is still much too much for many Mastodon users, also seeing as Hubzilla does not support line feeds in alt-text.
And yes, the actual image description goes into the post itself where hardly any Mastodon user would expect it and where Mastodon's culture sees no place for image descriptions. The post itself is always hidden behing a summary/content warning that reads, "CW: long post (<four or five-digit number> characters" in order to protect those on the official Mastodon mobile app and everyone else on Mastodon who doesn't want to see posts with over 500 characters.
For example, here's my most recent image post which was three months ago. Fair warning: When I say I write image descriptions that are tens of thousands of characters long, I mean it.
At a bit over 25,000 characters, the description is nowhere near my longest, and while it's the most up-to-date, I found it lacking after posting it already, and I already consider it outdated. If this link opens as a Web page in a browser instead of as a toot in your Mastodon app, then sorry, I can't show you how your Mastodon app renders this post. I could tell you a hashtag under which to find that post, but troet.cafe doesn't know any of my image posts.
@jupiter_rowland
Thank you for your insight - with every conversation I learn more about Hubzilla & the misunderstandings within our large network 😊...a picture description is always from the point of view of the contributing viewer, which is why there will certainly be interpersonal blurring in an identical picture...
@stefan
I always add your address manually at the moment, because otherwise you won't be displayed in the reply bar...
Add comment