@Nat Oleander It's just a pity that if you post about an obscure enough niche topic like I do, it's the more difficult to make image posts perfectly accessible to everyone, the more obscure the topic is. For the more obscure the topic is, the more you have to describe, and the more you have to explain. (Caution: Never put explanations into alt-text! They must go where everyone can access them.)

I currently write the longest image descriptions in the whole Fediverse by a wide margin. But they may not actually be accessible enough, even though I describe all my original images twice.

The short descriptions in the alt-text don't always contain text transcripts, especially not all of them, and being only short descriptions, they aren't full, detailed visual descriptions either. The long descriptions for the same images in the post regularly end up with a five-digit character count. They may not be accessible because they're way too long. But sometimes they're the only place where all text transcripts can be found. And they are the only place where explanations can be found.

So the consequence should be that I quit posting my original images because they're impossible to make perfectly accessible to everyone, at least as long as there is no rock-solid definition for what's actually required in image descriptions in my obscure edge-case. But there isn't even any consensus on whether text that's illegible or that's so tiny that it's basically invisible must be transcribed if it can be sourced.

#Long #LongPost #CWLong #CWLongPost #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

@Georgiana Brummell I'd say not everyone would consider almost 1,000 characters of image description in a 1,500-character alt-text accessible. And even fewer people would consider a long image description in the post itself that's tens of thousands of characters long accessible if your screen reader spends an hour or several rambling it down.

My most extreme case is a post with only one image. That one image is described twice like all my halfway recent original images. The short description in the alt-text is a bit over 1,400 characters long which barely leaves any room for the note that there is also a long description in the post itself. That long description is over 60,000 characters long. I'm not kidding. It took me two full days from getting up to going to bed to research for it and write it.

Now, there are a few dozen bits and pieces of text all over the image. At the resolution at which I've posted the image, two of them are ever so barely legible for sighted people. They're on a large logo on a building. Four more, two of them on that logo, too, two more on a sign on an easel, are illegible, but still visible. At least more on signs inside the building are visible, but they can't easily be identified as text. All the others are so tiny that they're invisible. It takes the long image description to even know where they are, for example, on the control panels of teleporters.

And yet, they are all within the borders of the image. And I can transcribe them. I can't read them in the image, but I can go to the place shown in the image and take closer looks.

Unfortunately, the rule or guideline that any and all text in an image must be transcribed verbatim does not take into consideration text that can't be read in the image, but that can be sourced and thus transcribed by whoever posts the image. No confirmation, no exception. And so I have to assume that I have to transcribe illegible text as well. And so I do transcribe them all.

But there's no way for me to put all these text transcripts into the alt-text, not if I want to keep Mastodon, Misskey and their forks from chopping it off at the 1,500-character mark. I'd also have to explain where all these pieces of text are, after all. And so the text transcripts are only available in the 60,000-character monster of a long image description.

It isn't really accessible to expect blind users to have their screen readers ramble and ramble and ramble for hours, just to get information that should actually belong into the alt-text which, in turn, shouldn't be longer than 200 characters.

On the other hand, it doesn't really seem accessible to me if I expect people to ask me to describe things in the image for them. It rather feels sloppy, if not out-right ableist to not describe everything that someone could possibly want to know right away.

The problem with my images is that they're renderings from very obscure 3-D virtual worlds. This means that nobody knows what anything in these images looks like unless they can see these images. This, in turn, means that I cannot expect anyone to know what something in my images looks like anyway. They don't.

At the same time, I can't expect everyone to not care about my images. In fact, I expect the very topic of 3-D virtual worlds that actually exist to make people curious. At this point, it doesn't matter what's important in my images within the context of the post. Sighted people will go explore the new and unknown world by taking closer looks at all the big and small details in the image.

But blind or visually-impaired people may be just as curious. They may want the same chance to explore this new world by experiencing what's in that one image. Denying them the same chances as sighted people is ableist. But giving them this chance requires an absolutely titanic image description.

Sure, I describe lots of details which a sighted person can't possibly recognise when looking at the image, especially not at the resolution of the image as I've posted it. But I simply can't keep telling blind or visually-impaired people that certain things in the image can't be recognised due to the image resolution. It feels lazy, like weaseling out. I mean, I can see all these details. Not in the image, but where the image was made, simply by walking closer to them or moving the camera closer to them.

If there are two dark objects inside a building that may or may not be plants, but that can't be identified as plants by looking at the image, why shouldn't I describe them as follows: "On the sides of the teleport panel, there are two identical açaí palms in square terracotta pots with wide rims. Like the other potted plants, these mostly dark green plants with long pointy leaves are kept at an indoor-compatible size, namely about three and a half metres or eleven and a half feet tall. Also, like the other potted plants, they are made of only four flat surfaces with partially transparent pictures of the plant on them, arranged in angles of 45 degrees to one another."

If there's room for improvement in my image descriptions, I improve my future image descriptions and declare my past image descriptions outdated. In fact, the 60,000-character-long description is outdated because it's bad style to describe dimension using measures. Instead, dimensions should be described by comparing them with something everyone is familiar with like body parts.

Right now, by the way, I'm upping my game at describing avatars, using rules and guidelines for describing people which I've discovered over the last few months. The last time I've described an avatar, I've done so in about 7,000 characters, but according to my new discoveries, I may have missed something.

However, I can't go into so much detail while still making my image descriptions short enough that a screen reader can read through them in under a minute.

CC: @Nat Oleander

#Long #LongPost #CWLong #CWLongPost #Metaverse #VirtualWorlds #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta