@Fediverse

Again, this post goes out to both the #Threadiverse and the rest of the #Fediverse.

I've decided that only writing about my problems with #AltText and enormous #ImageDescriptions won't work as well as actually demonstrating what I mean, and why it's a problem.

Preamble: Some of you may see me on #Lemmy. But I'm not on Lemmy.

Others may see me in their local or federated timelines on #Mastodon. But I'm not on Mastodon either.

I'm on #Hubzilla (official website) which is part of the Fediverse and federated with Mastodon, Lemmy and just about everything else. It has almost unlimited possibilities. But while I can do a lot here, especially Mastodon is deliberately incapable of displaying most of it.

For example, I can write posts that are tens of thousands of characters long, and I can write alt-texts that are almost as long as the posts can be. But while Mastodon can still show posts from outside unshortened, no matter how long they are, it has a hard cap of only 1,500 characters for alt-text which, as far as I know, is applied to alt-texts on images in posts that come in from outside Mastodon as well.

Also, I can embed as many pictures as I want in Hubzilla posts, and I can actually embed them. I can place them wherever I want in-between the paragraphs. I don't necessarily have to put them at the end. Mastodon, on the other hand, knows pictures only as file attachments which it puts below a toot. And Mastodon toots can only have a maximum of four file attachments.

Lastly, I know that the vast majority of Mastodon users use Mastodon through a dedicated app on a mobile phone. Whenever they tap a link, it will open their Web browser. I also know that mobile Mastodon users tend to see their Web browser popping up as a nuisance, and they'd rather avoid to use their Web browser and experience the Fediverse in its entirety in their Mastodon app without anything else opening.

These are limiting factors, some of which will play a role in this demonstration.

Now, to get to the topic which I've already ranted about here and, most recently, here.

I'm stuck in a situation that's a combination of these factors:

One, the Fediverse demands I comply with its #accessibility requirements at the behest of #blind and #VisuallyImpaired users, otherwise I'll be sanctioned in some way. And I'm not the one to skimp on this. I'd rather try to satisfy everyone's needs. I'd rather have people tell me that what I've done is complete and utter overkill than that what I've done isn't sufficient.

Two, while some are satisfied with a short and concise alt-text, others ask for full descriptions of pictures with everything in them plus explanations for those who are unfamiliar with what's shown in the picture.

To give you an example, here is an actual Mastodon toot from a few weeks ago. I have re-shared this post a few times already, but I can't expect everyone who reads this post to have seen it before. I've used Hubzilla's own built-in standard re-sharing feature to automatically put it here into this post:

----------

Stormgren schrieb den folgenden Beitrag Mon, 03 Jul 2023 18:20:44 +0200

Alt-text doesn't just mean accessibility in terms of low -vision or no-vision end users.

Done right also means accessibility for people who might not know much about your image's subject matter either.

This is especially true for technical topic photos. By accurately describing what's in the picture, you give context to non-technical viewers, or newbies, as to exactly what they're looking at, and even describe how it works or why it matters.

#AltText is not just an alternate description to a visual medium, it's an enhancement for everyone if you do it right.

(So I can't find any prior post of mine on this, so if I've actually made this point before, well, you got to hear a version of it again.)

----------

In case you didn't get a link to the account this post came from and/or to the post itself, here is a link to the post.

Besides, just look through posts with the #AltText tag on them, and you'll see many with very elaborate and detailed descriptions, albeit often of not-so-detailed pictures, but still. So this is actually happening, yes. Not only that, but fully-detailed image descriptions are often actually praised rather than criticised.

Three, alt-text and #ImageDescription rules demand all text in a picture be transcribed in their entirety, word by word.

Four, I often post pictures that, taking the above into consideration, require very extensive descriptions because there's just about absolutely nothing in them which my audience is familiar with. My pictures are usually taken inside a virtual 3-D world based on #OpenSimulator because that's what this Hubzilla channel is mainly about. But out of probably over 13 million Fediverse users, maybe two or three dozen are familiar with #OpenSim worlds in general, and all the others aren't. And I can often hardly expect even three or four of them to be familiar with that particular place where I've taken the picture. Let's say these places are far from being as well-known and as not requiring description or explanation as Times Square, the Eiffel Tower or the Sydney Opera House. And if people don't know something, they need it described.

Five, I don't always post pictures like on Instagram or Pixelfed. That's when you make posts with pictures, and the posts are about the pictures. I sometimes use pictures as illustrations for posts which are not about these pictures specifically. In fact, these pictures are actually optional. Unlike in the former case, full image descriptions in the visual part of the post are bad style in this case.

So much about my situation.

What I'm going to do now is demonstrate multiple ways in which a certain picture that requires a very extensive description can be described in a post. None of them will be perfect. Each one of them will have its shortcomings which I'm sure will discriminate against someone out there.

The image in question can be found through this link. I have deliberately linked to the picture rather than embedded it here in order not to have to provide an alt-text that's sufficiently satisfying for everyone in this post already. The follow-up posts will be about describing this picture. Thus, they will all contain a description of the picture, and at least one of them is very likely to provide a full image description in the post body that should be accessible to everyone on every Fediverse project.

The image was first used in a post from over a year ago (link to the post) in which I've mentioned that the Metropolis Metaversum, one of the oldest OpenSim grids, has finally shut down after 14 years of operation, a few days later than scheduled. The picture shows my Metropolis avatar waving at the camera one last time before the grid, and the avatar with it, comes to its end.

Due to how detailed the picture is, due to how many objects with text on them are in the picture, and due to how almost absolutely nobody who may come across this picture will know anything in it, a full description at a detail level similar to describing a single bird in front of a blurry background plus explanations where explanations are necessary plus a full set of transcriptions can only be enormously long.

In the original post, the picture doesn't have an alt-text.

So what I'm going to do now is create multiple remakes of this post with the same wording and the same hashtags. But this time, I'm employing different techniques from remake to remake to include an alt-text and/or a full image description.

For this purpose, I've taken an image description which I've written several days ago, which already had 10,985 characters. I had actually gone in-world and visited a static memorial copy of the location shown in the picture to describe details that are practically invisible, but still theoretically visible in the picture. I've re-worked this description a bit and and expanded it even further. I've also found pictures of the big black sign behind the tree trunk and managed to transcribe it. As what's written on the panel turned out to be in German, I also had to provide a full translation. The only remaining writings within the scope of the picture that weren't transcribed are all on the Windows "screen" of the laptop on the counter of the info desk which is actually a static texture.

When combined into one paragraph, the description has 13,215 characters now.

The variants I'm going to post:

• Variant 1: short alt-text that only mentions what matters in the context of the post; no description given at all

• Variant 2: full image description in the alt-text

• Variant 3: short alt-text announces image description available through a link; full image description available on a separate page

• Variant 4: short alt-text announces image description; full image description in the text body of the post itself and fully visibly right away

As you will see, each one of them will have serious drawbacks for Mastodon users, for mobile users, for the people for whom we should all write alt-texts and image descriptions in the first place, sometimes for everyone.

#A11y #Inclusion #InclusionMatters #Inclusivity

@Fediverse

This is going out to both the #Threadiverse and, because I can't keep this from happening, the rest of the #Fediverse where I've mentioned this issue before three months earlier.

In brief: I'm still not sure how much #AltText is optimal. And I tend to run into situations in which alt-text that describes everything in a picture will grow longer than any of you could possibly imagine in their wildest dreams.

Here's my situation:

• I don't have a problem with writing a lot. Unlike most of you, I'm not on a phone. I'm on a desktop computer, and if I'm not, I'm on a laptop. I've always got a full-blown hardware keyboard, and I can touch-type with ten fingers. And I like to rant.

• I'm on #Hubzilla. This means virtually no limit in post length and especially virtually no limit in alt-text length. The only limiting factor would be how much alt-text the instances where my posts are viewed can display. #Mastodon has a hard cap at 1,500 characters, for example.

• I'm not the one to skimp on #accessibility rules unless they're technologically impossible for me to follow. I'd rather do too much than too little. This includes full transcriptions of all texts in a picture unless privacy issues speak against it, or unless I've got no way to source the original of a text anymore, and said text in the picture is ineligible even for me. Yes, I transcribe text that's one pixel high if I can get the original.

• When I post pictures, I don't always post them Instagram/Pixelfed-style, i.e. posts that are about this particular picture. Instead, I often use pictures to illustrate the post. Hubzilla gives me all necessary means to write full-blown blog posts with all bells and whistles as regular posts. Describing a picture in the visible part of a post when the post isn't about the picture is horribly bad style. Doing so when there are multiple pictures in one post, regardless of whether Mastodon puts them in the right places (which it doesn't), is even worse.

• I usually post pictures taken in #VirtualWorlds. In comparison with pictures taken in real-life, they have a much higher tendency to contain things that need to be described, often to both sighted and blind or visually-impaired users, because they simply don't know them, be it objects, be it locations. It's one thing if a picture was taken on Times Square, and it's something else if a picture was taken in a place of which maybe not even five people in the whole Fediverse even know that it exists. Thus, more text is needed.

Now there are two schools of thoughts when it comes to alt-text.

One: clear and concise alt-text. Only describe what's necessary in the context in which the picture is posted. Screen readers can't handle long alt-texts well. You can't navigate alt-text with most screen readers, i.e. you can't stop it somewhere, rewind it to a certain point and listen to parts of it once more. All you can do is let the screen reader rattle down the whole alt-text in one chunk. If you need to hear it again, you have to hear all of it again.

The obvious downside of this is that most of the content of the image is lost to everyone who isn't sighted, and some is lost to those who can't identify it even by looking at it in that particular picture.

Two: full description of absolutely everything in the picture plus explanation if necessary. Denying non-sighted people the chance to experience everything that's in a picture, and be it through words, can be considered ableist. Also, tiny details that are barely visible in the picture could be described so that sighted people can identify them.

And besides, there's the idea that alt-text can help everyone understand what that is that they see (or don't see) in that picture if they're unfamiliar with them.

As I've said, extensive image descriptions in the visible part of a post may be okay when the post is about the picture, but not when the picture illustrates the post and even less when there's more than one picture illustrating the post.

Yes, this is a thing. Just read what @Stormgren wrote earlier this month.

Stormgren wrote the following post Mon, 03 Jul 2023 18:20:44 +0200

Alt-text doesn't just mean accessibility in terms of low -vision or no-vision end users.

Done right also means accessibility for people who might not know much about your image's subject matter either.

This is especially true for technical topic photos. By accurately describing what's in the picture, you give context to non-technical viewers, or newbies, as to exactly what they're looking at, and even describe how it works or why it matters.

#AltText is not just an alternate description to a visual medium, it's an enhancement for everyone if you do it right.

(So I can't find any prior post of mine on this, so if I've actually made this point before, well, you got to hear a version of it again.)

And I'm actually waiting for Mastodon users to refuse to boost posts that contain pictures with insufficient alt-text. Many refuse to boost posts that contain pictures without alt-text already now.

The obvious downside of it is: "DESCRIBE ALL THE THINGS" + lots and lots and lots of stuff in the picture + just about everything needs to be explained because nobody is familiar with any of it = alt-text the size of a rather long blog post.

I've tried that with this picture (no embedding although I could because reasons). I've written a detailed alt-text. I've spent more than three hours in-world in a preserved, static copy of this place, researching and transcribing text where probably none of you would even know that there's text otherwise. The picture alone wasn't enough of a source for an alt-text that I would have deemed sufficient.

Only description plus some transcriptions: 7,636 characters. Description plus everything transcribed, save for the big black panel in the middle background behind the tree which I couldn't transcribe because it no longer exists in-world, plus translations of everything that isn't English plus everything unfamiliar explained: 10,985 characters. If that panel had still existed in-world, and I could have transcribed it, I might have passed the 12,000-character mark. With an image description.

As I've said, Hubzilla doesn't have a hard cap for alt-text length. In theory, it could handle and probably display alt-texts much longer than this. I don't know how it'd display an alt-text of that size in practice, whether it'd be scrollable, whether it'd have a time-out before anyone could read it fully etc. Mastodon, in the meantime, has the hard cap I've mentioned above which probably also cuts alt-texts coming in from outside. That's where most of my audience is. And screen reader users might have no other choice than to sit through their screen readers rambling down alt-text for more than five minutes in one go, especially if they could get a hold of the original alt-text instead of one cropped at the 1,500-character mark.

Now, even though I'll probably kick off two separate threads, I'd like to read your thoughts about how detailed alt-text should be.

#Accessibility #A11y #Inclusion #Inclusivity #InclusionMatters

@test

Headline 1

Headline 2

Headline 3

Headline 4

Headline 5

Headline 6

Centred

Highlighted

Colour

Multiple-linecode blockwith blank lines

@Fediverse

I have a question to all of you who have subscribed to Lemmy groups from Fediverse projects that aren't Lemmy. Who follow Lemmy groups from e.g. #Mastodon, #GlitchSoc, #Pleroma, #Akkoma, #MissKey, what's still called #CalcKey, #FoundKey, #Mitra, #GoToSocial, #Socialhome, #Friendica etc., but also #kbin. And I sincerely hope that I'm not the only subscriber to this entire group who isn't on Lemmy.

My question is: If a Lemmy post contains an image or any other media, can you see it?

I'm on #Hubzilla. And when someone posts something with an image in it, I can see the post, but I can't see the image. When someone posts something with a video in it, I can't see the video either.

I can see images in comments with no problem. But I can't see them in posts.

How about you on Mastodon? Or on CalcKey? Or on /kbin? Or elsewhere?

I'm asking and hoping for replies because I need to find out if the issue is on Lemmy's or on Hubzilla's side so I can file a bug report.