Jupiter Rowland

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Duothematic channel. Primary topic is virtual worlds/OpenSim, secondary topic is the Fediverse beyond Mastodon. This channel is NOT about real life!

Ort:: The Hypergrid

Geschlecht:: Male

Homepage:: https://hub.netzgemeinde.eu/profile/jupiter_rowland

Thinking about how I should handle long image descriptions in the future

2024-09-27 22:38:02

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Some of you may know, some may not, but I write extremely long image descriptions. My records are over 60,000 characters for one image and over 76,000 characters altogether for three images. They have to be that long, yes.

By the way, they go directly into the post. I write separate descriptions for alt-text.

(1/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

Fediverse Accessibility

show all 24 comments

2024-09-27 22:38:30

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

However, my long image descriptions inflate my posts to sizes absolutely unimaginable to most Fediverse users, especially Mastodon users. They may actually be a contributing factor to my image posts being largely ignored: Nobody wants to deal with such massive posts. And their size makes their placement an issue.

(2/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:38:48

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Again, they currently go into the post. I've tried placing each description under its image as well as combining all descriptions at the end. Both have their disadvantages.

Now I've taken putting them into external documents and linking to these into consideration. But that's far from perfect itself.

(3/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:39:01

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

A few weeks ago, I ran a vote on whether people really preferred meme explanations in the post itself, seeing as they tend to grow very long, too. None of the voters wanted them, at least not in the post.

So my idea was to put my long image descriptions into linked external documents. But that's far from perfect.

(4/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:39:19

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Option 1: Image description block at the end of the post

Post text part 1
Image 1
Post text part 2
Image 2
Post text part 3
Image 3
Post text part 4
Image description preamble
Description of image 1
Description of image 2
Description of image 3

This is what I currently do.

(5/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:39:35

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Advantages:

Image descriptions roughly in the same place as the images
Makes a preamble possible with common explanations; I don't have to explain everything multiple times
Image descriptions further down can refer to previous ones; I don't have to describe everything multiple times
Post stays in one piece

(6/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:39:53

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Disadvantages:

Image descriptions not exactly in the same place as the images
A lot of text between the last image and its description
Inflates the post

(7/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:40:08

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Option 2: Each description right below the image

Post text part 1
Image 1
Description of image 1
Post text part 2
Image 2
Description of image 2
Post text part 3
Image 3
Description of image 3
Post text part 4
Image description preamble

I've done that last year.

(8/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:40:22

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Advantage:

Image descriptions right where the images are

Disadvantages:

Inconvenient; cuts the post itself into fragments
Hard for image descriptions to refer to other descriptions; I may have to describe everything multiple times
No preamble possible; I may have to explain everything multiple times
Still inflates the post

(9/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:40:38

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Option 3: Link to image descriptions

I'd describe each image in an external document, either a Web page (only possible on Hubzilla) or a plain-text document or a simple HTML document in the respective channel's file space. Then I'd put the link to the corresponding description below each image.

This is entirely untested.

(10/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:40:49

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Advantages:

The post will be dramatically shorter. Not necessarily 500 characters or fewer, but not necessarily tens of thousands of characters either
The post stays in one piece
This works not only with posts, but also with unfederated articles

(11/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:41:06

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Disadvantages:

Untested. I don't know if it'll work, e.g. how which browser handles plain text
Inconvenient for users because the descriptions are not where the images are
Extra inconvenient for app users because the description will be in a different app than the image
I'd have to explain and describe everything again and again for each image

(12/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-27 22:41:19

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

I'm planning a series of posts with up to four images in them and post text in-between the images. And now I'm wondering how to handle the full, long, detailed image descriptions, also seeing as they will have to include a whole lot of explanations.

All three options have serious disadvantages, especially in this case.

(13/13)

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-28 01:10:43

Scott M. Stolz

scott@loves.tech

One option is:

Image (with alt tag)
Short description underneath in plain text.
Link to a page that has both the image and the longer description.

It is enough where everyone knows what the image is about, and those who want or need more explanation can follow the link.

2024-09-28 12:02:04

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@Scott M. Stolz Disadvantages:

That'd be three separate and different descriptions per image. One in the alt-text. One redundant one under the image with no added value. And the long description.
Remember that even my "short" image descriptions are generally over 1,000 characters. That'd be quite a chunk of text under the image.
Even a very bare basic explanation of what OpenSimulator is would be too long for a "short" below-the-image caption. And it would look stupid both under one image and not under the others and under all images.
Full-blown Web page only works on Hubzilla with the Webpages app. This excludes all portraits or other images with eye contact because Hubzilla cannot make Mastodon blank them out (eye contact can trigger neurodiverse people). (streams) can, and so I only post such images on (streams), but (streams) has no Webpages app.
The alternative, an HTML document in the file space, is theoretically possible on (streams) as well, but entirely untested in practice.
A plain text file won't work as a fallback because I can't embed an image in a plain text file.
Due to the embedded image in the external document, each link to one of these documents would likely require its own content warnings, e.g. eye contact.

It's basically option 3, but even worse.

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-28 14:43:57

Scott M. Stolz

scott@loves.tech

@Jupiter Rowland

That'd be three separate and different descriptions per image. One in the alt-text. One redundant one under the image with no added value. And the long description

Let me clarify:

Image
ALT tag ONLY if different than the short description.
Summary / Short description underneath in plain text. Less than 500 characters.
Link to a page that has both the image and the longer description (any description longer than 500 characters).

You could go as high as 1000 characters for the short description, but a summary should be short (500 characters or so). Anything above 1000 characters should be in the article.

For example, if the image you post is a meme that clearly has the text written in it, then the description could simply be "A meme about [topic] by [author name]." You don't have to repeat what the text says in the meme. But the alt text would need to be longer since it would need to explain what the meme is. The alt text and the short description would not be the same.

If the image does not have text in it, then it would be something like this:

[image] (alt text = "An image about [topic]."
"This is a 500 word description of the image, and I will summarize the whole thing in less than 500 characters. if it is longer than 500 characters, it goes in an article."
More details: https://example.com/articles/name-of-article-about-image

In the article, you include the whole description, as long as you want. Since it is on the web, it will find additional viewers. You can ask them to comment on the original social media post, and/or allow people to comment on the article.

Basically, when you have super long descriptions (over 500 to 1000 characters), you should treat it as if you are promoting a blog post on social media (rather than the whole thing only exists on social media).

2024-09-28 22:10:02

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@Scott M. Stolz First of all: If your target audience is mainly on Mastodon, and you don't want to lose any more reach than necessary, you have to play by Mastodon's rules, whether you want or not.

My target audience is mainly on Mastodon, and I've already lost too much reach.

This means three things among several others.

One, there must be a sufficiently useful image description in the alt-text, full stop, no exception. For at least some Mastodon users, it doesn't matter how many image descriptions you've got elsewhere. Especially not if you have to hide your post behind a summary/content warning because it exceeds 500 characters, and they'll see your image before your post because they're still on Mastodon 4.2 or older which keeps images outside of content warnings.

Believe me, I spend more than enough time lurking on mastodon.social to know.

Two, if you put a post with more than 500 characters out in the open with no content warning to hide it, droves of Mastodon users will block you. Do that often enough, and it'll dramatically cut into your reach on Mastodon. Remember, Mastodon is your target audience.

Three, you must protect people from potential triggers and sensitive images. On Mastodon, that isn't limited to porn and gore. It includes eyes. Eyes can trigger autists and other neurodiverse people. Not only in facial portraits of people looking straight into the camera, by the way.

Point three, by the way, means you can't safely post pictures with faces or eyes in them from Hubzilla to Mastodon. Hubzilla has no way of making Mastodon blank sensitive images out.

A summary/content warning only helps on Mastodon 4.3 which hides images under content warnings. But Mastodon 4.2 or older, which means one solid half of the whole Fediverse, only hides the post text under a content warning and leaves the images dangling below the hidden post. And when a post comes from outside, the images are not blanked out. They remain in everyone's plain sight.

My meme images almost always have faces in them. The OpenSim in-world pictures I want to post are likely to have faces in them, too. So posting them on Hubzilla is out of question.

I've launched two (streams) channels for these while you were away:

@Jupiter Rowland's (streams) outlet for OpenSim pictures
@Jupiter's Fedi-Memes on (streams) for memes; this one doesn't do CWs for long posts because it isn't for the faint of heart, and everything's hidden behind CWs for eye contact etc. anyway

(streams) automatically sets Mastodon's proprietary sensitive flag on images if they're included in a post that contains the hashtag #nsfw or #sensitive. Hubzilla doesn't have this feature.

However, (streams) neither has the Articles app, nor does it have the Webpages app. If I wanted to use either, I'd have to post images on (streams) and then link to an article or a webpage on this very Hubzilla channel.

Now to some of your specific points.

Image
ALT tag ONLY if different than the short description.

Again, alt-text always. Unless you don't care for your audience on Mastodon. Again, I do.

So the short description under the image is redundant.

You don't have to repeat what the text says in the meme.

The transcript of the meme text belongs into the alt-text. Any text transcripts belong into the alt-text.

The only exception in my case are if there's too much text, and it doesn't fit into the alt-text. In this case, the transcripts go into the long description, and the alt-text mentions that the long description contains all text transcripts.

But the alt text would need to be longer since it would need to explain what the meme is. The alt text and the short description would not be the same.

Alt-text must never explain anything. No information must ever be exclusively available in alt-text.

That's because not all people can access alt-text.

Certain frontends don't support alt-text.
People with a severe tremor can't operate digital devices with a steady enough hand to be able to access alt-text.
Alt-text requires handling pointing devices. This means that quadriplegic people who use their computer by poking the keyboard with a headpointer strapped to their forehead or with a kind of pen that they hold in their mouth can't access alt-text either. They operate their computer only by keyboard.
And so on.

This means that explanations must be in a place that can be accessed by anyone who can access the post itself. This applies to anything that's part of the post text body, for example.

[image] (alt text = "An image about [topic]."

Nitpick, maybe, but starting alt-text with "image about", "picture about" or "photo about" = baaaaaaaaaaaad style. Other media are at least tolerated, but "image", "picture" and "photo" are deadly sins.

Basically, when you have super long descriptions (over 500 to 1000 characters), you should treat it as if you are promoting a blog post on social media (rather than the whole thing only exists on social media).

My image descriptions don't make for good blog posts.

I have put image descriptions into articles twice already:

demonstrational description of an OpenSim in-world picture, early 2023, highly outdated
description of a UI window from a Second Life and OpenSim viewer, 2023

You can see that they don't work as stand-alones that I could show random strangers on the Web.

If I were to do it again, this would be the style, only with a more up-to-date description.

Speaking of which, let me show you how I currently actually handle them.

First of all, some meme posts.

My first attempt at an updated meme post was One Does Not Simply Implement FEP-ef61 in early August.

Back then, I was still convinced that people find links to external information sources cumbersome and inconvenient to the point of being not accessible and uninclusive, and that they wanted everything explained in the post.

So I did. And I ended up with a whopping nine explanations:

what the image itself is about, namely the teething problems of the implementation of decentralised IDs as per FEP-ef61 as one of the first steps of adding nomadic identity via ActivityPub to the streams repository
the "One Does Not Simply Walk Into Mordor" meme template
snowclones
image macros
advice animals
Something Awful
4chan and, generally, imageboards
nomadic identity
Hubzilla, (streams) and their whole history from Mistpark to (streams)

All so that everyone in the Fediverse would easily have a chance to understand the meme post.

However, these nine explanations are over 25,000 characters long. The last two explanations are half of it.

I couldn't really imagine that everyone actually wanted to wade through so much text in one place. I couldn't really imagine that the six meme explanations in the post should be more accessible than one link to KnowYourMeme which, through its own links, would explain stuff like snowclones as well.

And so I ran a poll a couple of weeks ago. Call the poll rigged because I mentioned which way of explaining a meme would end up with how many estimated characters. I call it honest.

Without the character count, I guess just about everyone would have wanted me to explain everything in the post. That is, Mastodon users would have wanted me to explain everything in the alt-text; Mastodon does everything image-related in alt-text because that's how Mastodon does it, and because you can't do much with only 500 characters.

But with the character count, nobody wanted me to explain things in the post. Not if that meant 10,000+ characters.

As you wish. So I switched to linking what I could possibly link.

My next meme post, What if I told you the Fediverse is not only Mastodon? was very simple, so one link to KnowYourMeme for the meme template was sufficient. Otherwise, it went with no explanation at all.

GARGRON required an explanation for the image itself, but in addition to that, a link to KnowYourMeme was enough.

The Fediverse, late summer of 2024, my latest meme post, required some more explanation and a few more links. I had to explain the whole situation which the meme references.

The explanation had to include a very brief explanation of (streams) because the link to the streams repository (no. 3 out of a whole lot of links) isn't informative enough for someone who barely knows the Fediverse beyond Mastodon.

Firefish required a short explanation of the situation plus a link to the EOL announcement plus a link to the official Firefish repository plus a link to the official Firefish lighthouse instance.

/kbin required another short explanation of the situation plus a link to the repository to show for how long it has been abandoned plus a link to the dead website plus a link to the dead lighthouse instance plus a link to a post on Lemmy that gives some more insight.

In addition, at the top, there are two links to KnowYourMeme because I've combined two memes.

I could say that I've got meme posts down pat now. It only gets a bit difficult if I have to explain some more in the post because there are no external explanations for certain aspects of the post available. But I don't have to re-phrase entire KnowYourMeme pages anymore since people have voted against it.

My original in-world images are trickier because they work differently.

For one, they require much, much more detailed visual descriptions. Meme posts based on known and established templates hardly require any visual description. My in-world images, on the other hand, I have to re-paint with words because everything in them may be interesting for someone out there.

There also tends to be more text to transcribe. The basic rule is, "If there's text in the image, it must be transcribed." The rule does not handle any edge-cases, so I assume they have to be transcribed, too, if I can transcribe them. And I can. And "important within the context" doesn't apply if the image itself is the context.

Besides, they require much, much more explanation by me. There are far more things about them that the average casual on-looker is unfamiliar with. Thus, they have to be explained. But there are far fewer useful explanations for them readily available on the Web, if any. And before I force people to scrape together useful and comprehensible information from the official OpenSimulator website, I rather explain OpenSim myself in a way that people understand right away. And with fewer characters for more useful content.

I've written an article about this: Why descriptions for images from virtual worlds have to be so long and extensive.

My most recent OpenSim image post here on Hubzilla is this one from around April. It's only slightly outdated.

In numbers, it has

one image
one alt-text of precisely 1,500 characters, of which 1,402 characters are image description, and 97 characters are a too short note where a full description can be found
a full description of 60,553 characters in the post, right below the only image, including all necessary explanations and transcripts of all text within the borders of the image

The post is an absolute monster. It's my second-longest post to date. Only one post grew even longer at over 78,000 characters; I wrote it a year ago, and it contains three highly detailed images. But none of the image descriptions in that post manage to exceed 60,000 characters. And this post only has faint traces of a description preamble. Also, it places each description below the corresponding image.

This excessive size is one of the reasons why I considered parking the description someplace else.

My most recent OpenSim image post went to (streams) two months ago. Here it is..

It contains two images, and yet, it's much shorter. I've managed to reduce the necessary description effort by reducing the scenery in both images as much as I possibly could. Also, both images show the same avatar in the same outfit which I only had to describe and explain once.

This time, I put both images below one another in one block. The image descriptions are their own block right below. Also, this time, the preamble with common explanations for both images is the largest part, also because it contains the common description and explanation of the avatar for both images.

In this case, this was simply the easiest way to do it because both images have so much in common. However, this does not translate well to individual image descriptions per image, linked below the images themselves. If I were to do that, each image description would have to contain all necessary description elements and all necessary explanations. It might largely end up copy-paste, but still.

This was easy to do for another reason: The actual post text is only one block. The images are separate from it.

Now, however, I'm planning to write a blog-style post with up to four images within the post text, as in surrounded by post text above and below. I want to write a few things about the images that are not image description in the alt-text/accessibility sense. And that has to be close to each individual image.

This excludes individual long image descriptions per image right below each image. This also makes a big image description block at the end inconvenient because the described images would be so far away from their descriptions.

And that's why I considered putting the raw image descriptions into external documents without, however, completely overthrowing my way of describing images. I've honed it to a high level of quality over almost two years, improving my description style from post to post. And I'm fully aware just how detailed image descriptions for images from virtual worlds have to be.

I just need to find a way to make it work with a blog-style post with several images embedded within the actual post, both without ripping the post apart by placing humongous chunks of image descriptions into it and without placing the descriptions too far from the images within the post.

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Memes #Hubzilla #Streams #(streams) #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-28 22:44:42

Scott M. Stolz

scott@loves.tech

Well, I am not sure what to tell you then. You are working with a very picky audience and what is worse, there is no agreement on what the rules should exactly be. Plus there are some serious technical limitations. So no matter what you do, someone will complain.

You are going way beyond what accessibility standards require or recommend, and apparently some of your followers still don't think it is enough.

You do have to cater to your chosen audience, so I understand that. But just realize that what attracts one audience repels another. And you also have to consider whether the loudest complainers are actually the majority or not. Most of the time the loudest people are actually the minority... that's why they have to shout so much.

So you do have to consider whether you are catering to the silent majority or the loud minority. It is possible that most people don't want what the loud majority wants.

It is good that you are doing research and asking people about how to do it. But in the end, you will have to figure out what level of effort you want to put in, especially considering that you are providing labor for free to generate this content. If it is out of love, that is great. But even then, there are limited resources and time in the day.

2024-09-29 22:45:09

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@Scott M. Stolz It isn't only about my immediate followers.

It's just as much about users who may receive my posts as boosts.

It's just as much about people who happen to follow one of the hashtags I've put in one of my posts.

It's just as much about random strangers who stumble upon one of my posts on their Mastodon instance's federated timeline.

I can't choose my audience. Not when I post to the general public. My audience is whoever happens upon my posts.

I've chosen the way I describe and explain my images not because someone has complained to me directly that I'm doing it wrong. In fact, I barely ever get any feedback for my image posts.

But I pay a lot of attention to the topic of accessibility in the Fediverse. My main source of information is not my stream. It's the hashtag search on mastodon.social first and foremost.

I've learned a lot that way. I've learned that Mastodon loves long and extensive image descriptions. I've seen images with long and detailed descriptions in alt-text being praised for their long and detailed descriptions. I've actually come across posts by blind users who, when asked if they prefer brief or long descriptions in alt-text, said, the longer and more detailed, the better.

Do you know the Alt Text Hall of Fame? It's entirely about alt-text in the Fediverse, and it has its own presence on Mastodon: @Alt Text Hall of Fame. Take a look at it. Check the actual alt-texts found in the Fediverse and presented on the site. They aren't as short as your typical alt-text on your typical static website or blog. And yet, they count as the good stuff by Mastodon's standards.

Also, check out this poll which @Stefan Bohacek ran in August, 2023, about what blind or visually-impaired users like image descriptions to be like. Every one out of nine actual voters, about 30 voters, voted for "As detailed as possible".

Also, last year, I was pointed to this post when it was still fairly new.

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.)

It made sense to me. It made me realise just how obscure the primary topic of this channel of mine is, including the images. (I'm talking about my original 3-D virtual world renderings.) It made me understand just how much description and explanation they actually require.

@Stormgren was only wrong in one point: Explanations must never go into alt-text. They belong into the post text body where everyone who can access the post can access them as well. I've learned that a bit later when I tried to discuss just how detailed an image description for a picture from a largely unknown virtual world must be.

I was able to deduce quite a bit from all this.

If someone doesn't know what something in one of my images looks like, I must describe what it looks like. Just mentioning that it's there isn't enough.

And when I post a picture from a virtual world that's so obscure that even the underlying technology is only known to maybe 50 out of over 10,000,000 registered Fediverse users, I have to assume that nobody who can't see the picture knows what anything in it looks like. Especially not any of the the stuff that doesn't exist in real life.

Thus, it must be described.

Also, if someone doesn't know what something in my images is, I must explain it in a way that they understand it. Again, 50 out of over 10,000,000 Fediverse users even only know the technology underneath all these worlds, much less the actual places. So the unknown things in my images start with the technology and include the shown location as a whole.

Thus, it must be explained.

Everything that someone who happens upon one of my image posts on the federated timeline of mastodon.social is not necessarily familiar with must be explained.

A lot of what I'm doing is based on assumptions. That's all I have. What I'm doing is an extreme edge-case. Something whose existence next to nobody in accessibililty circles is even aware of. Something that no rules exist for. Something that no guidelines exist for, which meant that I had to adapt existing guidelines, based on deduction and assumptions.

And I'm still largely basing what I'm doing on assumptions. I barely get any feedback, especially not from blind or visually-impaired users. Not unless I single one of them out, get them to interact with me, and then send them a link to one of my image posts and ask them for feedback. I've done that exactly once. With rather positive feedback.

The way I handle text transcripts is a combination of taking a rule literally that doesn't even cover my situation and, again, assumptions.

The rule says that any and all text within the borders of an image must be transcribed 100% verbatim. The rule does not make any exceptions for text that's illegible in the image, but whose content can be sourced by whoever posts the image. It doesn't take such text into consideration at all. As if such text doesn't exist.

And so, in compliance with this rule, I transcribe signs that are four pixels wide and three pixels tall in the image. They can't even necessarily be identified as signs in the image at the resolution at which I post it, much less read. I can read them, though. I don't read them from the image. I read them in-world. And what's four pixels wide and three pixels tall in the image may pretty well be a texture of 1024x768 pixels in-world and therefore perfectly readable.

Things would be a lot easier for me if I could actually talk to people about all this. If Mastodon had been useful for actual discussions. If it had known threads and conversations, and if it had had discussion groups from the get-go, but at least from New Year's Day 2022 on, before millions of Twitter users came flooding in.

I could actually go and discuss such things with dozens, hundreds of Mastodon users. Blind Mastodon users. Visually-impaired Mastodon users. Those Mastodon users who defined Mastodon's accessibility standards. Mastodon's alt-text police.

I could ask them questions such as:

If there's already a very detailed image description in the post, is it still absolutely necessary to put a useful image description in the alt-text?
What if the post is hidden behind a content warning that conceals the full description, but that says that there's a full description in this post?
Do blind or visually-impaired users really need to be described what something looks like if you mention that it's there, but they're likely enough to not know what it looks like?
Let's assume I post an image from what amounts to a completely alien world of whose mere existence none of you has even heard of. Sighted users will be likely to let their eyes wander around the image and explore this new and unknown world that way, regardless of what in the image is important within the context of the post. Do blind or visually-impaired users demand being granted the same opportunity to explore this unknown world by taking in all the small details, only in written words?
Assuming that I, as the poster, can read the text in all these following cases, because I actually can: Does text really have to be transcribed if it's so small that it can't be read? If it's so tiny that it can't be recognised as text or that it can't even be seen? If it's half-obscured, e.g. by a tree trunk in front of a sign?
Is the post text body really the best and the most accessible place to put an image description of tens of thousands of characters in length? Or would alternative placements outside the post be more convenient and more accessible in this case? (And no, putting the full, detailed image description into the alt-text is out of question for multiple reasons.)
Let's assume there is a picture somewhere within the borders of my image that shows another place. The resolution of that picture is so low that details can't be made out in it. Let's also assume that it'd be fairly trivial for me to travel to that other place and see enough details to give a better and more accurate description of this picture.
If I already describe my image at an extremely high level of detail, would I have to describe this picture in the picture at a similar level of detail? Or would I be spared by the low resolution because more than that isn't visible from where the on-looker in my image is located, not without considerable travelling?
If so, how many levels of pictures within pictures would I have to travel for? Picture within a picture within a picture? Or even deeper?
When I explain where the picture was taken, can I assume the concept of 3-D virtual worlds to be understandable without explaining it?
If there's a teleporter in my picture, and I explain it, do I have to list all its possible destinations and then explain them, one by one? (This would be difficult with an OpenSimWorld beacon which also serves as a teleporter with over 1,500 destinations that constantly change.)
Let's assume an image at 1080p resolution. It's a scenery shot that's cluttered with all kinds of stuff, buildings, plants, decorations and whatnot. In the far distance, there is what amounts to a person, a human, so far away that this person takes up four pixels in height and one pixel in width in the image. This means the head is about half a pixel high and less than half a pixel wide. Could an autistic person deduce from the colour of the top pixel that this person has their face oriented in such a way that at least one of the eyes is revealed, and could this cause an eye contact trigger for the autist? (Before you ask, I've basically had exactly this actually confirmed to me.) Would my image count as sensitive and require a content warning and being blanked out?
Likewise, if there's a pale red blot of 2x2 pixels somewhere in the image that's a depiction of a strawberry cocktail, could this be an alcohol trigger? Would it become an alcohol trigger by me mentioning in my image description that this blot is a strawberry cocktail?

Alas, it's nigh-impossible to discuss these topics and get satisfactory answers from enough people. The only way to discuss these topics at all is by going to mastodon.social, searching for blind or visually-impaired or autistic people and then asking them these questions in DMs, one by one.

Also, that wouldn't even be discussions because they wouldn't interact with their own likes, only with me. What I need is someone to tell me, "Yes, it makes sense, and it sounds useful, do it," and six other people to chime in and tell both me and that someone, "No, by all means, don't do it, it's completely unnecessary and total overkill!"

And so, all I have is what I've learned so far plus deduction from what I've learned plus assumptions based on what I've learned. Plus a little hope for feedback whenever I post an image.

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-30 00:52:05

Scott M. Stolz

scott@loves.tech

@Jupiter Rowland

I can't choose my audience. Not when I post to the general public. My audience is whoever happens upon my posts.

You may not be able to choose who views your posts, but you can and should choose your target audience.

Basically, who do you want to reach? Who would resonate with your content? Those are the people you write for and you write in a style most of them will accept.

People who complain about technicalities but who have no interest in the content you post are not as important as the people who actually want your content.

And so, all I have is what I've learned so far plus deduction from what I've learned plus assumptions based on what I've learned. Plus a little hope for feedback whenever I post an image.

I am glad that you are working on accessibility and experimenting on how to do it best.

But it is hard to help you when you ask for feedback and then reply that our feedback is wrong.

I am not saying I am right, because I probably am not, since I am not your target audience. But when I said I can't help you, it means it feels like you wouldn't accept my feedback anyway. Others in your target audience are probably more suited to give you feedback.

2024-09-30 09:23:03

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@Scott M. Stolz

You may not be able to choose who views your posts, but you can and should choose your target audience.

Basically, who do you want to reach? Who would resonate with your content? Those are the people you write for and you write in a style most of them will accept.

Sure, from a Hubzilla point of view, the ideal solution would be to put everyone in my target audience into a privacy group or, if different content has different target audiences, into several privacy groups and then only post to them.

But I can't do that. It's impossible. I'm not posting to defined, fixed bubbles.

My target audience is not fixed. My target audience is everyone who may be interested in what I have to post. They're somewhere out there. They could be anywhere and everywhere out there. I can't pinpoint them.

My target audience includes the random stranger who comes across one of my OpenSim image posts on the federated timeline of their Mastodon instance and goes all, "Wow, the metaverse lives? It's real? Decentralised even?!"

My target audience for OpenSim image posts explicitly includes people who have never heard of OpenSim. I want them to see that "The Metaverse" is not Meta's creation. And not all blockchains and crypto and NFTs either. And that it's alive.

My target audience includes the random stranger who comes across one of my meme posts on the federated timeline of their Sharkey instance and laughs their butt off because they can relate to it.

In either case, my target audience includes the random stranger who just happens to follow one of the hashtags I've added to a post. A feature that Hubzilla doesn't have, but that lots of other places from Mastodon to (streams) do have.

I can't simply sit and wait until people discover me without even having seen a single post from me. Mastodon doesn't have a user discovery, much less one like the one on (streams) that covers all kinds of Fediverse server applications instead of like the one on Hubzilla that only covers Hubzilla. People find people through their public posts and not through keywords.

I may be the only one on Hubzilla whose target audience is way more on Mastodon than on Hubzilla. But that's the way it is.

I am glad that you are working on accessibility and experimenting on how to do it best.

But it is hard to help you when you ask for feedback and then reply that our feedback is wrong.

We both have one thing in common: If we don't know something, all we can do is assume, based on what we do know.

What you've written to help me was based on assumptions. I've shown you that your assumptions were wrong.

Others in your target audience are probably more suited to give you feedback.

I guess so. But they, too, are somewhere out there. My target audience for accessibility topics are blind users, visually-impaired users, users who know more about Mastodon's accessibility culture and the way the greater Mastodon community wants to push it upon the non-Mastodon Fediverse than me, members of Mastodon's alt-text police etc.

And as I can't simply gather them around me to discuss such things, all I can do is shout into the void and hope someone catches it.

#FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta

2024-09-30 15:45:56

Scott M. Stolz

scott@loves.tech

@Jupiter Rowland

Sure, from a Hubzilla point of view, the ideal solution would be to put everyone in my target audience into a privacy group or, if different content has different target audiences, into several privacy groups and then only post to them.

That is not what targeted audience means.

Let me give you an example. Let's say my target audience is people who play Star Wars: The Old Republic (SWTOR). My posts will contain information about that game.

My posts are still public and anyone can see them, but I only care about reaching people who are interested in that specific game. So my content is written in a way that they would understand.

My mom, on the other hand, does not play these types of games, and does not play this game in particular. She is not my target audience. She may follow me and see the post, but she has no interest in the topic and will just scroll by.

As such, I am not going to rewrite my content so that my mother understands the game, because she won't even read the post and will just go to the next one if she happens to see it.

My mom may say that my posts on SWTOR are too technical for her, and she would be right, but since she has no interest in the game, it does not matter. The post is not for her. If I were writing for her, the post would be completely different.

And your audience may be different for different posts. If I were writing for current players, I would write one way; If I was writing for new players, I would write a different way... because they are different audiences.

So "target audience" has nothing to do with making your posts public or private. It is who your intended audience is and speaking in a way that they understand, even if others do not.

2024-09-30 18:31:49

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

Still, my intended target audience is very widespread, regardless of whether I'm posting about OpenSim, or whether I'm posting about the Fediverse.

I mean, sure, I could write my OpenSim posts in such a way that only those who know OpenSim inside-out because they're active users understand them. It'd be easier for me.

But there are probably fewer than 50 registered users in the Fediverse who know more about OpenSim than that it exists, if even that. Of these, maybe half a dozen is active. That wouldn't exactly be a big target audience.

At the same time, there are lots of people out there who receive my posts, too, be it by following me, be it by following one of the hashtags I use, be it by discovering one of my posts on their federated timeline.

Now, these people may pretty well have their curiosity piqued upon receiving one of my posts with an image from within a virtual world, the kind of which they would never have imagined exists since the Metaverse hype ended.

At the same time, the post doesn't tell them anything about this kind of world. Nothing is explained, and nothing is described, or if something is described, it's the very bare minimum requirement for the most basic accessibility. Even if they're sighted, they'd really like to know what kind of place this picture is from, and what the picture actually shows. There may be items in the picture that they've never seen before.

It's even worse if they're blind. Sure, you may say that blind or visually-impaired people have no use for virtual worlds, thus, literally not a single one of them can even be interested in that topic.

I wouldn't count on that. I'd say a blind or visually-impaired Fediverse user may just as well be curious about virtual worlds as a fully sighted user, including what everything looks like. If they are, they've got the same right to learn about it as a fully sighted user. Everything else would be ableist.

Both need my help. No-one else can really help them. No, not Google either. Not on this level of detail, and I'm not even talking about the first hits for OpenSim probably leading them to the human body simulation of the same name.

Both need explanations. And the blind or visually-impaired user needs visual descriptions. Only I can provide them on a sufficiently detailed and sufficiently accurate level.

Not providing sufficient description and explanation for image posts on this level of obscurity is about on the same level as not providing any image description at all whatsoever for images in general. Hubzilla doesn't care. Mastodon, on the other hand, calls that ableist. And Mastodon calls it out.

My target audience for posts about the Fediverse, including memes, is even wider. It's basically everyone, especially my target audience for Fediverse memes.

But not everyone is on the same level of knowledge. And if an image is included which is almost always the case in my meme posts, I have to describe the image anyway.

I guess you've read that shared post by @Stormgren. Image descriptions are not only good for blind or visually-impaired people, but they're also good to help sighted people understand an image which they wouldn't understand without the description.

So I describe and explain the meme image.

But explanations don't belong into alt-text. I've already written about that.

So I describe the meme image in the alt-text, and I explain it in the post where explanations belong.

Sure, I could say, "If they don't get it, they don't get it. I don't care." But doing so would make these posts inaccessible to the vast majority of their intended target audience.

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta

2024-09-30 18:44:28

Scott M. Stolz

scott@loves.tech

@Jupiter Rowland

I guess you've read that shared post by @Stormgren. Image descriptions are not only good for blind or visually-impaired people, but they're also good to help sighted people understand an image which they wouldn't understand without the description.

That was actually a point I made months ago. Accessibility isn't just about blind people. I am not the only person who believes that. You keep telling me that I am wrong, but you sometimes repeat what I told you months ago, perhaps because someone else said the same thing.

I am not saying that you should ignore people's needs. In fact, I am saying the opposite of that.

What am saying is that you prioritize blind and sighted people who are interested in what you talk about over people who don't care about what you talk about.

Blind people, deaf people, people who have various sensitives, etc. are all part of your target audience. But only those who are actually interested in the topics you write about.

Prioritize these people:

"I read your posts and wish you would make these changes so it is easier for me to consume."

And deprioritize these people:

"I have no interest in your topic, and don't even read your posts, but you violated this rule that I arbitrarily made up, and I will force you to change even if your actual followers don't even want it."

That is all that I am saying.

2024-09-30 20:22:58

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

In an ideal Fediverse, I'd provide at least four to six different descriptions per original image so that as many people get a description that's as close to what they need as possible. And I'd link to them, even though that makes things more inconvenient for reasons I've already explained, and hope I'll get away with it.

Of course, this isn't feasible.

Still, I do have to find a way to satisfy as many people on Mastodon specifically as possible. Even if they aren't interested in what I write about, they may still cut into my reach.

It's incredibly difficult to get away with sub-standard accessibility for images on Mastodon. You can quickly be ostracised for not providing an alt-text as well as for providing an alt-text, but it isn't useful enough because it carries too little information about the image.

This starts at being lectured about having to provide good alt-text. It continues via refusals to boost posts with undescribed images. And it ends at people who post images without describing them being muted or even blocked. If your only connection on a particular instance blocks you, you also disappear from the instance's federated timeline, and your posts may no longer be delivered to those on that instance who follow a hashtag that you use.

Besides, reputation matters. If you don't take accessibility by Mastodon's standard seriously enough, and your posts appear on Mastodon often enough, you're likely to earn a bad rep as someone who's lazy and careless and basically ableist. Even useless alt-text is ableist by Mastodon's standards.

I already have a very active member of Mastodon's alt-text police among my connections, so I can't neglect this.

On the other hand, I don't think you'll get a bad rep for trying too hard, for doing over 1,000% of what's required and still trying to improve and optimise and max out your accessibility game.

What I can say is that I'm not constantly being scolded for giving too long and too detailed image descriptions, nor am I scolded for parking them in the post. I guess my alt-texts already keep the alt-text police satisfied for now, and they should still know that if my alt-texts aren't sufficient, there's still the long description.

The

"I have no interest in your topic, and don't even read your posts, but you violated this rule that I arbitrarily made up, and I will force you to change even if your actual followers don't even want it."

people may still cut into my reach, so I can't ignore them. Besides, they don't enforce rules which they themselves have just pulled out of thin air right there right then. They enforce rules which the greater Mastodon community has already firmly established, and which Mastodon is expected to live by. And when I say "Mastodon", I mean the typical Mastodon user's perception of Mastodon which includes everything that happens on Mastodon, regardless of where it comes from.

You aren't exempt from these rules just because you're on Hubzilla. Besides, Mastodon users can't see where you are anyway. So they'll assume you're on Mastodon unless they know better.

I'd say my current ways of describing and explaining my in-world images have been working quite well. I usually don't get any feedback, but this also means I don't receive any negative criticism about neglected accessibility.

If I were to do any major changes such as moving the full descriptions into linked external documents, it'd be a gamble.

For one, the technical side is largely untested. All I know is that a blind user once told me that a Hubzilla Article didn't work with her screen reader, so Hubzilla Articles may actually not be accessible at all. I'm not going to post the majority of my future images on Hubzilla anyway, only on (streams) which can make Mastodon blank them out, and (streams) doesn't have the Articles app.

So all I could do would be a simple HTML document uploaded to the Files app that contains the full description, maybe with the described image embedded as you've suggested. However, I don't know how which browser on which OS handles an HTML document dealt to it not by a Web server, but by a file server. Will it open and display the document, or will it download the document to the device like any other file without opening it?

Besides, what else is untested is the acceptance of linked image descriptions, especially when they're the only available source of transcripts of the bits of text in an image. After all, this means that something that's defined as absolutely mandatory, the transcripts, is not available in the post at all, only in some external document that requires extra effort and, on mobile devices, an extra app to open.

Externally linked descriptions may improve the acceptance because they drastically shorten the image posts, and they make image posts with multiple images within the post text possible in the first place.

But they may just as well have the opposite effect because they remove a detailed image description from the post as well as any and all explanations and, worst of all, every last text transcript, and instead, they park everything someplace that requires extra effort and, on phones, letting another app open to access them.

Yes, Mastodon doesn't like long posts. Some Mastodon users actually block everyone whom they catch posting over 500 characters upon first strike. But this can be partially mitigated with a long-post content warning.

And there are way more complaints about missing or useless image descriptions than about too long posts. So I dare say the former cut into your reach more deeply than the latter.

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta