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

Source code Sponsor

Scott M. Stolz

Entrepreneur & Author

Authorship Studio is part of a decentralized social media network powered by Hubzilla.

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

2024-09-27 15:38:02

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

2024-09-27 15:38:30

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 18:10:43

Scott M. Stolz

scott@authorship.studio

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 15:44:42

Scott M. Stolz

scott@authorship.studio

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 15:45:09

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 wrote the following post Mon, 03 Jul 2023 11:20:44 -0500

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

2024-09-29 17:52:05

Scott M. Stolz

scott@authorship.studio

@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 11:31:49

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 13:22:58

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

Sorry, you have got no notifications at the moment...