What alt-text can really be for

Thu, 27 Jul 2023 11:52:03 +0200

jupiter_rowland@hub.netzgemeinde.eu

#AltText is only for people who can't see properly, they say.

A short and concise alt-text that's reduced to what really matters within the context of a post is enough, they say.

Non-sense, I say. And so do others.

There's more than enough justification for giving full and detailed #ImageDescriptions including #transcripts of everything written in them and explanations of things which your audience may not know or understand otherwise.

Here are some excerpts from another #Mastodon thread (https://mastodon.social/@KydiaMusic/110777644794322792):

KydiaMusic wrote the following post Wed, 26 Jul 2023 02:51:54 +0200

A gentle reminder that this is Disability Pride Month—a perfect time to get into the habit of adding ALT text to any pictures you post. There are many sight-impaired individuals who would appreciate it, and even those with excellent vision sometimes need clarity on what is going on in a picture, or what is significant about it. ALT text is also a great opportunity to add more context, humor, or commentary.
*see addendum below

#ALTText #DisabilityPride #Disability #SightImpaired #VisionImpaired

👩‍🦯The Blind Fraggle wrote the following post Wed, 26 Jul 2023 03:10:56 +0200

@KydiaMusic And if you're an aspiring writer, it's a great chance to work on your skills!

This also goes for screenshots of text. Screen readers see them as images, which means they can't read them. So yep, that means you've gotta type them all out. Sorry, even the long ones.

Smythe wrote the following post Wed, 26 Jul 2023 04:38:31 +0200

@KydiaMusic Another benefit of ALT text is that you may be waiting for a slower instance to update an older image, but the ALT-text will let you know if you really need to wait, but telling you what the image is.

impulsenine wrote the following post Wed, 26 Jul 2023 10:36:25 +0200

@KydiaMusic shout-out to my fellow sighted people who are just really easily confused 😅

Zazzoo 🇨🇦 wrote the following post Wed, 26 Jul 2023 19:43:20 +0200

@KydiaMusic I love alt-text, it's a brilliant feature. It's not only useful to people with screen readers but it's a fun place to be creative and add additional commentary about the image that otherwise wouldn't fit within your instance's character limit.

(Disclaimer: These excerpts were neither hand-typed nor copy-pasted. They were generated semi-automatically by the Share functionality that has been built into #Hubzilla since it's inception. They are guaranteed to be full, verbatim quotes.)

#Fediverse #ImageDescription #Transcription #Accessibility #A11y #Inclusion #Inclusivity

show all 7 comments

Thu, 27 Jul 2023 12:59:06 +0200

Patrick H. Lauke

patrick_h_lauke@mastodon.social

@jupiter_rowland "explanations of things which your audience may not know or understand otherwise" however, as a counterpoint: if you only provide this extra context/information in a text alternative, it may not be available to keyboard users if it's *only* available on mouse/pointer hover (depending on their particular client/browser)

Thu, 27 Jul 2023 14:52:00 +0200

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@Patrick H. Lauke I'm kind of torn here. Torn between a good post style and easy availability of information, at least as long as an #ImageDescription doesn't end up so massive that it's too long for #AltText anywhere.

On the one hand, when I post something with a picture, it isn't always Instagram-style/Tumblr-style/Pixelfed-style where the post is about the picture, in which case describing the picture in plain sight doesn't hurt the post, style-wise.

It may just as well be an announcement or a bit of news or a commentary, and the picture is only illustration/decoration and technically unnecessary. If I describe an image in the text of a post which is not about the image, in details even, that's terrible style.

On the other hand, I've heard about enough cases in which alt-text is plain inaccessible for users for various reasons. Some of them might even find it hard to navigate to an image description outside the post itself which would also have the disadvantage that it's impossible to access the image and its description simultaneously. So #accessibility basically demands to throw any traces of good post style overboard, put the full image description into the visible part of the post and have the alt-text only mention that the image in question is being described in the post.

Sometimes I have to do this anyway, namely when following all rules of good and detailed and informative image descriptions to a tee results in truly enormous image descriptions like you've probably never seen one. I'm not only talking about too long for Mastodon (1,500 characters). I'm talking about so long that no front-end for any Fediverse project that has been launched since 2010 can display it in a satisfying way.

I'm on Hubzilla. I can put an image in a post and give it a 50,000+++-character alt-text if I want to. Practically no limit here. But even Hubzilla couldn't display anything nearly that long. And most screen readers would have to spend an eternity rattling the whole thing down. Not to mention that Mastodon would cut everything beyond the 1,500-character mark off.

I've been in a situation which might be an absolute edge-case, but which still is a case.

To sum up the situation:

Picture with lots and lots and lots of details in it.

Some Fediverse users want image descriptions that go beyond what's absolutely necessary within the context of the image. And some Fediverse users actually write highly detailed image descriptions, only about images with not nearly as many details that were taken in real-life settings which people can be supposed to be halfway familiar with.

The picture was taken for and is used in a very short kind of news post which is not about the picture. It is only there to "illustrate" the news while not providing any extra information.

The picture was taken in a virtual 3-D world. Maybe four out of 13,000,000+ Fediverse users know the location. This isn't Times Square, and this isn't the Eiffel Tower either. If people are unfamiliar with a place, it needs to be described. So this place needs a full description.

Since the picture was taken in that virtual 3-D world as opposed to real-life, mentioning what's in that picture doesn't cut it. People don't know the stuff that's being mentioned. A lot of it needs to be explained. In fact, they have to be introduced to that virtual 3-D world system first and foremost.

Lots and lots of signs and logos and stuff with text on them. Even disregarding surfaces that face away from the on-looker, thus being completely invisible, the image still shows more than 20 instances of text in some way which, from the perspective of the image, can be seen.

Good image description style demands that any and all text in an image be transcribed. The only exception I've ever read about is when a transcription compromises someone's privacy. This does not apply anywhere in the picture.

If you take a look at the image at the resolution I've used, you may see that most of that text is so small that it can't be read anyway. Three letters "M" are large enough. Nothing else is larger than two or three pixels. Some text is in the sub-pixel range and therefore partially actually invisible. There's even a big sign in the middle that's partly obscured by a tree trunk.

However, there's no rule against transcribing text that's illegible to begin with, at least none I'm aware of. If I can transcribe it, I will. And I did. In fact, illegible text should definitely be transcribed because absolutely everybody can't read it, not only those with poor or without eyesight. I only made one exception: I didn't transcribe what's written on the "screen" of the laptop computer at the info desk which you wouldn't even know that it's there if I hadn't told you. The resolution is too small even in-world.

In addition, more than one third of all that text is not even in English. Some is in German, some is in French, some is in Latin. Since transcriptions aren't transcriptions if they aren't 100% verbatim and identical with the original, I transcribed them all word by word, letter by letter, including special characters and diacritics, only leaving out line breaks. Then I translated them into English. The only time when I didn't translate something which was also available in English right above or below was when the English version in the picture was already a direct translation of the German version, and it directly followed the German version on the very same sign.

After following all rules of good, detailed, informative #ImageDescriptions I'm aware of, I had something that might be the longest image description in the history of the Fediverse: 13,215 characters. That's almost nine times as long as Mastodon's alt-text limit at which the image description has only just barely begun describing the location. And that's over 26 times as long as Mastodon's toot limit. If I were on Mastodon, posting the image description as a thread might have taken me half an hour.

Research, including in-world research (I visited a static preserved copy of that very same place to transcribe most of the text and see more details in detail), writing and redacting must have taken me some five or six hours altogether. Only for this one picture, only to comply with existing standards and people's requirements/needs/wishes.

The first time I used the picture was in early July 2022. That was before I wrote alt-texts, so the image doesn't have one. Here's the post I used the picture in:

https://hub.netzgemeinde.eu/item/ae405262-2ca1-4622-a10e-771f57e23372

A few days ago, I remade the post four times in various styles with alt-text plus an explanation, just to show four different ways of handling the situation.

Variation 1: short alt-text which some Fediverse users ask for; absolutely no detailed image description.

https://hub.netzgemeinde.eu/display/b64.aHR0cHM6Ly9odWIubmV0emdlbWVpbmRlLmV1L2l0ZW0vYzMzYzA4NTAtNTMzNC00NTE5LWJlMTUtNTFjNDlkZTU4ZTk2

Variation 2: full image description in the alt-text, just to show that it doesn't even work. The only difference between the image description here and that in variations 3 and 4 is that this one doesn't have line breaks and links; otherwise, they're identical.

https://hub.netzgemeinde.eu/display/b64.aHR0cHM6Ly9odWIubmV0emdlbWVpbmRlLmV1L2l0ZW0vZTRiNjU2NjItNDJmZC00ZjBmLWE2M2QtYmM2MDk0NTgxMzY1

Variation 3: full image description in an external article, alt-text refers to the link that takes you to the description. No visible image description disturbs the post yet; it looks fairly good. Still, an image description is available.

https://hub.netzgemeinde.eu/display/b64.aHR0cHM6Ly9odWIubmV0emdlbWVpbmRlLmV1L2l0ZW0vYzFmOTFlYzMtZTBlYS00ZjQ4LWE1NGMtYTMxMTY5NGJhNGY4

Variation 4: full image description in the post itself. That's 115 characters of post text, an image and 13,000+ characters of image description. For Mastodon users, it's 115 characters of post text and 13,000+ characters of description for an image they won't even see until they've scrolled all the way down this massive description. Still, it's the most accessible variant.

https://hub.netzgemeinde.eu/display/b64.aHR0cHM6Ly9odWIubmV0emdlbWVpbmRlLmV1L2l0ZW0vY2RlNTk2YWEtZjUyNi00NDkyLWFmNTAtM2MyZDc2OWRhNWQy

Judge for yourself which one looks best, and which one works best while being informative.

Thu, 27 Jul 2023 15:53:53 +0200

Patrick H. Lauke

patrick_h_lauke@mastodon.social

@jupiter_rowland thank you for sending me back an essay *grin*

not disputing any of your points, just wanted to add the caveat about being careful not to lock up necessary/important context information *just* in the text alternative, as that then may make it inaccessible to other users.

Mon, 31 Jul 2023 00:31:26 +0200

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@Patrick H. Lauke Don't worry about that, at least not in my case.

My #ImageDescriptions have the tendency of growing so long that they don't work as #AltText at all. I can put almost anything of any size into an alt-text, but Mastodon crops everything down to 1,500 characters even when it comes in from outside, and those projects that don't crop alt-texts eventually can't show oversized alt-texts in their entirety anymore. I've tested that actually. Not to mention that screen readers will go haywire.

So it's likely that I put an image description into the text body of the post itself, no matter how bad style that may be. I think it doesn't get much more accessible than that.

Sat, 05 Aug 2023 00:07:42 +0200 last edited: Sat, 05 Aug 2023 00:10:43 +0200

_jayrope

jrp@hub.kliklak.net

@Jupiter Rowland I'd very much appreciate more great (image-describing) texts instead of meaningless images :) But ... how to describe this? https://social.anartist.org/@Solarphunk/110824408940773527

Sat, 05 Aug 2023 18:55:45 +0200

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@_jayrope I'd never even try to describe abstract art. I leave that to others.

Sat, 05 Aug 2023 18:59:02 +0200

_jayrope

jrp@hub.kliklak.net

@Jupiter Rowland Well here's one example that might make for good rehearsal