> 15.4% "Jake Archibald"

> 13.8% "Avatar"

I declare those 29.2% unequivocally wrong.

“Jake Archibald” is the worst option, completely useless as alt text, because the name “Jake Archibald” follows immediately afterwards. This is one of my pet peeves, avatars that have alt text of the name that immediately follows in normal text, especially in HTML emails where images are much more likely to not be rendered, and so you end up seeing things like “Jake ArchibaldJake Archibald” or “{Jak / e / Arc / …} Jake Archibald” (spreading the alt text across many lines, either then truncated or overlapping with something else).

“Avatar” is almost as bad, because it doesn’t help screen readers in any way—the image is behaving more decoratively than functionally.

In fact, I think the HTML email stuff I mentioned suggests a good way of assessing alt text: turn off image loading so that you actually do see the alt text, and decide then whether it was a good idea or not. In this case, “Jake Archibald” is definitely right out, and “Avatar” isn’t useful.

So yeah, the two options that remain are empty alt text, and something custom. In web systems in general where the name appears beside the avatar, I’d say empty alt text is absolutely the right choice: the avatars are then purely decorative and not useful to non-visual users. (If the avatar doesn’t have a name shown alongside it, you should almost certainly use the name as both the title and alt attributes.) In specific cases like this, “Jake, cheekily peering from behind a plant” is good, where “Jake Archibald hiding behind a plant” didn’t feel right. Mind you, this is where my remarks about seeing the alt text fall down a little, because although I might want to hear it, I actually wouldn’t want to see this particular alt text, so at least in an HTML email I’d probably try to visually hide the alt text, e.g. `<img style="color:transparent">` should work in most clients, though I don’t know if it’d do for the MSO renderer that Outlook/Windows Mail use.)

The following phrase from the linked article by Léonie Watson [1] jumped out at me:

> blind people, particularly those who once had sight (and that's most of us)

I wonder if those of us who have been blind from birth tend to forget that. I'm pretty sure I do. And while I don't have data on this, my guess is that the largest proportion of blind people who are active in advocating for accessibility have been blind from birth or early childhood. That's certainly true of most, if not all, of the blind people I know online.

BTW, while I'm legally blind (i.e. can't drive), I'm not totally blind. But while I can read text up close, I seem to have very little ability to make sense of pictures even up close, presumably because I was never trained to do that at a young age; my parents and teachers were probably satisfied that I could read print. So I'm certainly unaware of any emotional content I might be missing in an image, and that's why I classify myself as blind for the purpose of this discussion.

[1]: https://tink.uk/text-descriptions-emotion-rich-images/

I wish alt text could have more internal structure. A string is not good enough: you can't separate the text into paragraphs, or have bullet points, or emphasize text or link to anything.

Something closer to the figcaption[0] element than a string.

[0]: https://developer.mozilla.org/en-US/docs/Web/HTML/Element/fi...

Emotion can matter but should not be required. Information is most important, the rest is nice to have in certain contexts.

I kept looking for the emoticons then realized I'd misread the title.

But how about emoticons and emoji?? Screen readers should get them correctly and they're also a visual reminder about what's in the image if you just see the text :)

“The best and most beautiful things in the world cannot be seen or even touched. They must be felt with the heart”

― Helen Keller

I would have liked the article to get opinions from the people who actually use screen readers. They would have the most experience on what makes alt text significant.