Post #4424795
2026-08-06 19:38 UTC
@fleuron@mas.to Great question. OCR is super handy for stuff like this (just double-check the output to avoid any funny typos). If the goal is for everyone to "see" the table and all of its content, then yeah, you're gonna want to describe it as well. The way I would do it is to OCR the text, and then add the description around that.
For example, when I run OCR on that image, I get a pretty gnarly mess of output, but at least the content is all there. And it's pretty easy to describe the table.
Starting with the OCR:
"Live? Coding?
Alignment Chart
output
traditional
output
neutral
output
radical
..."
And then turning that into useful alt text:
"A four by four grid titled 'Live? Coding? Alignment Chart'. The first row consists of a blank space, 'output traditional', 'output neutral', and 'output radical'. The second row consists of..." etc.
Make sense? Just paint a picture of what's there so that someone who can't see it can still follow along, and you're good to go!
Replies (1)
-
@fleuron@mas.to 2026-08-06 19:43
@neatnik@social.lol Oh sorry, maybe I misphrased, but I think you still answered my question: - I should not assume that anyone uses OCR on pictures of text - even if they do, it might not be usable output - it's better to include the full text of an image of text in the alt text did I get that right?