Clarification needed about the definition of non-text content and "programmatically determined"

[A11yTea]

Hi everyone,

I used to pass images who were coded as decorative (meaning they had a null alt attribute) as long as there was visible or invisible text that made the image redundant. But now, I re-read the definition of "text alternative" and I see that the concept of "programmatically determined" is stressed.

In addition, the definition of "Decoration, Formatting, Invisible" looks like it does not cover redundant information either.

Would you fail an image under 1.1.1 Non-text Content if it was coded with a null alt attribute but had a descriptive caption that contained all of the image's information?

[patrickhlauke]

Would you fail an image under 1.1.1 Non-text Content if it was coded with a null alt attribute but had a descriptive caption that contained all of the image's information?

in short, no (of course, provided all other things are equal ... if the text itself was in a completely different part of the page, then it'd likely be a problem, assuming that then a user who can't see the image and needed whatever info/context that image was providing won't get the equivalent in the same-ish spot as the image)

[A11yTea]

Is there a problem with 1.1.1 Non-text content with redundant content being assumed but not necessarily explicit in the understanding document, techniques and definitions within the SC criteria.

If the caption uses a <p> tag, does it fail because its not programmatically determined? @patrickhlauke

[mraccess77]

I've always taken the association bit to be flexible in the sense that it could be in close proximity, in a caption, or even a reference - but it's always been a point that has raised some question in my mind to what is required. However, we have historically allowed same page alternatives as well as a conformance alternative - and thus if you look at the text as an alternative to the image marked decorative then it could be met that way as well.