Scott O'Hara digs into the
<figcaption> elements. Gotta love a good ol' HTML deep dive.
I use these on just about every blog post here on CSS-Tricks, and as I've suspected, I've basically been doing it wrong forever. My original thinking was that a
figcaption was just as good as the
alt attribute. I generally use it to describe the image.
<figure> <img src="starry-night.jpg" alt=""> <figcaption>The Starry Night, a famous painting by Vincent van Gogh</figcaption> </figure>
I intentionally left off the alt text, because the
figcaption is saying what I would want to say in the
alt text and I thought duplicating it would be annoying (to a screen reader user) and unnecessary. Scott says that's bad as the empty alt text makes the image entirely undiscoverable by some screen readers and the figure is describing nothing as a result.
The correct answer, I think, is to do more work:
<figure> <img src="starry-night.jpg" alt="An abstract painting with a weird squiggly tree thing in front of a swirling starry nighttime sky."> <figcaption>The Starry Night, a famous painting by Vincent van Gogh</figcaption> </figure>
It's a good goal, and I should do better about this. It's just laziness that gets in the way, and laziness that makes me wish there was a pattern that allowed me to write a description once that worked for both. Maybe something like Nino Ross Rodriguez just shared today where artificial intelligence can take some of the lift. But that's kinda not the point here. The point is that you can't write it once because
alt do different things.