Some Things About `alt` Text

I’m sure you know about alt text. It’s the attribute on the image tag that has the important task of describing what that image is for someone who can’t see it for any reason. Please use them.

I don’t want to dimish the please use them message, but some interesting alt-text-related things have come up in my day-to-day lately that are related.

When you don’t

Hidde de Vries wrote You don’t always need alternative text recently:

But when an icon has a word next to it, for example ‘Log out’, the icon itself is decorative and does not need an alternative text:

<button type="button"><img src="close.svg" alt="" /> Close</button>

In this case we can leave the alt attribute empty, as otherwise a screenreader would announce ‘button – close close’.

I would think in a perfect world because that icon is entirely decorative that applying it via CSS would be ideal, but the point stands: if you have to use an image, alt text hurts more than it helps here.

Can we get it for free?

Computers are pretttttty smart these days. Perhaps they could look at our images and offer up descriptions without us manually having to type them.

🔥 I think I used machine learning to be nice to people! In this proof of concept, I’m creating dynamic alt text for screenreaders with Azure’s Computer Vision API. 💫https://t.co/Y21AHbRT4Y pic.twitter.com/KDfPZ4Sue0

— Sarah Drasner (@sarah_edo) November 13, 2017

Sarah’s demo uses a Computer Vision API to do that.

What about figcaption?

I hate to say it, but I’m not particularly good at writing alt text descriptions for images in blog posts right here on CSS-Tricks. It’s a problem we need to fix with process changes. We do often use <figcaption> though to add text that’s related to an image. The way that text is often crafted feels like alt text to me. It describes what’s going on in the image.

I was asking around about this, and Zell Liew told me he does the same thing:

I actually have the same question. Most of my figcaptions are used to describe the image so readers understand what the image is about.

In my mind, I figured it would be worse to drop the exact copy from the figcaption into the alt text, as someone who was reading alt text would then essentially read the same description twice.

I also talked to Eric Bailey who had an interesting idea.

<figure>
  <img src="screenshot.png" alt="Screenshot of Chrome displaying a split view. On the left is a page full of image thumbnails comparing pre and post-optimization filesize. On the right is Chrome developer tools showing paint rasterize duration for the images. With a 6x CPU slowdown, the longest Paint Raster took 0.27ms, AKA 0.00027 seconds.">
  <figcaption aria-hidden="true">
    With a 6x CPU slowdown, the longest Paint Raster took 0.27ms, AKA 0.00027 seconds.
  </figcaption>
</figure>

This:

• Preserves figure styling
• Avoids nulling alt, which can be problematic for some screen readers
• Keeps the description close to the content and communicates point to SR users
• Communicates significant takeaway to visual readers without duplicating reading for SR users
• Uses an aria property designed to be used outside of forms

I’d stress that he considered this just an idea and it hasn’t been heavily vetted by the larger accessibility community. If there are any of you out there reading, what do you think?

Eric’s demo used a more verbose alt text than the figcaption, but it seems like the pattern would be fine even if they were identical.

Little reminder: Twitter has alt text

But you need to enable the feature.