Thoughts on embedding alternative text metadata into images
You can’t read anything by Eric Bailey without taking away a big ol’ bag of gems. I personally have not heard buzz around the idea of embedding alt text in images along with the rest of the metadata that’s packed in them. At first glance I was like, UH-HUH! There’s been more than one situation where I’ve written alternative text several times for the same image and thought there must be more of a silver-bullet for that kind of repetitive work.
But that’s where my thinking is already off the rails. It’s all too easy for me to think of *my* convenience over a user’s convenience. It’s the classic UX-DX struggle for balance.
The lustre of embedding alt text is lost the more you poke at it. Thank goodness we have folks like Eric who are not only willing to do the hard work of picking this stuff apart, but openly sharing it for the rest of us to learn from.
What I like most about Eric’s post is that it is less prescriptive than it is asking the questions that someone like me would likely fail to consider before excitedly jumping on an idea that might make my life better:
- Can the placement of an image alter its context such that the embedded description is no longer relevant?
- Can we trust someone — or anyone — to provide high-quality descriptions beyond “screenshot”?
- As a content author, should I be able to edit a description that someone else embedded if I am repurposing an image file?
- What happens if marketers come to see this as a promotional opportunity for injecting superfluous “brand identity” into an asset, or making a sale of the thing in the image?
- As long as we’re talking about marketers and bad faith, what about embedded descriptions providing new phishing avenues for malware, ransomware, and the ilk?
- How are embedded descriptions translated for multilingual support?
Another nugget I’m taking from Eric is the concept of optical character recognition, you know, that relatively new-ish feature on many phones that can not only recognize, but select, copy, and paste text that appears in an image. There are times I’ve loved and loathed the feature, but now I’m looking at it as an opportunity for correcting alt-less images rather that have more chance to improve user experiences than yet another possible be-all, end-all solution for making my own life easier.