Should transcriptions of text-based images be included (only) as alt-text?

Question

Often on Skeptics.SE, we get questions about some viral meme taken from social media which is a graphic containing nothing but text.

This makes it difficult for search engines to find the page when people search for the text inside it to see if the claim is true. It also makes it difficult people who use screen-readers to be able to understand it.

When editing such questions, my long-standing practice is to simply transcribe the words in the graphic directly into the alt-text part of the markdown, cramming it together as one line.

I assume that screen-readers and web-scrapers will happily be able to find it then, and pat myself on the back for making the Internet a more usable place.

Today it occurred to me: I don't use a screen-reader. I am not a web-scraper. Maybe I am not really helping at all. I should check.

I've read How can I write good alt text? which is more about describing a photo or info graphic rather than transcribing a text box. I've read Five golden rules for compliant alt text and I think I am following them.

But, I would love to hear from someone who uses, or is an expert, in screen-readers wade in on whether simply trancribing the text is sufficient. I would also be interested to hear whether it is necessary to include the text in the body of the question, or having it only as alt-text is sufficient.

Answer 1

Please consider that alt text is only rendered when the image is not (for example, with screen readers, or if corporate IT blocks the image). However, there are also people who sometimes have trouble reading text in images -- if it's small or uses a challenging font, if the image is blurry (like photos), if the contrast or color choice is poor, etc.

For that reason I think it is better if we transcribe text in the body of the post (in a blockquote) to make it more accessible to everybody, even when the image is delivered to the browser. If you fully transcribe an image, then the alt text would be "".

Of course we should use judgement here; if an image is of a page in a book and the important part is one sentence, that's the part that's most important to transcribe, and in that case the alt text should describe the image more generally.