[Someone]

Mac: VoiceOver (free with MacOS). Windows: probably Jaws (https://www.freedomscientific.com/products/software/jaws/. Expensive, but Narrator is too limited). Linux: wouldn’t know.

As a first rough test (for a blog; if you’re going to have lots of images things get more difficult), do select-all, copy, paste into a text editor. If you see all the text, it is in reading order, paragraph breaks (line breaks in code sections) are where you expect them, words aren’t cut into pieces and there are no noise characters in-between, chances are decent you’re good (I’m sure you still can mess things up if they do, for example by adding a zillion invisible pictures to a page, but doing so accidentally isn’t likely) (disclaimer: browsers may have gotten smarter with this. If so, this rough test has lost value)

If you see problems there, those are areas where you want to pay attention to when looking how screen readers behave (doesn’t mean you should ignore the rest)

Also, look at screen colors and contrast. Many legally blind users have some remaining vision that they may use.

[PaulHoule]

Contrast is important for many people who are not blind but who have vision that is less than perfect. People who live a normal lifespan get Presbyopia which means you can’t have everything perfectly focused at all distances.

I set the design rules for my three sided cards so that people don’t need to hold them close in order to read the text. You can trade off contrast vs sharpness in terms of ease of perception.

Given the vast variation of screens and situations in where you might view a print the ideal image is meaningful when viewed in a poor situation, but reveals rich depth under good viewing situations.