From time to time I activate Narrator screen reader at Windows-11 to test accessibility. I want to “hear” how a tiddler will be read as a whole.
Here is what I did at the GitHub preview:
Adds a new alt-text field to every image tiddler I could find. Especially those used for HelloThere - flex cards.
Improve flex-card procedure to render alt-text for images, so they can be read by screen readers.
The PR “aria-hides” the new image-badge because it causes problems – Narrator reads the full card text 2 times
So for everyone of you, who use screen readers, or know how to use them properly, I would like to know, if the changes made actually are an improvement. Or did I make it worse.
I’m not a regular user of screen readers, and I suspect those who really rely on them know which ones are state of the art. So I certainly can’t comment on whether your changes improve things or make them worse. Still, I think it’s good to practice orienting to how a site might come across with accessibility tools.
Turning to the first suggested accessibility reading plugin for chrome (“Read Aloud”), I just tried it on tw-com. It starts by treating the Contents (sidebar tab) items as an ordered (numbered) list. (Perhaps it imposes ol so that it’s easier to navigate/choose.) Quite jarringly — surely because of the implicit period in the ordered-list mechanic — the voice separates each number from its content, then runs each list-item together with the number for the next-in-sequence, like this: “One.… HelloThere two.… Learning three.…” (etc.)
I don’t know how much savvy is built into good reading algorithms, and how much we can do things to help the voice synthesis with better punctuation/timing hints. At any rate, that’s my first small bit of feedback.
That’s interesting. I also had to add a “full stop” to the alt-text field. Otherwise the alt-text and the “flex-card” text body are read as 1 sentence.
I did not change anything in the sidebar. I did only change something in the HelloThere tiddler.