I wonder if at any point browsers will offer a low level accessibility API for you to manually describe components. I’ve worked in the web for years and I’m a big believer but it’s also indisputable that Canvas offers more performant UI rendering than HTML when done correctly. I don’t think it should ever be used for web “documents” but web apps already bastardize HTML and CSS to achieve their aims anyway. Accessibility remains the missing component.

As far as standards is concerned, that API is ARIA [0].

W3C already offers guides for accessibility and canvas. But no one who opts for canvas turns around and remembers to do their landmarks.

> But no one who opts for canvas turns around and remembers to do their landmarks.

Not completely true. Flutter has been adding some accessibility for web canvas target. [1]

I think Avalonia is in in the make it work phase. Accessibility will probably be added in the make it right phase.

[1] https://docs.flutter.dev/ui/accessibility/web-accessibility

Then I’m showing my ignorance… how do you add ARIA landmarks to Canvas elements?

You would create transparent DOM elements in the right places with the right ARIA attributes and content, I suspect.

I guess that’s what I’d like to see a better API for, then. Mapping on click events for invisible elements feels like a hack.

It's HTML imagemaps from the 90s, when we could not style buttons and navbars where GIFs with links in the right places. Browsers still have the code to render them.

https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...

https://caniuse.com/mdn-html_elements_map

OTH, we are still failing to provide a bare minimum for accessibility. Heck, we even needed a law (in the EU, that than needed to be translated to national law), so that companies providing crucial end user services would care about accessibility.

That's possible. But it is difficult to get right, and can have poor performance if you you have many such elements.

> I wonder if at any point browsers will offer a low level accessibility API for you to manually describe components

Accessibility Object Model:

https://wicg.github.io/aom/spec/

It's very slowly coming together, but it won't be rone for many years yet. Especially since what you want is Phase 3.

I'd just like the PgUp, PgDn, and arrow keys to reliably scroll a web page.

Oh my yes.

I want every app and every web page to be 100% navigable if I do not have a pointing device attached to my computer.

And I want this enforced by law, by large rich countries. Accessibility to people with disabilities would be a good way: if your product or service is not accessible to people who can't see, can't use a mouse, or can't use their hands at all, then you can't sell it.

I'm not aware of any screen reader that works by continuously feeding screenshots of user interfaces into a remote expensive image LLM, which is an absolutely insane and impractical idea for many reasons, but I used standard TalkBack on Android

That’s not exactly fast for people who need these tools though.

I can't code, I'll use an LLM to write one!

I can't use your app, I'll use an LLM to read it!

Slow and expensive. You’re just proposing yet another “disability tax”.