Thank you for bringing this up. It's I who apologize for being ignorant of these expected keyboard interactions. I'll make an issue in the repo to address this.

This is an interesting project that looks useful for making React components more accessible: https://react-spectrum.adobe.com/react-aria/index.html

"React Aria provides accessibility and behavior according to WAI-ARIA Authoring Practices, including full screen reader and keyboard navigation support. All components have been tested across a wide variety of screen readers and devices to ensure the best experience possible for all users."

Nonsense. A deep study about web accessibility some 2 years ago revealed that 98.8% of web pages have accessibility issues. So saying "might as well not exist" is a dramatization of reality.

The significant majority of accessibility issues found by those sorts of large-range studies are comparatively minor, and often completely spurious. I have been decidedly unimpressed with every such study that I have delved into the results of.

(I know you’re arguing against someone claiming 100% accessibility coverage as a requirement, but I wanted to point out that figures like 98.8% are grossly inflated and quite meaningless.)

Those minor or spurious issues may well disable a disabled person even more.

That said, the entire accessibility scene is fundamentally broken on the technical side and full of snake oil salesmen on the business side.

Anxiously waiting for WCAG to finish an actual standard, and for vendors to actually implement it.

I’m well familiar with the issues of accessibility, and I use the words “minor” and “spurious” deliberately. Most especially, when I say “spurious”, I mean false positives.

I have yet to see a false positive. It might appear that way to someone only testing on a single platform, a single browser and a single reader, macOS + Chrome + VoiceOver for some reason being the usual combo even though VO doesn’t work right with Chrome.

Typically devs will also ignore reader shortcuts and instead use tab navigation, which is not how screen reader users do things because it bypasses screen reader navigation altogether.

Do you know for a fact that the ”false” positives weren’t actual problems for someone using NVDA or JAWS shortcuts to navigate on Firefox on Windows, or some other combination of reader + navigation + browser + OS?

Everything is broken and nobody can test or fix everything.

Not all web pages are owned by public or commercial entities. Those that are were not under legal obligations two years ago, but are now or are in the grace period.

> without 100% accessibility coverage

If that's the standard for open-source projects, it's a wonder any FE components get used

It is a legal requirement to provide equal access to all features and information for people who use accessibility features. Any non-trivial deviation from this must be declared in an accessibility statement, or the site owner risks legal action and large fines.

100% is not hyperbole, even though ”not exist” is.

As someone who has led WCAG remediation projects, I can firmly assure that it is hyperbole.

100% “accessible” is meaningless: full adherence to what standard? 2.1 AA? That’s an incredibly high bar, and would require constant vigilance and expensive audits on a rotating basis. Much of the guidelines are up for interpretation and can’t easily be expressed in a binary pass/fail besides.

2.1 AA is mandated by law in the European Union for public entities, as well as commercial entities (soon). And yes, it is a high bar, for a reason.

Just tried it with NVDA on Windows and it appears that it does not handle keyboard accessibility at all and does not communicate the relationships between nodes.