At some point I should actually dive into it in more depth, I haven't given it more than a quick look yet 🙈 I like the idea a lot though, I constantly have to tell folks not to wrap a whole card or region in a link or button, and to use a JS click event / trigger a link click instead.

Though from your comment it seems I could make that more clear & go more into how the underlying semantics can support a number of different visual permutations

The name is on purpose because on a technical side, the accessibility API mappings and the way this type of control is exposed to screen reader users is essentially identical to a static list, but with interaction & keyboard behavior 😅

I really should've just asked you for a good github example before I posted, I was just being lazy 😂

Generally I don't sweat specific focus order things like that too much, as long as the accName of the wishlist button makes it clear what product it belongs to. But I do think you're right that if interactive list is otherwise a good choice, it'd also make that specific issue go away.

It could! Ecommerce is tricky because it has a different balance between learning curve vs. efficient keyboard & AT browsing vs. something like an email app. But also I could see use cases where an interactive list would desirable depending on context / checked with user testing 👍

(please excuse the APG salt, the layout grid and role=menu nav examples have been longstanding frustrations. Actually hoping / have discussed interactive list being a reason to finally retire layout grid)

Heh I was just talking about exactly that with someone last night. I think it'd work really well in combination with a whole bunch of the interactive list examples in the post. Interactive list + focusgroup + link/clickDelegation could make a no-JS single-tab-stop nav or set of cards really easy.

Ahhh, now I get what you mean. That makes sense!

Think I might add that to the post or github issue w/ credit to you later 😊

Generally I don’t think the combobox role is as flexible as your repo’s readme makes it out to be, but there are other avenues you could explore — aria-activedescendant without the combobox role works fine on a search input or multiline textbox, for instance

Combobox is a bit more tricky — it needs a screen reader user to stay in forms mode to work properly, so semantic or interactive nested content is not going to work. Also since the pattern is fundamentally based on selection, I don’t really see the use case for list over listbox for the popup.

If the cards are just a stylistic thing and there’s no extra info beyond the name of the checkbox or radio, then it could also just be checkboxes, radios, or a listbox

Depends on context, but if arrowing is called for and there’s also semantically relevant stuff inside the cards, then that sounds like it could be an interactive & selectable list!

Oooo yes! Pagination is a great example for interactive lists, especially when there can be a ton of pages.

Though I do agree that devs generally are bad at accessibility for anything complex. Dunno how to fix that 🤪.

That wouldn't address the mode-switching issue, since the semantics of lists +buttons wouldn't support arrowing (actually have a line about that in there because it's come up before 😅). Using focusgroup also requires choosing a token, likely `toolbar`, which doesn't really cover the use cases.