Re: [csswg-drafts] Pseudo classes for the `interestfor` API (#12154)

Great discussion today - thanks all. Here's what I heard:

- There is some concern (which I agree with) that `possible` likely introduces trivial circularity, if it is implemented in a useful way. Folks seemed to agree that `possible` should only match if interest was actually "possible", which would mean the `interestfor=foo` idref should be a valid reference, and **if the element is focusable** for example. And that means that `:interest-invoker(possible) {display:none}` would be circular.
-  (Fix for above is that the developer can easily use `[interestfor]` as a stand-in for at least `:interest-invoker(possible)` that doesn't have circularity problems.)
- There were concerns (@fantasai) about the naming of `:interest-target()` because "target" is confusing. I don't disagree, but would love suggestions for an alternative. The only other one I've seen on this thread is `:target-of-interest` which I don't love. I also wonder whether it's possible to have two arguments to a functional pseudo class? If so, we could do something like `:interest({invoker|target}, {partial|total})`?
- There was a discussion of the mutual exclusivity of `partial` and `total`. The current Blink prototype behavior is that `total` is a superset of `partial`, because the most useful behavior seems to be just "does it have any kind of interest". @emilio suggested perhaps making `partial` and `total` be exclusive, but have `:interest-invoker` (without parens) match `partial` **or** `total`. There's a question about what to do with future/other values like `possible`. My thought would be that `possible` should definitely not be included in `:interest-invoker` because then it just devolves to `[interestfor]`.
- @LeaVerou suggested a new idea, about a combinator that gets you "from" the interest invoker "to" the target of that invoker. I like that idea, and I'll open an issue for it.
- There was quick mention of the relatively new OpenUI idea of a `::interest-hint` pseudo *element*, which is [detailed in the explainer](https://open-ui.org/components/interest-invokers.explainer/#option-35---add-an-opt-in-info-button-pseudo-element). Perhaps I'll open a CSSWG issue for that one also, so we have a place to discuss.

My tentative takeaway is that:
- The general shape is "ok", with two functional pseudo classes (or a single multi-argument one?) that have values for `partial` and `total`, and a version without parenthesis that matches *either* `partial` or `total`.
- Open question about whether `total` is a superset of `partial` or is exclusive.
- There are issues with `possible` that at minimum likely mean it should be deferred for later, to see if there are real use cases.
- Naming might still need work - suggestions appreciated.


-- 
GitHub Notification of comment by mfreed7
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/12154#issuecomment-3028591124 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 2 July 2025 16:58:03 UTC