split & tidy up the Invokers proposals, scoping for v1 invokers

[keithamus]

This split the Invokers proposal into 2 separate sections; "Invokers" (what might also be known as invoketarget/invokeaction and related behaviour) and "Interest Invokers" (what might also be known as interesttarget). It also starts to demarcate chunks of the "Invokers" proposal into what we're initially shipping, vs what will come as a follow-up.

(I'm aware this is a big diff but... it's quite a bit to tease these apart, I can try and split it into smaller PRs if people would feel better about that?)

Closes #964, #956

[mfreed7]

I'm happy to review this (unfortunately not in the next few days, I'm otherwise swamped), but I wanted to ask: does this indicate anything about the urgency or priority of landing interesttarget? Personally, that's almost a bit more interesting (ha!) to me than invoketarget, because it unlocks tooltips and unblocks popover=hint. Are you still invested in the interest target proposal?

[keithamus]

There's no change in the prose that either prioritises or defers interesttarget. I'm still very much interested in pursuing it, it (hopefully) answers some as-yet-unanswered a11y patterns (see w3c/aria-practices#128, I'm participating in the aria-wg in the hopes that we can thread this needle).

I'm hoping that we can ship invoketarget to use as learnings for interesttarget, scoping invokers down to dialog/popover pushes that forward so my hopes for interesttarget making progress are higher than they were last week 😉

[lukewarlow]

I very much hope we can get interest as a fast follow up. I do think it has more as yet unanswered questions but I'm hopeful we can make progress on it.

I believe in Chrome there's already a vast amount of code that can be reused, from the popover hint and hover experimentation.

[mfreed7]

Great, thank you both! I'm happy to assist with code reviews or whatever you need.

I believe in Chrome there's already a vast amount of code that can be reused, from the popover hint and hover experimentation.

I would agree - Chromium has a full prototype of the hover-trigger and hover-detrigger code, plus CSS properties for controlling the delays, plus testing. That'll all likely need modification, and it doesn't handle non-mouse triggering, but it should hopefully be a good start.

[keithamus]

@mfreed7 could I get your eyes on this again? I think it'd be useful to separate these explainers so we can iterate more on the interest parts.

[lukewarlow]

I've left some comments but overall this looks good.

[mfreed7]

Sorry for the delay. Lots of comments, but this LGTM to land. We can iterate after landing.

[keithamus]

Thanks for the review from both! Aside from the various typos/nits, the changes of substance are:

• Pulling out the descriptions on alternative HIDs and explicitly calling out that this is an in-progress area that's out of scope for the initial release.
• Removing mention of adding delays, punting to another discussion/issue/PR.

[mfreed7]

• Removing mention of adding delays, punting to another discussion/issue/PR.

That's ok by me, but let's definitely not forget about that part. I do think it's important.