[css-contain-2] Should CSS highlights make content relevant to the user?

[nt1m]

This section is ambiguous: https://drafts.csswg.org/css-contain/#relevant-to-the-user , it only mentions selection, but also links to https://drafts.csswg.org/css-pseudo-4/#highlight-pseudos

More-over, scroll to text fragment seems to have some integration in https://searchfox.org/wubkat/source/LayoutTests/imported/w3c/web-platform-tests/css/css-contain/content-visibility/content-visibility-048.html

So it's not obvious which CSS highlight pseudos should trigger content relevancy. Can we get a clarification on which of these trigger content relevancy?

• ::target-text
• ::spelling-error / ::grammar-error
• ::highlight(*)

cc @frivoal @tabatkins @vmpstr @fantasai

[nt1m]

cc @rwlbuis @delan @schenney-chromium

[vmpstr]

I suspect ::selection is the important one here. The reason for making selection cause content to be relevant is because this is the copy target when the user hits cmd-c or ctrl-c, so even if content is off-screen, it should still be selected with up to date rendering.

It's not as straightforward to say that the element with ::selection pseudo element is the right one, since the highlight pseudos can be parented arbitrarily. This is according to https://drafts.csswg.org/css-pseudo-4/#highlight-pseudos:

The highlight pseudo-elements do not necessarily fit into the element tree, and can arbitrarily cross element boundaries without honoring its nesting structure.

I'm not sure if there's a better spec text to explain what selection means though

[schenney-chromium]

Maybe instead link to the selection API spec? So if JS has selected content the user knows what they are going to get with copy commands.

I think target-text is another case, but again not because of any CSS that's applied, but because the document is going to scroll there (which I guess brings it into the viewport and makes it user relevant regardless).

[nt1m]

So from what I'm hearing, CSS highlights and grammar / spelling pseudos should be excluded? That's fine by me, but would be good if we didn't link to a section that's ambiguous.

[vmpstr]

Maybe we can link to https://w3c.github.io/selection-api/#dfn-selection

[tabatkins]

Yeah, agreed, we shoudl be a little clearer and link to a better target. The Selection API looks good.

[css-meeting-bot]

The CSS Working Group just discussed [css-contain-2] Should CSS highlights make content relevant to the user?, and agreed to the following:

• RESOLVED: Change the line about selection to link to the selection API instead of highlight pseudo element

The full IRC log of that discussion

<dael> vmpstr: In contain2 spec we have text what 'relevent to the user' means. One point is if element or content is selected it's relevent and then links to a highlight pseudo
<TabAtkins> +1 to this
<dael> vmpstr: I think that's misleading. Instead I think it should link to the selection API which goes into details on how a selection acts. I think it's more appropriate
<schenney> +1
<dael> Rossen_: Sounds reasonable
<dael> Rossen_: What do other folks think?
<schenney> q+
<Rossen_> ack schenney
<dael> schenney: Clarification: if there's target text it will be scrolled to and will open up? That's the only other pseudo taht came to mind as potentially useful.
<dael> schenney: If we have target-text from URl target is that considered user relevant already?
<dael> vmpstr: Yeah, that should be. For content-visibility:auto where relevance matters the skipped contents are still available to features like tab order navigation and find in page.
<dael> schenney: Okay, so I think we're good with the proposal
<dael> Rossen_: Other opinions?
<dael> vmpstr: Prop: Change the line about selection to link to the selection API instead of highlight pseudo element
<dael> Rossen_: Objections?
<dael> RESOLVED: Change the line about selection to link to the selection API instead of highlight pseudo element