Panelset isn't panelset - update accordion.explainer.mdx #759
Conversation
correct some things about what is written here about panelset
bkardell
requested review from
gregwhitworth,
andrico1234,
mfreed7 and
dbaron
as code owners
| It seems likely, however, | ||
| that the consistent use of heading+section markup pattern | ||
| will make it difficult to do good CSS layout of tabs, | ||
| and may also pose other issues with tabs | ||
| with both developer expectations for tab markup and | ||
| possibly assistive technology expectations for tab structure. | ||
|
|
||
| Panelset is a single proposal | ||
| for multiple UI components | ||
| that present things that are semantically headings and sections. | ||
| This differs from the approach proposed here which builds technology | ||
| for just a single presentation. | ||
| The approach of building one thing has clear advantages | ||
| because it is a single control for a single underlying semantic, | ||
| but also presents issues because the different interaction models | ||
| have different accessibility characteristics, | ||
| which creates some difficult problems | ||
| as demonstrated in the work on CSS toggles. |
There was a problem hiding this comment.
Maybe this part should be kept? I think it's still true.
There was a problem hiding this comment.
Maybe... I'm not sure what specifically you mean there. Line 364 'preferreddisplay' thing is from the old proposal and I think if you want to draw a comparison with toggles. The preceding 6 lines, I believe, cover the relevant bits. The different a11y characteristics of each are handled in the panelsets demo as well. If there is a problem there it is well expressed in my opinion in the linked #559 which is, as you can see, complicated.
re: 366, if you look on https://tabvengers.github.io/spicy-sections/demonstration/#tabset-state-list on a wide enough screen you'll see tabs using this. It's not just CSS, it's minimal shadow dom. I don't think styling is an issue - @mirisuzanne or @jonathantneal can perhaps agree or deny, but we had several examples. Yay shadow dom?
I thought leaving in the
The panelset proposal has enough open questions that it's a bit hard to assess
and linking in that issue kind of covered it well enough.
But if we want to compare them, I'm not sure if we should compare the whole thing vs the affordance part? That is, there are two levels to our proposal, an attribute and the css property. The reasons for the css property seem to apply equally well to this proposal which just sidesteps it and says "use some JavaScript". That seems an equally easy change to make for now to ours, so let's leave that aside and compare just ... the markup? Because I mean, if we are changing attributes then some characteristics might change on summary details with name too, right?
Sorry I'm not sure how I missed what was actually in 748 when merged, but there is some real confusion there that is probably easier to correct in a PR and a comment.
This PR removes references to the old 2015 proposal which was also called panelset. I'm sorry 😆
While there are some surface similarities in that what was defined was a general structure on which several views could be laid, they are in almost every practical way very different things and thus most of the content/critique in there is not talking about the actual work this group did. That work was preliminary referred to as "spicy-sections" because: a) it followed hixie's original idea in the HTML spec of just putting a marker element around some sections b) we didn't want people to take it too seriously yet or get into bikeshedding names.
Eventually we did bikeshed names, and (not because I advocated it, I didn't) people chose 'panelset' again
Sorry. I hope this PR helps, I'm happy to help answer questions or fill it in or review a different request instead.