W3C

– DRAFT –
ARIA deep dive summary element

27 July 2023

Attendees

Present
Adam_Page, jamesn
Regrets
-
Chair
-
Scribe
jamesn

Meeting minutes

Scott: goes over a brief history of the summary element and some of the issues in HTML

scott: what can we do to expose this consitently and home up with a pathway forward

html doesn't want to make changes to the content model

jcraig: is the reason to talk about this in context of role parity

scott: 1st can we do this consistently? then - would it be existing or new

jcraig: so we need a pushable or clickable with open/closed state

matt: 1st respond - tab is all manually done. there is no implicit relationship. Could just as easily use button - no taxonomy changes needed

matt: could be useful for us to have a list of the distinct problems

<jcraig> parity role should be similar to a clickable with a expanded/collapsed state.... tab or toggle button seem most likely: disclosure -> *[role=button][aria-pressed]

matt: a clickable element needs to be exposed and think it is reasonable to assume that that clickable element itself isn't the element with descendant content. and some way of naming it needs to be specified

matt: can't think of any place in aria where you have an interactive element with nested interactives

matt: just a problem as can't have nested interactives

<jcraig> remaining problem with mapping "disclosure -> *[role=button][aria-pressed]" is the children presentational issue with button.

jamesn: want to allow to solve the other nested interactive cases too

mu-an: as someone who uses the details elements - lots of places have disclosure removed - resetting those is a bit of a mess. to expose interactive content - if the triangle has been removed the interactive would be gone

scotto: in reviewing the a11y of these things - there are some issues with some browsers where if the marker is removed it can end up being a nothing element without it

scotto: whatever we end up doing we need to makes sure that css is not required and if the marker is removed it should not make it inaccessible

melanie: are we asking if we would like to more uniformly announce summary - want to get scope right here

scotto: one thing would be good to solve would be a single way that these are implemented

in some browsers would navigate by buttons and in others not

<jcraig> This is the article melsumner was referring to: https://melsumner.github.io/details-as-a-menu

discussion around bending these elements - usage has been expanded over the years. find in page will expand details elements

can create silly user experiences when using for dropdowns for example

jcraig: remap cases where summary -> menu would be uop to the author to provide the actual mapping

jcraig: initial comments - forgot to mention that disclosure could map to button or tab but the key problem with button is children presentational baggage. Links for example can have nested interactives

mario: can have an exposure of the summary as for example the summary role and tabbing would be the first thing - disclosure button could name that with heading or content

https://codepen.io/scottohara/pen/RwqQgML

<jcraig> nit: if we go Mario's direction, I prefer “disclosure” as a new role name over “summary”

<melsumner> +1

mario: not possible to say all of summary is exposed as a button - as can have interactive content within it

<melsumner> casual comment: The children of interactives are presentational. Genuinely don't want that to change. Maybe I need to make a t-shirt "no nested interactive" ;)

matt: like the general direction - does skirt the children presentational issue. don't know if it is really a problem

matt: stilll wondering about a use case - where the button that doesn the expanding and collapsing and essentially labels the discolsure - where that button is wrapped in a heading. can put a heading inside the summary but if want to create something more like accordions in APG - can then press enter to expand the content

scotto: regardless there will be work for the browsers

scotto: since html doesn't want to change the content model means there will be needed changes in the browsers

scotto: allowing headings to wrap them could be allowed

matt: could put anything in the summary - seems counter intuitive

Minutes manually created (not a transcript), formatted by scribe.perl version 221 (Fri Jul 21 14:01:30 2023 UTC).

Diagnostics

No scribenick or scribe found. Guessed: jamesn

Maybe present: jcraig, mario, matt, melanie, mu-an, Scott, scotto

All speakers: jamesn, jcraig, mario, matt, melanie, mu-an, Scott, scotto

Active on IRC: Adam_Page, jamesn, jcraig, melsumner