You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
associationlist and related roles have tediously long names #1662
associationlist and related roles have tediously long names. These look like ObjC naming conventions: associationlist:item:value: ;-)
associationlist: A section containing associationlistitemkey and associationlistitemvalue elements.
associationlistitemkey: A single key item in an association list.
associationlistitemvalue: A single value item in an association list.
I don't have a particular preference other than "not what's there" but in the interest of forwarding the discussion, I'd suggest the following:
Remove associationlist entirely. I don't think there needs to be any distinction between list and associationlist in the Accessibility APIs.
listitemkey and listitemvalue may be sufficient. It's possible listitemvalue could be an alias of listitem… I'm not certain, but the existence of a key may be all that's needed to distinguish between standard lists (~html:ol/html:ul) and these (~html:dl)
It would simplify matters to permit the key/value roles inside list, whatever they may be called. (I agree the currently suggested names are too long).
What happens if people use key & value elements and plain listitem elements in the same list? Is that a problem?
If this is about parity with HTML, it should not be allowed, but I can imagine cases where it would be useful to do so. (e.g. if there is a placeholder listitem for an as-yet-unknown key-value pair).
I have a slight preference for keeping listitemvalue (or whatever name we come up with) separate from listitem because of the different (possible) structure of 1 DT followed by multiple DDs.
@aleventhal discovered we have some leftover expectations related to this, including invalid computedrole and invalid name-prohibited computedlabel tests
The current term and definition roles used for dt and dd in html-aam are incorrect per HTML's defined semantics, so it would be good to update them to something else. (See w3c/html-aam#347)
Activity
pkra commentedon Dec 16, 2021
Related: #961
smhigley commentedon Dec 16, 2021
+1 to the first option
medeaoblongata commentedon Feb 1, 2022
It would simplify matters to permit the key/value roles inside list, whatever they may be called. (I agree the currently suggested names are too long).
What happens if people use key & value elements and plain listitem elements in the same list? Is that a problem?
If this is about parity with HTML, it should not be allowed, but I can imagine cases where it would be useful to do so. (e.g. if there is a placeholder listitem for an as-yet-unknown key-value pair).
pkra commentedon Apr 28, 2022
I'm in favor of the OP.
I have a slight preference for keeping
listitemvalue(or whatever name we come up with) separate fromlistitembecause of the different (possible) structure of 1 DT followed by multiple DDs.scottaohara commentedon May 5, 2022
@jnurthen going to add myself here as a second assignee. we can discuss / i can do the associated HTML AAM stuffs.
associationlist and related roles have tediously long names
jnurthen commentedon Oct 10, 2023
Lets discuss 3 options:
cookiecrook commentedon Oct 16, 2023
+1 for removing association lists [entirely]
jnurthen commentedon Oct 26, 2023
As discussed in https://www.w3.org/2023/10/26-aria-minutes.html#t07 will remove for now but may re-add a subset later on
cookiecrook commentedon Aug 15, 2024
@aleventhal discovered we have some leftover expectations related to this, including invalid computedrole and invalid name-prohibited computedlabel tests
editorial: update roleInfo.js (manually) (#2624)
zcorpan commentedon Sep 24, 2026
The current
termanddefinitionroles used fordtandddin html-aam are incorrect per HTML's defined semantics, so it would be good to update them to something else. (See w3c/html-aam#347)