ariaNotify tops an accessibility-heavy Interop 2027 ballot

Voting for Interop 2027 closes on 19 November 2026, and one developer has published a fifteen-item ranked ballot that puts ariaNotify() at the very top. The ranking was posted via Firefox's Interop ranking page — a tool Mozilla built for the second year running to let anyone order their preferences for the annual cross-browser interoperability effort.

The author's reasoning for the top slot is blunt: years of watching developers get ARIA live regions wrong, coupled with persistent confusion about how live regions are exposed and supported across browsers and screen readers. The ARIA Working Group has invested significant effort in specifying ariaNotify() in the W3C ARIA specification, evolving it from what the author calls an early, odd proposal. The hope is that a simpler notification API will close gaps that live regions have left open — gaps the author partly attributes to companies not employing screen reader users in critical roles.

What fills out the top five?

• HTML focusgroup attribute (number 2): better keyboard outcomes for users and less scripting for authors.
• Shadow DOM referenceTarget (number 3): the author has seen broken label and name associations across shadow boundaries leave screen reader and voice control users stranded, while ARIA reflection and the proposed AOM remain unavailable today.
• HTML headingoffset and headingreset (number 4): headings from user-generated or third-party content are often the wrong level, though rarely full blockers. The fear is authors will get cavalier and treat this as a document outline algorithm.
• Device Posture API (number 5): foldable phones have existed for years, but Apple's newest hardware is already breaking design systems that rely on fixed screen sizes — a mistake the author says limits user pain caused by author error.

The Viewport Segments API lands sixth for the same reason, "with the added bonus of numbers."

Which author-facing features made the cut?

CSS dominates the middle of the list. Linked parameters would let authors pass colors into SVGs referenced via <img>. A change making the CSS AccentColor system color reflect accent-color would repair a break that left design systems stuck with low-contrast browser defaults after fingerprinting concerns killed the user-preference pattern they'd leaned on.

The line-clamp property addresses years of card designs that visually clip text or relocate interactive controls, confusing keyboard and screen reader users — though the author is wary of CSS generating content because of localization risk. random-item() would remove scripting, with a caution that authors might misuse it in ways that break accessibility API exposure. Anchor positioning with transforms and animations gets a personal endorsement: the current script payload "seems to cause my phone to overheat and my fans to spin." flex-wrap: balance would fix what the author calls a glaring, years-old gap in flexbox.

HTML proposals include <input type="week"> — motivated by the sheer number of poorly coded, inaccessible date pickers the author has seen — plus responsive iframes, flagged as an accessibility problem once critical content gets visually hidden and scrollbars turn embeds into a functional mess. Pointer events interoperability on touch devices closes the list, justified by the news that Apple is finally building a touchscreen Mac; the author predicts Apple will ignore what it learned on mobile and "probably invent new interactions because it knows better."

Who else is voting?

Eighteen other developers have published rankings, listed alphabetically: Jeff Bridgforth, Declan Chidlow, Joshua Comeau, Chris Coyier, Ky Decker, Roland Franke, Paweł Grzybek, Jeremy Keith, Christopher Kirk-Nielsen, David Marland, Eric A. Meyer, Stuart Robson, Ana Rodrigues, Remy Sharp, Sam Smith, Sara Soueidan, Tyler Sticka, and Cyd Stumpel. The author notes the list is, so far, exclusively people with last names — "kind of doubling down on the Western Wide Web phenomenon." There are also plenty of items the author would happily down-vote if the process allowed it. With the deadline three weeks out at time of writing, those ballots will help decide which of these features browsers commit to testing together in 2027.