W3C

– DRAFT –
(MEETING TITLE)

15 July 2026

Attendees

Present
Dr_Keith, JoncCohn, matatk, Neha, PaulG, Roy_Ruoxi
Regrets
-
Chair
Matthew
Scribe
matatk

Meeting minutes

title: APA Weekly Teleconference

Agenda Review & Announcements

Web & AI IG

https://www.w3.org/2026/07/14-webai-minutes.html

matatk: We had a key question: what does good AI training from an accessibility perspective look like?

TPAC 2026

Reg is now open: https://www.w3.org/news-events/tpac/2026/

Funding info & application for IEs: https://www.w3.org/news-events/tpac/2026/registration/#registration-fee-waiver

New Charters

WebAppSec

w3c/a11y-request#168

matatk: Maybe good potential to improve experience for PWD, especially cognitive. Will need to keep an eye on it.

PaulG: They have a lot of deliverables - lots for any advice we give them to permeate through.

New on TR

Verifiable Credential Forgery Defense v1.0

https://www.w3.org/TR/2026/WD-vc-forgery-defense-1.0-20260630/

matatk: Low level

PaulG: Agree

SHACL 1.2 Profiling

https://www.w3.org/TR/shacl12-profiling/

matatk: There are i18n considerations - so should there be accessibility ones? SHACL UI seems interesting but not sure how relevant.

PaulG: Also trying to get my head around SHACL UI

Spec Reviews

WebMCP

<PaulG> (possibly relevant https://github.com/webmachinelearning/webmcp/issues/65)

Declarative WebMCP: https://webmachinelearning.github.io/webmcp/#declarative-api

Explainer for declarative part: https://github.com/webmachinelearning/webmcp/blob/main/declarative-api-explainer.md

matatk: Declarative is tied to form filling where imperative is not. Brings some concerns but has some opportunities. Should be done first?

PaulG: Imagine a workflow which is separate steps, each step is a form. For accessibility, if I've hooked up the WebMCP, it's aware of them all even if I'm only on one stage. Say I'm on stage 3, and the UI is problematic, but I know there's a way to get to my history...
… Instead of navigating away from the workflow to history, and back, I could ask my agent to surface the workflow history. This could be done in a separate window/tab/siderail/extension. It's augmenting a somewhat lacklustre UI with some real-world help.
… If it's all declarative under the hood, then the 'functions' may be learned/advertised by the agent. There's definitely some legs here for this.
… I remember 'web services' being explained way back... it's just another web page but it's for the machine and it's used to construct a query that is sent to the database. It's exactly the template the server is expecting, maybe with some guiderails.
… This is doing the same thing - giving shape to the input call.
… With that we can probably make more use of things like HTTP Query (which is new). The server understands these new things but the UI doesn't (yet).

matatk: Anything new based on this for our comment?

PaulG: I think the declarative approach has a better chance of _not_ creating two lanes, where there's a disparity between the capabilities available to AT and those avaiable to agentified UAs.
… The declarative approach is just HTML with some extra attributes to indicate this is a tool. This is available to AT, all users, and the agent.
… If we have purely the imperative structure done first, that's only available to one modality, the agent. Worries about creating a 'class' system as a result.
… Also drift between what the UI user and the agent are capable of.
… I think this will leave fewer people behind.

<Neha> +1

CSS Image Animation Module Level 1

https://github.com/w3c/a11y-tracking/issues/333https://github.com/w3c/a11y-tracking/issues/333

w3c/a11y-tracking#333

Related 'web platform gaps' issue: w3ctag/gaps#15

PaulG: + page transitions to that list in gaps#15

Minutes manually created (not a transcript), formatted by scribe.perl version 248 (Mon Oct 27 20:04:16 2025 UTC).

Diagnostics

No scribenick or scribe found. Guessed: matatk

Maybe present: title

All speakers: matatk, PaulG, title

Active on IRC: Dr_Keith, JoncCohn, matatk, Neha, PaulG, Roy_Ruoxi