Meeting minutes
<janina> Date 05 Aug 2026
<JonCohn> open the zoom link
Agenda Review & Announcements
zakiom, next item
TPAC 2026
re: accessibility and privacy, some features which were proposed and could have benefit were not taken up because the fear of privacy risks
matatk: are there any features that are vital enough to accept the privacy risks?
… we want to hear about those features so we can organize their architectures into any themes that might be present to help build them in a way that better respects privacy concerns
… I will make the breakout proposals
… there are 9 proposed and we expect an influx of accessibility-related ones
… last week we talked about accessibility and society mini-workshop and I need to figure out the process for suggesting a lightning talk to them
janina: we'll get our proposals in early but details will continue to flow until late
New charters review
matatk: web apps security charter
Web App Sec WG
<matatk> w3c/
<Roy_Ruoxi> review history
matatk: we haven't reviewed this group's work very often. However, there's a tie-in to TAG work and well-known URLs, this group has 2: password and passkey
… there was a lot of discussion in the last TPAC about migrating between devices, credentials, and use cases about the tech and security of that. And we have guidance about the accessibility of the authentication portion of those cases.
janina: I think we're signing off without objection
New on TR
CSS Linked Parameters Module Level 1
<Roy_Ruoxi> w3c/
<Roy_Ruoxi> https://
<matatk> PaulG: It's a cool spec that will allow parameters from CSS to be injected and flow into SVGs. It can even penetrate an <img> tag that is hosting an SVG.
<matatk> PaulG: So you can do stuff like alter colours, sizes, etc. If you work on design systems you'll love this.
PaulG: My take is this is going to extend capabilities we already have. You wont' need to have as many assets; you'll have one that CSS can modify.
PaulG: The only thing I'd be looking out for is if they want use cases - that could inspire devs to pick it up and use it. Maybe symbols as you say janina, and or high contrast and symbols. Worth making a demo.
matatk: +1 to what PaulG said, adding use cases is a specialty of ours. I believe TAG is reviewing this. This is clearly something that may not specifically add end-user stuff but it makes building so much easier that it could help accomodate more use cases.
Spec review requests
EPUB Accessibility 1.2 Explainer
<Roy_Ruoxi> https://
<Roy_Ruoxi> https://
matatk: I'm reminded that we have the proposal for ..
matatk: we did just review some epub items and there were a couple of key issues being addressed--fixed layout
<gpellegrino> https://
gpellegrino: epub a11y 1.2 is CR defines what it means to have accessible epubs. Essentially, following wcag and have page numbers and time text media
… the explainer was requested by the community to help people understand wcag's relationship to epub. For example, authentication isn't something I've never seen.
… others like multiple ways may be tricky to implement
<Zakim> matatk, you wanted to add other specs in to this topic before we move on (feel free to jump ahead in the queue if you want to talk about this topic)
- 2026-07-14: Digital Credentials
- 2026-07-29: Attribution Level 1
- 2026-07-31: CSS Image Animation Module Level 1- 2026-07-14: Digital Credentials
- 2026-07-29: Attribution Level 1
- 2026-07-31: CSS Image Animation Module Level 1
Issue tracking
matatk: there need to be 2 MCP comments that we get into shape
… one of the questions is about the axtree use by MCP and we aligned with others saying "no"
… we still need to post the comment but the other question was relating to the accessibility considerations which is a section to be drafted as a general "what do we think" comment.
WebMCP: accTree topic
draft comment: w3c/
matatk: I think PaulG comments need to be raised in the general comment. This issue will reflect the initial concern about axtree redundancy.
matatk: we're worried about abuse of the axtree. Authors stuffing keywords meant for agents.
… making MCP additive fits well. I can add that. But I don't want authors to fiddle with aria just to signal agents.
janina: semantics are useful and if some values need to be stuffed, just make your own tree. Don't get in the way of the accessibility by trying to overfit the existing structure for a different purpose.
matatk: I'll make those edits in the next couple of days.
WebMCP overall proposed APA comment
w3c/
matatk: largely it says we're excited for the possibilities of this technology. However, we're concerned about the API in its current form.
… the declarative version of the API looks like a better starting point and we'd like to see that experimentation. We don't want it to cause interference with the axtree. We'd like to see the form submission behavior to allow review by the user.
… please review that comment and put your feedback on the thread.
https://