Meeting minutes
matatk: The group discussed the ARIA meeting, and impasse there. Also the WHATWG and CSS meetings, and possible way forward with multiple attributes being exposed as one in the accessibility tree.
PaulG: ACK re attributes approach
mhakkinen: The multiple attributes approach was discussed a while back, and taken as too much of a lift for tools that parse the DOM directly, rather than use the accessibility tree.
APA TPAC planning page (mentions Pronunciation a couple of times): https://
ARIA minutes: https://
mhakkinen: Has our technical approaches doc been shown to WHATWG?
matatk: Can check the minutes; we can show it to them if/when we request attributes.
matatk: CSS thought pronunciation was not a thing that should be in CSS, and said there were no developments on CSS Speech in the past year.
PaulG: Sounds like the path forward is likely going to be single or multi-attribute approach. We will still have to work out the benefits. If we're going to convince some platforms to come along, we need to figure out how to make it easier for them.
PaulG: [ suggested an opt out in some pathways, to hand over to the Web Speech API, if they want to opt out ]
PaulG: Thinking about the Custom Element POC I made when joining the group. Wish we had a way to bridge to authoring in HTML.
… There are discussions at the moment about supporting ARIA across shadow DOM boundaries. That could help.
mhakkinen: Custom Elements, or custom attributes?
PaulG: Custom Elements (but attributes could also be used for this).
… The POC I posted to Glitch had a short script that took the markup that was Custom Eements and turned it into SSML and fed it to the Web Speech API.
… That is the gist of what we're trying to do, but taking a round-about way to get there, and probably falls short of what we've been talking about for the past couple of years, but don't want to take anything off the table.
… If people are picking up Custom Elements again and talking about how they can work with accessibility, it may be worth exploring.
matatk: (1) there was some discussion of crossing shadow DOM in WHATWG meeting IIRC; (2) AOM develoment is picking up; (3) TPAC reminded me of the power of prototypes and how important and persuasive they are.
PaulG: The POC approach could be treated by some ATs as if it were just SSML. We are not asking to change rendering, but may have to accept not having SSML in the browser. The challenge is that even if we don't have SSML as first-class, there may be gaps in the accessibility tree.
<mhakkinen> FYI, we do have videos with demos created for previous TPACs: [1] https://