Meeting minutes
New Issue Triage
spectranaut: scribing is fun!
New PR Triage
Deep Dive planning
Deep Dive next week is about popups.
TPAC Agenda - add tag F2FCandidate by August 16th
Siri: Who will attend TPAC
spectranaut MichaelC JamesN Bryan and jcraig will attend in person.... possibly Siri
CurtBellew can only attend remotely
spectranaut: any remotes?
Scott, mark, Peter, and Jaunita
spectranaut: GitHub tag created for F2F candidates. please add. Will discuss on the 16th
"Whitespace characters" underspecified
spectranaut: this is a blocking issue for 1.2
JamesN researched; discovered this was a formal decision by ARIA in 2016?
<spectranaut> 2016 decision on aria-roledescription: https://
<Cyns> Partial regrets, hoping to join later
What def of whitespace is being used in WebKit?
jcraig: will look it up
spectranaut: how to triage since this is blocking 1.2 and this was a formal decision previously
jcraig: motion to triage to 1.3 since this is late in 1.2
scotto: NVDA will suppress role descriptions using this technique on Windows; JAWS ignores and speaks it anyway
aaron: my concern is that considering all the ways authors can go wrong will be a pain for us.
jcraig: right. I don't see how it blocks 1.2
aaron: could just say "space" and figure out in 1.3
jcraig: we still need to verify implementations
aaron: we're currently implementing it
jcraig: underspecified leaves potential for implementers to do what they think best in the meantime
aaron: I'm mainly concerned about overcomplicating the code.
… were implementors invovled in the original decision?
… feels like not
spectranaut: if AT can still decide what to do (since it's exposed separately), then there's no real loss.
… I'm moved by the argument that it's an author error and AT can still decide what to do with it
… but we want to focus on getting 1.2 out the door.
… we should work out where else we refer to whitespace in our specs.
… e.g., aria-braille
aaron: Firefox on Mac seems to only check length 0.
spectranaut: in ARIA spec, whitespace is only mentioned in braille props.
scott: but there must be something in, e.g., accname.
jcraig: what if role=button and roleDescription is empty string.
… implementation might want to invalidate use of role.
… e.g., if it's not focusable.
aaron: if we find out safari and firefox just check empty string, can we change the spec?
michael: it would be good to match both specs. but also good to put changes until later
spectranaut: so we need to know what webkit does.
… and work on braille props.
pkra: braille props were following the regular ones.
… so I'd suggest to keep doing that
spectranaut: makes sense.
jcraig: that's probably fine
spectranaut: so perhaps we can decide on the empty string. At least we should not slow down the release of 1.2
… of aria
spectranaut: how about a formal decision to change it to empty string.
aaron: +1 we'd have to update Chrome.
scott: this way, we're not cutting off the discussion for later.
… I'm concerned about people actively suppressing roles
spectranaut: I'll want jnurthen and Firefox's opinion on this.
aaron: james teh commented on the issue
spectranaut: right, he wrote it sounds like an authoring error.
spectranaut: I'll talk to jnurthen and suggest a formal decision on the mailing list.
1.3 triage
spectranaut: first step: going through all issues in the milestone assigned to you.
… tag them with a few different tags as per my email.
… all: reminder to take a look