Meeting minutes
New Issue Triage
jamesn: some DPUB aria issues.
… they're mostly editorial.
pkra1: I commented. Tried to answer, Some editorial issues.
jamesn: aria 2479. ariaNotify
… alison seems to be on it.
… does this need to be agenda'd ?
clay: Alison addressed a lot in comments.
jcraig: have been discussing this separately, too.
… also about extending WPT to cover things. Dominic will likely become more active.
jamesn: great. we can revisit them later then?
clay: yes.
jcraig: +1
clay: I'll keep an eye on them.
jamesn: I had responded. Can we close this?
melsumner: we should point to html-aam more.
jamesn: maybe a generic at top?
melsumner: yes. I'll take it on
New PR Triage
keith: I'll review pkra's
jamesn: and the other one has review
Francis_Storr: there was a question to jcraig or Rahim.
jcraig: I'll look
zakim. next item
WPT Open PRs
jcraig: nothing right now.
… does giacomo-petri's need review?
giacomo-petri: I can try adding more examples.
jcraig: seems good to go with graphics role
… then continue
Deep Dive planning
jamesn: any new deep dives?
… also: do we need the whitespace one at this point?
melsumner: I think we might not need to. I created a project so people can review.
… I can pull that together. We need to know & share this.
jamesn: that sounds good. It helps attract contributions to be async.
Group as allowed acc child of role menu if acc child of menuitem
jamesn: where we stopped last week.
giacomo-petri: the big problem for me is the inconsistency across specs. They seem contradictory.
… menuitem says it needs group with menu or menubar parent.
… but menu doesn't
jamesn: can we change the spec to remove this contradiction?
giacomo-petri: I could try.
jamesn: That would be great.
Should the 'row' role really be necessary for parents of 'gridcell' and other cell role elements? agendabot]
jamesn: in January we had hoped aardrian might find it interesting.
… any chance you could take a look?
aardrian: I'll take a look.
Document interop of misspelled aria-labeledby and its conflict resolution
jcraig: I brought it back to the group because when I originally filed it, I thought engines already did this. Turns out, Firefox does not.
… webkit and chromium do
… jamie and alice questioned if we really want this.
… subsequent issues surfaced that its use has reasonably low usage in data.
… there is some complexity in supporting this in webkit and chromium.
… some concern about future implementations, e.g. shadow boundary
… so even though the group had decided to support this mispelling, we should re-consider with the new / corrected information.
jamesn: maybe a console warning instead?
jamesn: ideally.
keithamus: I'd try a use counter and do an intent to deprecate in Chrome.
aardrian: I see dev typos all the time.
… is there anything we already have that could guide our process?
jcraig: in those cases, we have some alias examples.
… but one point was that we don't support other typos, e.g. title.
… and the complexity with shadow dom, it changes it.
… it's unclear why this change was made in webkit. I think it predated the blink fork. People probably never noticed.
… the complexity of asking Firefox to implement this raises questions, in particular some affecting webkit and blink.
… e.g., accname
… e.g. object reference.
… so my initial perspective was "we already have this, let's standardize" but now see things differently.
<aardrian> I feel bad for jcrag offering the full history. I am on board with a console warning.
scott: not arguing for either but I'd have to file many issues where we found these spellings.
… I just need to know which direction we're going.
jcraig: we could also acknowledge that it's something that exists.
… alice mentioned the trouble working around it.
… but we're not necessarily talking about taking it out.
scott: right, trying to fix these in code bases will be low priority.
<Daniel> The aria-labeled by [sic] discussion (including Jamie/Alice): w3c/
<Daniel> Alice said “…it adds a lot of complexity and potential fragility to the codebases”
jcraig: I liked Keith's proposal.
keithamus: doing nothing doesn't improve the situation but it seems like a waste of resources to ask Firefox to support this.
… I don't think we have an alias that's another spelling.
<melsumner> fwiw there are code linters that already check for invalid aria, and it's pretty common in JS land to alias things so I'm not super worried about this compared to other issues we're trying to solve
keithamus: it will take some energy to deprecate but it seems worth the energy.
jcraig: history fun fact we tried to fix it early but the implementation in JAWS had just finished...
jamesn: so what are we doing?
keith: I pushed a use counter patch for chrome.
… if that's merged, we have some more data.
jamesn: ok. let's revisit once we have some data.
melsumner: also: let's all add this to linter rules.
keithamus: right. and with deprecation we can surface console warnings.
jamesn: sounds good.