W3C

– DRAFT –
ARIA Editors

27 November 2023

Attendees

Present
-
Regrets
-
Chair
-
Scribe
spectranaut_

Meeting minutes

synonyms

summary of discussion so far: we should link from one role section to the other

summary of current discussion: how easy/hard should it be to find the synonym we don't want used?

issue for this topic: w3c/aria#2073

mcking: the non-preferred section links to the other

mcking: in characteristics table, where name is prohibited on role `none`, do we also role `presentation`?

mcking: how does this effect aria.js

jamesn: would the easier way to do this by adding an extra script to post process

jamesn: to fix the few links

mcking: role reference to synonym to instead of PR

spectranaut_: if you link to the synonym, if will just redirect you to the other role

spectranaut_: so it doesn't seem like a problem to me

PR/Merge process - next steps after deep dive?

spectranaut_: I was going to update the process document with more clarity on the process, re james craig's confusion

mcking: also we need to discuss next steps for the monorepo, so resolve james craig's concerns

spectranaut_: anyone have experience with combining git repos

(resounding silence)

mcking: but we don't want to combine issues, also, there is the challenge with the history

peter: you can merge repos and keep the histories

peter: they are working on different files/subfolders

spectranaut_: we were discussing keeping the other repos open for issue tracking, alone

peter: you can have a PR that closes issues in a different repo

peter: I'm worried that we have repos that we don't officially maintain, like svg-aam and dpub-aria

peter: we can't just put those into our monorepo

mcking: most of our issues are only related to core-aam and html-aam and accname

peter: james wants everything done in one PR

spectranaut_: if we build a pr-preview for a PR that has changes to core-aam and aria, will the links between the two of them be links to the pr-preview? I think not

peter: right, interlinks won't work, but PR preview will link to all of the specs revised by the PR

mcking: sounds feasible to make those links work

jamesn: maybe, but it might be too hard

mcking: right now we uses classes for those links to accname. so if you are linking something in accname in the ARIA spec... it depends on how you build the preview

jamesn: we don't build it, we use pr-preview, respec runs, and the code to generate links happens in there

james: we want to remove the special linking that we do, the way we link between specs, there is a ton of code to make sure the editor drafts link to editors draft

jamesn: those concepts don't exist. maybe we should just use xref, and export our definitions, we can use standard respec

jamesn: if we use xref, the standard, then we have no way to interlink the specs

peter: is there agreement that if we at least have preview links, that would be a sufficient first step?

mcking: I don't know if it would resolve the issue james craig has for implementors

mcking: you can't just send the implementor this one link to this preview

mking: if that aria actions preview has a link to core-aam in it... you can't use the link to core-aam, you have to hunt down the relevant information using this preview. it sounds very complicated.

jamesn: it sounds like need to talk to james and see if it will solve his problem, if it doesn't, we need to investigate more

mcking: doesn't seem like there is a ton of value in a mono repo other than having a single PR, doesn't add any value other than that

jamesn: it allows someone to review the PR with all the changes in one place. I think it makes the review and maintenance processes easier.

mcking: so there is a huge advantage to us, the working group members

mcking: maybe some advantage to the implementor

roll out prettier setup - tracked via aria-common#99

peter: should i merge these?

jamesn: yeah it is scary

ISSUE: w3c/aria-common#99

peter will open a PR on aria-html

jamesn will review the change on svg-aam

retiring contributors.md across specs - tracked via aria-common#103)

<pkra> w3c/aria-common#103

matt garish doesn't like it for dpub

jamesn: he is welcome to maintain his own

peter: it's fine if they don't want to use our automated stuff....

jamesn: I think it's possible to only add stuff after a certain date... I could make an enhancement to respect to do that

jamesn: for example, you would want, from this version forward

spec markup for advice for AT (jnurthen)

to bug james :)

modernizing aria.js (aria-common#104)

peter: after talking to james, I created the issue... I thought I'd make some progress but I haven't

peter: I'm going to start refactoring just to make it more readable.. so intertwined... hard to understand how things work

ISSUE: w3c/aria-common#104

peter made notes in the issue

peter: I have a dream where we have less of this data extraction and html stuff

peter: cleaner more understandable code

jamen: we know what the logic should be

jamesn: we start with list of roles, a hierarchy, assign states and properties as we go down the hierarchy

peter: some things like alert, there are no local props, it gets them through parental properties -- on structure and role type, its hard to compare whats currently in the role info.js info file and what the spec would generate...

(scribe is not really catching this)

peter: it is fun!

Summary of issues

  1. w3c/aria-common#99
  2. w3c/aria-common#104
Minutes manually created (not a transcript), formatted by scribe.perl version 221 (Fri Jul 21 14:01:30 2023 UTC).

Diagnostics

Succeeded: s/monorepos/combining git repos/

Maybe present: jamen, james, jamesn, mcking, mking, peter, spectranaut_

All speakers: jamen, james, jamesn, mcking, mking, peter, spectranaut_

Active on IRC: pkra, spectranaut_