Adrian Roselli, pH0

1/2

A few years too late, but ARIA Authoring Practices Guide finally has (only) four support tables:
https://www.w3.org/blog/2023/04/answering-what-aria-can-i-use/

Mind, a year ago it re-branded as a copy-ready pattern library without any data on support.
#a11y #accessibility

Answering “What ARIA can I use?” | W3C Blog

Apr 17, 2023 at 6:05pmWeb
Show threadAdrian Roselli, pH0Apr 17, 2023

2/2

Those four patterns are generally safe, too (button, link, alert, radiogroup).

However, the AT support tables:
• are all listed as “Unapproved Report”;
• only list desktop browsers/SRs;
• do not include Firefox.

So, IMO, worth asking for input but not calling a win.

Show threadAdrian Roselli, pH0Apr 18, 2023

3/2
I agree with Eric:
https://yatil.social/@yatil/110218482529480904

Not super useful as a go / no-go.

Eric Eggert (@[email protected])

@[email protected] @[email protected] I like how those tables are made for people who already know, too. “Toggle buttons are supported 63% in Safari/VO macOS.” is effectively non-information for implementers. It does not answer the question of “can I use this pattern?” (And of course you might not even want to use some of the patterns over native HTML or hybrid solutions.)

yatil.social
Show threadAdrian Roselli, pH0Apr 18, 2023

4/2

Not a “sea change”. That statement is pure hyperbole.

It is literally a year too late (4 SRs & 1 browser short) for something that announced it was an accessible pattern library (which it is not) last GAAD.

• https://w3c.social/@w3c/110220013802529659
• https://twitter.com/w3c/status/1648315190373433345

*grumble*

World Wide Web Consortium (@[email protected])

W3C blog: "Answering 'What ARIA can I use?'" by Matthew King “Assistive Technology Support” tables in the ARIA Authoring Practices Guide show how 3 screen readers support 4 UI pattern implementations represents a sea change in accessibility engineering https://www.w3.org/blog/2023/04/answering-what-aria-can-i-use/

w3c.social
Show threadAdrian Roselli, pH0Apr 28, 2023

5/2

I missed this other write-up:
“‘Can I Use…’, but for ARIA!”
https://bocoup.com/blog/can-i-use-but-for-aria

How is knowing NVDA 2021.3.5 (two years old) supports something at 96% useful? Does the 4% the mean no keyboards?

What about my NVDA 2023.1.0.27913? Or that I use it with Firefox?

‘Can I Use…’, but for ARIA! - Bocoup

Web Platform Consulting Services

Show threadEric EggertApr 28, 2023

@aardrian I want to tear everything down. This is so useless. Nobody understands why Can I Use? is useful and beloved. (Hint: It’s because you get Yes/No answers. Not 96%. “Can I eat this?” – “96%” – “Yeah, maybe not.”)

ARGH.

Show threadAdrian Roselli, pH0Apr 28, 2023
@yatil I would not mind if these announcements were framed better. Essentially, “Hey, we are finally trying to gather some limited notes on support.”
Show threadEric EggertApr 28, 2023
@aardrian Yeah. Or if there was a summary like “For most use cases the button role is well supported, just be aware of X.” But just saying X out of Y tests work is not useful. And then sell it as if they just invented sliced bread.
Show threadAdrian Roselli, pH0Apr 28, 2023
@yatil Sliced bread? One step away from toasts!
Show threadEric EggertApr 28, 2023
@aardrian Well, toast is sliced boring bread… which… also fits.
Show threadPatrick H. LaukeApr 29, 2023
@yatil @aardrian it's infuriatingly pointless. i don't doubt that whoever was involved with it worked really hard or whatever, but yes the conclusions/framing of results is supremely pointless. why not actually present it like ... caniuse do for browsers? instead of fucking numberwang https://youtu.be/0obMRztklqU
Numberwang Episode 1

YouTube
Show threadPatrick H. LaukeApr 29, 2023
@yatil @aardrian is this supported in browser X with AT Y? yes, no, partial - and for partial, clearly list what currently fails/is missing. it's not rocket science
Show threadPatrick H. LaukeApr 29, 2023
@yatil @aardrian just expand the approach and presentation of https://a11ysupport.io/ basically...
Accessibility Support

Show threadEric EggertApr 29, 2023
@patrick_h_lauke @aardrian Yeah, that’s so much more useful. The other information that can I use has are links to tutorials, work-arounds and polyfills. So it also answers the question “it’s not supported in all browsers, what can I do to start using it now and make it work everywhere.”
Show threadEric EggertApr 29, 2023
@patrick_h_lauke @aardrian A lot of that has to do with the W3C mantra of “vendor neutrality”. They are just not allowed to show preference or that one browser/screen reader/tool is better than another. That is why the “Web Accessibility Evaluation Tools List” lists everything and puts the onus on visitors to make distinction based on data. It’s essentially useless.
Show threadPatrick H. LaukeApr 29, 2023
@yatil @aardrian wonder if it's worth me just crafting a diplomatically worded (HAH) issue on their github repo...just to at least put it down "on paper" how currently the resource is less than ideal and what would *actually* help authors. sure, it'll be ignored or "we'll look into it at some point" ... but at least it's there for reference
Show threadEric EggertApr 29, 2023
@patrick_h_lauke @aardrian I mean, you can try. I have given up, to be honest.
Show threadPatrick H. LaukeApr 29, 2023
@yatil @aardrian just to put it on record...i think i might. then in future i don't have to repeat myself and just point to the issue (like i do with most #WCAG things that are borked...sigh)
Show threadAdrian Roselli, pH0Apr 29, 2023
@patrick_h_lauke @yatil Can you link it here when you do? I am partway through a blog post putting a stake in the ground that this resource is not what they are advertising.
Show threadEric EggertApr 29, 2023
@aardrian @patrick_h_lauke Ah, I guess I can then check off my todo. 😜
Show threadAdrian Roselli, pH0Apr 29, 2023
@yatil
Meh. We may have different approaches. Mine will not be done today anyway; gotta get to a book fair!
@patrick_h_lauke
Show threadEric EggertApr 29, 2023
@aardrian @patrick_h_lauke Oh, mine is under someday because it’s Snooker World Championship! Priorities!
Show threadHiddeApr 29, 2023

@patrick_h_lauke @yatil @aardrian please do!

For me the most important goal would be that a web developer can easily find and understand if they should or should not use a pattern. Not because they deserve an easier life or anything, more because it would make it less likely they do right things (as in build accessible/accesssibility supported UIs)

Show threadAdrian Roselli, pH0Apr 29, 2023
@hdv
I maintain that if the support charts are robust and accurate they will help keep devs from implementing many of the APG patterns as written.
@patrick_h_lauke @yatil
Show threadHiddeApr 29, 2023
@aardrian @patrick_h_lauke @yatil yep and that's a good thing!
Show threadEric EggertApr 29, 2023
@hdv @aardrian @patrick_h_lauke It’s probably a bit of too much of a conspiracy theory, but the reason that this testing takes so long might be that it would show how few of the “patterns” are properly supported…
Show threadHiddeApr 29, 2023
@yatil @aardrian @patrick_h_lauke I (and you) seen plenty of other ways slowness is embedded in every step of the process in this organisation (often for good reasons, sometimes not so much)
Show threadDennisLApr 17, 2023
@aardrian It’s at least a good start. I know someone on the inside who is doing the best he can. And as you know, anything with the W3C is a long, painful, and political process.
Show threadAdrian Roselli, pH0Apr 17, 2023

@dennisl A better start would not have been to re-position APG as a pattern library (outside of W3C process AFAIK) when it had no support data at all (other than all the open issues pointing out lack of support).

So yeah, progress is good but still too late for me to give it a pass.

Show threadEric EggertApr 18, 2023
@aardrian @dennisl I like how those tables are made for people who already know, too. “Toggle buttons are supported 63% in Safari/VO macOS.” is effectively non-information for implementers. It does not answer the question of “can I use this pattern?” (And of course you might not even want to use some of the patterns over native HTML or hybrid solutions.)
Show threadJoe HumbertApr 18, 2023
@yatil @aardrian @dennisl As one of the people who has supported this group since it formed in 2018 and helped provide that test data, yes I wish it would have moved faster. Also, more data on more AT, more browsers and OSes are coming. But a small group of volunteers can only do so much
Show threadEric EggertApr 18, 2023

@joehumbert @aardrian @dennisl WAI is putting a lot of effort into underfunded, understaffed projects, burning out volunteers in the process. That’s WAI’s choice. These patterns should be better supported or not published at all.

None of my criticism is of the people involved but of the intentionality of the work y’all do. This should not be a volunteer effort.

Show threadJoe HumbertApr 18, 2023
@yatil @aardrian @dennisl
I agree that the patterns should need better supported. The group has actually gotten Vispero to make changes based on the test results. To my knowledge this effort was suggested by mostly volunteers and a lot of work has been paid for by sources outside the W3C. As this is my only involvement in any work related to the WAI that is as far as I can comment
Show threadEric EggertApr 18, 2023

@joehumbert @aardrian @dennisl Yes, and it is good for what it is. But Test the Web Forward (aka Web Platform Tests, wpt) exist since 2014. W3C just never considered assistive technologies as part of the web platform, so no tests. And even the ARIA specification seems to be under-tested in wpt.

Accessibility spec testing must be where mainstream spec testing happens.

http://testthewebforward.org

Test the Web Forward

Show threadAdrian Roselli, pH0Apr 18, 2023

@joehumbert
My issue is consistently with how it is framed. Not the lack of time or effort from volunteers. This is not easy work and pitching it as a pattern library was, IMO, a mistake.

@yatil @dennisl

Show threadJoe HumbertApr 18, 2023

@aardrian @yatil @dennisl

Well Matt King and Michael Fairchild would know more, but I believe the goal is to test more than the APG in the future. I can't comment on the framing as I have not followed that aspect closely

Show threadMichael 🧐 FairchildApr 18, 2023
@joehumbert @aardrian @yatil @dennisl y’all have valid points and feedback. At the end of the day, I think this is a step in the right direction and I’m optimistic that it will continue to grow given enough support. There is a lot happening behind the scenes, beyond just these support tables, to lay the foundation for even more meaningful results.