Notices by Adrian Roselli (aardrian@toot.cafe)

@aral When possible, you can use `aria-labelledby` to point to an existing span of text that already tells the reader what it is.

But yes, it can be a box-ticking exercise. At least until Safari shows interest in supporting keyboard users and Chrome actually deploys the feature.

@aral FYI that without an appropriate role and accessible name, you might be at risk of failing 4.1.2. depending who’s checking.
Ref: https://adrianroselli.com/2022/06/keyboard-only-scrolling-areas.html

@aral Oh man there’s a browser!

I made a kawaii browser for one post and kawaii toast for another and this would have been so much easier (well, for the browser at least).

@gvlx
While I hate to pile on, it’s important to understand that the messge is from frustration with the ongoing and genuine harm that these kinds of barriers introduce.

It is valid to be angry and suggesting otherwise is dismissive of that community experience as a whole. That dismissiveness is part of the problem. Your lack of sympathy as well.

Historically, accessibility affordances only come after using this kind of advocacy. Which sucks.

@aral

@aral No idea if you saw this passibly-related thread about piping output to NVDA in a WINE setup: https://dragonscave.space/@meatbag/112528902037585952

@aral
Any chance it is being recorded and posted later?

@laura

@aral This would be a VoiceOver bug, then, since Safari only passes the text along. VO makes a decision on whether to call a hyphen character a “minus” on the fly.

Anyway, wasn’t trying to reply-guy your frustration. But if something changed from when I worked on this with my clients then I need to get ahead of it.

@aral I put a dollar sign in there because I thought you were doing US currency. Sorry, I mis-read.

I updated the pen and they all still announce with “minus one”.

@aral For poops and/or giggles:
https://codepen.io/aardrian/pen/yLryobo

NVDA/Firefox, JAWS/Chrome, Orca/Firefox VoiceOver/Safari all announce each div as “dollar [sign] minus one”.

VO was the only one to add “sign”.

It’s also possible I did not code this at all like in your scenario.

@aral You did not ask, and my Mac is too far away to just go get and make a test case, but did your negative numbers use the minus character (`−` / − / −) or a hyphen / dash?

Asking because I recall dealing with this with a client and I _think_ we resolved with the character. Could also be hopped up on maple syrup.

@aral @marjon @nachtfunke A11ySupport is an amazing resource, but largely maintained by one person (so not always as current as the rapid pace of browser releases).

With Google’s Baseline completely excluding accessibility support (https://adrianroselli.com/2023/12/baseline-does-not-really-cover-baseline-support.html), I have been trying to get it to commit to throwing some resources at A11ySupport (https://github.com/web-platform-dx/web-features/issues/498#issuecomment-1878781911) to cover that gap.

If you have any levers to pull to make that happen… well, we could all benefit.

@aral @LALegault @ProcessParsnip
I cannot see the original post (only a reply), but captions are for users with hearing and/or language needs. They are a visual affordance that are not helpful to blind users.

For low-vision users, closed captions can be scaled, colored, back-plated, etc. Open captions are burned into the video frames and cannot be adjusted.

A screen reader has no reason to access closed captions since that would conflict with audio from the video.

@pixelcode
Mathematicians don’t write math that way because not all math uses simple operators like that.

Hence MathML

Either way, it is not a simple NLP problem, it is also about context:
https://adrianroselli.com/2021/10/blaming-screen-readers-red-flag.html

The good news is you can contribute to NVDA (a free, open source screen reader) if you think you have a viable approach.

@odddev @thilo @aral

@pixelcode
Then you should pitch that. Or, if you are comfortable writing the code, you can file a PR:
https://github.com/nvaccess/nvda

NVDA is a free and open source screen reader, so community contributions are pretty much how it works.

You can also contribute to Orca:
https://github.com/GNOME/orca

Improvements to those screen readers put pressure on those from Apple, Google, Microsoft, and Freedom Scientific.

@odddev @thilo @aral

Happy Friday!

#AudioEye is suing me
https://adrianroselli.com/2023/05/audioeye-is-suing-me.html

I am not taking questions.

#a11y #accessibility

Welp, the WCAG3 update session at #CSUNATC23 is full and just like the good ol’ days of San Diego CSUN I was denied entry. Too small room, which is weird given the demonstrable interest.

Reminds me of that Viking & Lumberjack sit-in a few years ago.

Just like we should always ask “Could this meeting have been an email?”

…we should always ask “Could this PDF have been a web page?”

• The color contrast formula in WCAG 2 was written for a different technology;
• WCAG 3 is not well underway;
• WCAG 3’s contrast formula is still in development;
• You likely still need to conform to WCAG 2 because laws;
• ADA is silent on web accessibility.

Drafted in March 2022, got a cease & desist letter before posting, sat on it while that played out, sat on it some more, finally hit the ‘publish’ button:
https://adrianroselli.com/2023/02/audioeye-will-get-you-sued.html

Some of it is a bit dated, but the pattern of behavior is still clear. #accessibility #a11y

A welcome overview of the current state and suggested adjustments to Shadow DOM from @sundress:

“How Shadow DOM and accessibility are in conflict”
https://alice.pages.igalia.com/blog/how-shadow-dom-and-accessibility-are-in-conflict/

I also want `autocomplete` attribute support too, weird though that SC is.