Developer Perception of Baseline Widely Available re: Accessibility

[lolaodelola]

Assumption

Developers assume a feature being Baseline widely available also means it's accessible in a general sense, i.e. it shows up in the a-tree as expected, it works as expected with assistive technologies.

Considerations

• The DOM tree and Accessibility Tree are sperate in the browser, a web feature being presented as expected in the DOM tree as expected doesn't preclude it from having issues in the accessibility tree.
• Widely Available web features can still have browser interoperability issues as it concerns accessibility, e.g. dialog was made widely available in 2022 however it had major focus issues which led to advice to use custom dialogs instead. Changes were made in 2023 which addressed the major focus issues, which made up a lot of (but not all) the interoperability issues.

Potential Questions

I implemented some of the feedback from the original questions but suspect these could still use some workshopping. We may not need all three either, we could potentially lose q3.

---

1. An HTML element is labeled as "Baseline widely available." Mark all that apply to this feature
   a. It works in all major browsers latest releases
   b. It works with assistive technologies (e.g. screen-readers, voice control, keyboard control, etc)
   c. It works in Web Views
   d. It's available in major browsers' releases over a year ago.
   e. None of the above
   f. I don't read Baseline support data

2. How confident are you that a web feature marked as "widely supported" would work as expected with screen-readers?
a. Very confident
b. Somewhat confident
c. Unsure
d. Somewhat doubtful
e. Very doubtful

3. How likely are you to assume a web platform feature marked as "widely supported" also has full assistive technology support (i.e. screen readers, voice control, etc.)?
a. Very likely
b. Somewhat likely
c. Neutral
d. Somewhat unlikely
e. Very unlikely

[captainbrosset]

Let me ping a few folks who participated in the live discussion during the WebDX call, about the above survey proposal, so they can provide feedback if they have any: @cyns and @atopal (obviously, anyone is welcome to provide feedback too).

[cyns]

This looks good to me. I'm very curious about whether developers assume baseline support includes accessibility

[ddbeck]

I'm very curious about this! But I worry a little that the questions, as written, are a bit leading. Instead of asking about the likelihood of behavior, it might be more straightforward to ask the respondents to tell us what they think about an example, including some things we know to be falsehoods and contradictions. Something like this (just spitballing here):

1. An HTML element is labeled as "Baseline widely available." Mark all that apply to this feature.

   a. It's available in major browsers' latest releases.
   a. It works with screen readers.
   a. All browser vendors are planning to build this into their browsers.
   a. It's available in major browsers' releases over a year ago.
   a. etc.
   a. etc.
   a. None of the above

[tidoust]

For question 1, as currently phrased, I suspect I would choose a. or c. even if I don't know what Baseline is just because the uppercase AND looks suspicious, and the other answers look smarter. I like @ddbeck's multi-choice approach that could help dismiss answers from people who would just check boxes randomly. As an added bonus, it could also tell us a bit more about other possible misunderstandings (for example, one choice could be "It's available in Webviews").

I don't recall the original questions and feedback received but looking at questions 2 and 3, I'm not sure I understand the difference between them. They seem to be asking roughly the same question with different terminology:

• "web feature" vs. "web platform feature"
• "how confident are you that" vs. "how likely are you to assume"
• "work as expected" vs. "has full support"

Question 2 seems scoped to screen readers while question 3 seems broader in scope. Is that the intended difference?

[lolaodelola]

@ddbeck Thanks for this, I was concerned about leading questions, I think your suggestion guards against that. I'll change the format of the questions.

@tidoust

As an added bonus, it could also tell us a bit more about other possible misunderstandings (for example, one choice could be "It's available in Webviews").

Is there anything else the group would be interested in checking for? I know accessibility & webviews are two that we've discussed recently, are there others that have come up in the past?

I don't recall the original questions and feedback received but looking at questions 2 and 3, I'm not sure I understand the difference between them.

I agree that they may be too similar, I don't think we need both. 2 is specifically about screen-readers while 3 is about assistive technologies in general including voice control, keyboard, braille, etc.

[lolaodelola]

@captainbrosset @ddbeck @tidoust I've updated the questionnaire, I basically lifted Daniel's suggestion and included an answer about web views too. I think the one question is enough. cc @cyns @atopal