Archive index

A11y Slackers Gitter Channel Archive 21st of September 2017

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    @zakim-robot
    Sep 21 00:03
    [ghanek] The gulf I have witnessed is between an occasional user of screen readers declaring that all the components pieces could be heard and properly identified. True.
    But then the actual user experience failed hard, when everything was put together.
  2. zakim-robot
    @zakim-robot
    Sep 21 00:44
    [sugi] @ghanek what do you mean?
  3. zakim-robot
    @zakim-robot
    Sep 21 00:57
    [caesar] Sounds like they did unit testing on individual components, but failed to test the entire page
  4. zakim-robot
    @zakim-robot
    Sep 21 01:14
    [karlgroves] Sorry but I have to strongly disagree with a lot of what has been said above. Amateurs testing with SRs is more likely to lead to getting trapped in quicksand than to insight.
  5. [karlgroves] As soon as they start doing so, they start chasing their tails on what they think SRs should “say”
  6. [karlgroves] For instance, I’ve had client devs tell me they had to add tabindex="0" on all <td> elements because they didn’t know how SRs worked with tables
  7. zakim-robot
    @zakim-robot
    Sep 21 01:19

    [karlgroves] We see, even in this slack, people asking how to “make” a SR read numbers, phone numbers, abbreviations, etc. correctly.

    IOW without some more nuanced understanding of how SRs work - and how PWDs interact with them - all we’re doing is adding more confusion around a11y.

    We’re waaaaaaaay better off, IMO, giving devs exposure to real PWDs using AT and giving them the necessary guidance on proper technical implementation of a11y

  8. [quidkid] PWDs?
  9. [quidkid] oh!
  10. [karlgroves] PWDS === People With Disabilities
  11. [quidkid] nvm haha yeah i got it
  12. [karlgroves] thumbsup emoji
  13. [clairewebber] I can see both sides to this. Sometimes I have to do heaps of testing for developers who are not sure if there code will work for tiny changes and they wish they could figure out SRs to test themselves. However even some developers who have a great understanding of A11y in my team use SRs more often and still do stuff like @karlgroves mentioned above, making more work for themselves. It's something we discuss often which way is best and whether more people should be testing with SRs and we are still not sure which way to go with it.
  14. zakim-robot
    @zakim-robot
    Sep 21 01:26
    [karlgroves] I guess my point is that unless someone actually understands a11y, understands SRs, and knows a lot about how PWDs interact with computers, they really have no business testing with SRs
  15. zakim-robot
    @zakim-robot
    Sep 21 01:40
    [quidkid] Yeah at first my company expected me to do the screenreader testing since I took initiative to learn accessibility. But I don’t trust myself to do it right and refused so we ended up hiring freelancers/agencies(trying different options right now). And even now with getting reports of screenreader issues, I have a hard time replicating them. And while I definitely see the benefit of having me also be able to screenreader test, I hope that we will hire someone with a disability that actually uses a screenreader on a daily basis.
  16. [karlgroves] Doing this for 15 years now and I still don’t trust myself with SRs. Usually I can get everything done I need to, but sometimes when SRs do something weird, I grab the help of a person who uses SR full-time
  17. zakim-robot
    @zakim-robot
    Sep 21 01:48
    [quidkid] what boggles me is that we are hiring abled bodied people to do our screenreader testing. but I’m not quite sure how would we advertise trying to find someone who is using a SR fulltime. another thing I’ve been wondering is whether fulltime SR users know all the WAI ARIA design patterns for the widgets. like how would they know for tab panels you would need arrow keys? and that sort of thing.
  18. [svinkle] I once saw someone put tabindex="0" on just about everything because, "how else would someone tab to read the text?" We promptly had a discussion on how to use a SR and how PWD typically use them. That said, I'm still not very good with them. Very much rely on usability testing and real people's opinions on things.
  19. [ugi] @quidkid ARIA keyboard patterns mirror native components.
  20. [svinkle] @quidkid This is a reality; people don't know about the require content switch. It can be tough to get UI right.
  21. [svinkle] @ugi Native components on desktop apps, not usually on the web.
  22. [svinkle] The web is sometimes a true wild west where anything goes.
  23. zakim-robot
    @zakim-robot
    Sep 21 01:54
    [svinkle] That said, it's good to have standards to follow. The ARIA practice guidelines are a great resource.
  24. [svinkle] Eventually people discover and get used to the UI, but would be ideal to avoid these pain points from the get-go.
  25. zakim-robot
    @zakim-robot
    Sep 21 02:07
    [ghanek] Hi @sugi !
    A form that didn't use semantic elements well, and had silent and non-focused progressive revelation. So the form elements could be heard. Then, if you knew where to listen for them, the newly revealed elements could be heard, but their sudden appearance wasn't made explicit. Non-sighted sr users couldn't complete it.
  26. zakim-robot
    @zakim-robot
    Sep 21 02:31
    [caesar] @svinkle I saw the "tabindex=0 everything" once too, because the devs only had experience with VoiceOver and they thought that's how all SR should behave.
  27. [caesar] The other thing that blew my mind recently was during the discussion about hacking an abbreviation to read properly, when somebody pointed out that it would essentially screw up a Braille Reader "read"!
  28. zakim-robot
    @zakim-robot
    Sep 21 03:17
    [shawn.henning] @quidkid You could hire a screen reader user that was not a developer. I'm a project manager for instance. Though I am primarily a Voice Over user so might be susceptible to some of the issues mentioned above.
  29. [shawn.henning] The abbreviation thing makes for some amusing listening such as "St. Elmo's" being read as "street Elmo's". Or the India English voice Deepa reading battleaxe as battlie axe I'm not sure I can convey the pronounciation, but try reading the D&D 5e manual in Voice Dream with the Deepa voice for some amusing things.
  30. zakim-robot
    @zakim-robot
    Sep 21 03:23
    [shawn.henning] *note the st example above did not work just now likely due to context.
  31. stevefaulkner
    @stevefaulkner
    Sep 21 12:11
    hi folks :smile:
  32. zakim-robot
    @zakim-robot
    Sep 21 13:51
    [quidkid] @shawn.henning our freelancer testers are testers or ui/ux designers but they aren’t full time SR users. i think the last time I discussed it with someone, I remember “subject matter expert” as a term being thrown around but otherwise, not too sure how to go about finding/hiring
  33. zakim-robot
    @zakim-robot
    Sep 21 14:09
    [shawn.henning] @quidkid First you need an open job, then there is a #jobs channel here as well as the jobs mailing lists for ACB and its affiliates and NFB. Posting to these places is more likely to get you a screen reader user.
  34. [shawn.henning] So many blind people have trouble finding jobs that will be able to get their heads around how to accommodate them that many have just stopped trying; yet there is a ton of talent sitting idle. Some of it is inaccessible software (cough slack) which means going in a blind person has no idea if they can work with the tools being used. I'll step off my soapbox now...
  35. zakim-robot
    @zakim-robot
    Sep 21 14:47
    [quidkid] @shawn.henning good point so maybe slack isn’t the best place to post then. I was also considering reaching out to local long-term care facilities but not sure if that would be the most effective as preferably we need someone who is already familiar with the wcag aa guidelines.
  36. Michael Fairchild
    @mfairchild365
    Sep 21 15:07
    Question for y'all: if a web application emulates navigating to a new page with javascript (it just replaces the main region of the page), but fails to set focus after the new content loads (current focus is lost because the button that activated the change is now gone), which success criteria of WCAG is being violated?
  37. In other words, the change of context is not announced by AT
  38. zakim-robot
    @zakim-robot
    Sep 21 15:13
    [rselzer] I recall the criteria for modal windows mentioning returning focus when closing them. Checking that section might lead you to a more general rule
  39. Michael Fairchild
    @mfairchild365
    Sep 21 15:15
    I went down that path, but couldn't find anything that specifically addressed a change of context like that.
  40. zakim-robot
    @zakim-robot
    Sep 21 15:16
    [tink] I'm not sure there is a WCAG SC that directly deals with this.
  41. [rselzer] maybe google around for something like "accessibility and progressive web apps" or "accessibility and PWAs"
  42. [tink] Best practice is to update the <title> element, and place focus on the container for the chunk of the content that represents the new page (prefrably with an associated heading).
  43. [tink] But this isn't an a11y fail IMO because it's the same thing that happens (more or less) when a new page loads in the browser.
  44. Michael Fairchild
    @mfairchild365
    Sep 21 15:25
    In my experience, when a new page loads, a screen reader will announce that a new page has loaded. For example, VO will announce the title of the new page after it loads. In the example that I provided earlier, after a link is clicked, nothing is announced. While I'm sure a screen reader user could figure out that context did actually change after navigating around the page for a bit, it seems like a bad user experience to not announce anything after a link was activated (could potentially lead to confusion). Or am I way off?
  45. zakim-robot
    @zakim-robot
    Sep 21 15:29
    [tink] @david.laforest it wouldn't be focus visible because there is no expectation/requirement for anything to have focus.
  46. [tink] @mfairchild365 that's right, but there is no WCAG SC requirement for that.
  47. [rselzer] you can use role="dialog" do have the sr announce when content inside the element has changed
  48. Michael Fairchild
    @mfairchild365
    Sep 21 15:32
    role="disalog" is an interesting solution. I will have to look at that.
  49. zakim-robot
    @zakim-robot
    Sep 21 15:32
    [tink] The best solution is to use .focus() to take keyboard focus to the element that is the container for the body, assuming it's something like the <main> element use aria-labelledby to associate it with the heading that is at the start of the content area.
  50. Michael Fairchild
    @mfairchild365
    Sep 21 15:32
    Thanks, Tink, that was the solution that I was going to recommend.
  51. However, I worry that because there is no WCAG requirement for this, the vendor might not feel obligated to fix it.
  52. zakim-robot
    @zakim-robot
    Sep 21 15:33
    [rselzer] there's also role="alert" the difference between the two is that if the SR is reading something and content inside a role="alert" changes it will interupt what the sr is reading to read the alert content. It's literally for alert type content. IIRC Dialog will finish what it was reading before announcing the content change
  53. [tink] @rselzer it wouldn't be an appropriate use of either role to apply them to the content container.
  54. [rselzer] agreed, but is there some other way to announce the change of the page inside the content container?
  55. [tink] Yes, see ^^
  56. [tink] <main aria-labelledby="this"><h1 id="this">New page title</h1>...</main> then use .focus() to move focus to the main element.
  57. [tink] Plus updte the <title> element in case a screen reader user queries it after the new content has loaded.
  58. zakim-robot
    @zakim-robot
    Sep 21 17:24
    [lauraf] @mfairchild365 2.4.4 link purpose; if they don't know what the link does when they click it, and nothing appears to happen, then its a problem. For a single-page application we had here that a button submitted a question and took you to the next question, the aria design pattern for button was appropriate: "If the button action indicates a context change, such as move to next step in a wizard or add another search criteria, then it is often appropriate to move focus to the starting point for that action."
  59. Michael Fairchild
    @mfairchild365
    Sep 21 17:25
    Thanks, that is helpful. I landed on that, 1.3.1 - Info and Relationships and 1.3.3 - Sensory Characteristics as somewhat relating to the problem.
  60. zakim-robot
    @zakim-robot
    Sep 21 17:48
    [lauraf] Also it may help to look up some of the many talks/presentations by Marcy Sutton on single page apps.
  61. zakim-robot
    @zakim-robot
    Sep 21 17:55
    [jamesn] +1 on Marcy's talks. Also this is a situation that WCAG2.1 3.2.7 is intended to cover I believe. (of course wcag 2.1 is far from final at this point)
  62. Michael Fairchild
    @mfairchild365
    Sep 21 18:01
    ah, yes https://www.w3.org/TR/WCAG21/#change-of-content is perfect. I was going to ask if it was covered in 2.1