Archive index

A11y Slackers Gitter Channel Archive 25th of August 2015

What fresh hell is THIS now? - Patrick Lauke
  1. powrsurg
    Aug 25 00:09
    What about Windows-Eyes :p
  2. powrsurg
    Aug 25 00:10
    Real question: how significant are the differences between VoiceOver for Safari on OS X vs iOS?
  3. jonathantneal
    Aug 25 00:13
    @powrsurg I tested Windows-Eyes as well, which warns the user not to use Edge. I also tested Safari on OSX vs iOS and the results were the same. The interesting thing about Safari for iOS is that you swipe right to advance and you swipe left to go back, which felt a bit counterintuitive to me at first.
  4. powrsurg
    Aug 25 00:15
    last I tested VoiceOver (on my ipod touch) it was skipping some stuff with overflow-y: auto
  5. powrsurg
    Aug 25 00:15
    but that wasn't thorough testing, could have been a few other things
  6. jonathantneal
    Aug 25 00:17
    http://webaim.org/projects/screenreadersurvey5/ - Windows Eyes seemed to be much lower on the totem poll, but it was the most consistent, producing the same result in IE11, Firefox 40, and Chrome 44.
  7. powrsurg
    Aug 25 00:21
    yeah, it's not one of the major ones, but it's free if you have an MS Office installed
  8. powrsurg
    Aug 25 00:24
    I literally forgot about SAToGo =/
  9. MichielBijl
    08:52

    a11ywin: designer took an interest in accessibility!

  10. MichielBijl
    08:52
    Woah
  11. MichielBijl
    08:52
    Don't start sentences with a hash…
  12. StommePoes
    09:35
    why?
  13. StommePoes
    09:35

    foo

  14. MichielBijl
    09:59
    Gitter supports markdown, so a # is a h1
  15. LjWatson
    10:22
    Post by Marco Zehe (from Mozilla) on changes to Firefox that will expose more information to VoiceOver on OS X https://www.marcozehe.de/2015/08/25/improvements-in-accessibility-for-mac-os-x-in-firefox-41/
  16. LjWatson
    10:57
    @rodneyrehm @jonathantneal I stand completely corrected about access to MSEdge with third party Windows screen readers... Looks like NVDA 2015.3 has initial support :)
  17. StommePoes
    11:07
    woohoo!
  18. LjWatson
    11:09
    @StommePoes Indeed. This is a huge step forward. Huge props to Jamie and Mick at NVAccess for doing it.
  19. StommePoes
    11:12
    I know they were constrained by a lot in edge tho so I'm assuming dstorey adn jacob rossi and those guys have well changed something to help, otherwise I think they were stuck
  20. StommePoes
    11:12
    but it's definitely cool to see the FOSS guys ahead of FS :P
  21. iandevlin
    11:27
    I keep forgetting about the existence of this place, which is poor form.
  22. StommePoes
    11:27
    we have a word for that here: stout
  23. StommePoes
    11:27
    dunno if it's the same in German?
  24. iandevlin
    11:34
    a word for what?
  25. StommePoes
    11:35
    for the "poor form" :D
  26. iandevlin
    11:36
    Oh, no idea, to me, stout is either fat or a drink like Guinness!
  27. StommePoes
    11:37
    haha
  28. StommePoes
    11:38
    it's like when you tsk-tsk someone, or a child or puppy misbehaves, or it could mean mischievious... but usually it's just, bad you and your poor form :P
  29. MichielBijl
    11:43
    @LjWatson did James ping you about the APG Wiki pages?
  30. MichielBijl
    11:50
    I feel ashamed I only just found RFC 1766 and have always filled the lang attribute incorrectly :(
  31. MichielBijl @MichielBijl sits down in the corner, tries not to cry; cries a lot
  32. StommePoes
    11:51
    what were you doing wrong?
  33. StommePoes
    11:52
    this was the newest one? https://tools.ietf.org/html/rfc5646
  34. StommePoes
    11:54
    we don't use it but the 3-letter ones are more accurate
  35. MichielBijl
    11:56
    I always used stuff like nl_be instead of nl-be :x
  36. StommePoes
    12:01
    we do <html lang="nl"> but in the backend in python it's always xx_XX
  37. StommePoes
    12:01
    but we don't do the country in our languages
  38. StommePoes
    12:02
    esp since we do that wrong too (all english is "US", usually shown with a British flag)
  39. StommePoes
    12:02
    ug
  40. StommePoes
    12:02
    hate flags==languages, so wrong
  41. iandevlin
    12:10
    It is very wrong, and often very inaccurate!
  42. iandevlin
    12:10
    I use en-IE on my website, because that is what I speak and write.
  43. LjWatson
    12:18
    @MichielBijl No, but @IanPouncey did... although I can't quite remember what it was we wanted as gh-pages?
  44. MichielBijl
    12:29
    @LjWatson this page: https://github.com/w3c/aria/wiki/Aria-Authoring-Practices-Patterns-Status
  45. MichielBijl
    12:30
    So anyone can make a PR if they start work on a code example (add their name to the 'code example' column).
  46. jonathantneal
    12:53
    I currently have 14 combinations of screen readers available for testing with the latest VoiceOver, NVDA, JAWS, Windows Eyes, and ChromeVox. I’m working to document how I setup the tests, and I’ve put the results of some initial tests @ http://sandbox.thewikies.com/aria-tests/results.html
  47. jonathantneal
    12:54
    I was mostly interested in what is read in different "hidden" text scenarios, but I’m hoping to expand the tests to include how labelledby and describedby are picked up.
  48. jonathantneal
    12:54
    Or, if there are other common patterns you can think of, I’d be interested in testing them.
  49. jonathantneal
    13:15
    @LjWatson thanks for the update on NVDA. I will see if I can test it later.
  50. LjWatson
    13:17
    @MichielBijl Thanks. When I looked at it after the last call, I couldn't see a way to make a single file within a repo sync to a gh-pages branch.
  51. LjWatson
    13:17
    Not sure if it's possible. I've only ever switched a repo to use gh-pages as master, not just a single component within the repo.
  52. MichielBijl
    14:00
    @LjWatson I'm thinking we could just move it to a separate folder within the repo. Then we could reach it via gh-pages right?
  53. MichielBijl
    14:01
    So translate it to html and we should be good to go. If you need any help; you know where to find me.
  54. MichielBijl
    14:02
    Could do some styling or something.
  55. powrsurg
    14:50
    does anyone know a service that can be used to create audio transcripts for audio that is obtained via getUserMedia?
  56. powrsurg
    14:56
    We've got a tool users upload their own audio files via the mic. Got that all working ... would be good if we could somehow make it more accessible via having the a transcript automatically generated
  57. stevefaulkner
    16:04

    A11ySlackers chat log 24th August 2015 http://w3c.github.io/a11ySlackers/archives/2015/08/2015-08-24.html

  58. LjWatson
    16:23
    @powrsurg Might be worth talking to 3Play Media? They have a captioning/transcription service, and it might have an API for integration into other services perhaps?
  59. aardrian
    18:00
    I think @LjWatson mentioned this earlier, but NVDA's brand new release has some Win10 support: http://www.nvaccess.org/post/nvda-2015-3-released/
  60. jonathantneal
    18:30
    @aardrian I’ll definitely try it and update http://sandbox.thewikies.com/aria-tests/results.html since the current notes show Edge as being unusable.
  61. jonathantneal
    19:07
    @aardrian @LjWatson NVDA 2015.3 has issues in Edge. It reads content in [aria-hidden="true"] and [hidden], but then fails to read content in [style="position: absolute; clip: rect(0 0 0 0) (the only time this test has failed in any scenario).
  62. powrsurg
    19:15
    Maybe I'm misreading that, but it looks like that is more for calling generated transcripts through an API, and it looks to me like it requires video when we just need audio.
  63. stevefaulkner
    19:21
    @jonathantneal edge is badly broken in accessibility support, NVDA does its best to work around the brokenness. I don't think it has been suggested that trying to support AT users on Edge is a practical proposition at this time. Microsoft has some serious work to do on implementation of its accessibility API...
  64. jonathantneal
    19:25
    @stevefaulkner, so it’s safe to say that Edge is the least accessible browser ... at least of highly campaigned browsers?
  65. jonathantneal
    19:27
    Is it really less accessible than, say, Internet Explorer 6?
  66. powrsurg
    19:28
    Edge basically has most (all?) of the accessibility features of IE stripped from it, but has not finished on adding the new accessibility features they (say) they are working to add in
  67. powrsurg
    19:29
    The list of features on the first platform update on the Edge dev blog does not mention any accessibility related features there either
  68. stevefaulkner
    19:30
    Edge is largely a black hole in regards acc support its worse than IE but without the ability for AT to hack around its deficits, which IE allows
  69. powrsurg
    19:30
    well, I guess you can say the oninvalid event handler is sort-of an accessibility improvement
  70. jonathantneal
    19:35
    I must be late to the party, but that’s wild. That’s really wild. https://twitter.com/jon_neal/status/636260600922939392
  71. jonathantneal
    19:37
    Well, Internet Explorer 11 tested really great most of the time, so I hope it’s not a complete step back once the next few months of Edge updates are released.
  72. stevefaulkner
    20:21
    IE sucks, but AT works around its suckiness by hacking the DOM, Edge does not provide API access to DOM for client software like IE does
  73. cptvitamin
    20:28
    is there a way to use display:table without (some) screenreaders reporting the container as a table?
  74. aardrian
    20:29
    Got into an interesting Twitter convo this morning about making Tumblr images more accessible. As Tumblr doesn't provide facility for alt text in photo posts, and requires HTML knowledge for images in text posts, does anyone have techniques they use? I have my own, but it's not ideal.
  75. aardrian
    20:29
    Context for Twitter convo: https://twitter.com/LeonieWatson/status/636120493515493376
  76. kensgists
    20:40
    @cptvitamin did you try role="presentation" ?
  77. cptvitamin
    20:53
    @kensgists I did not. I tried role=“region” to see if providing an explicit role would trump the one from CSS and that didn’t work. Let me try role=“presentation"
  78. powrsurg
    21:32
    @stevefaulkner Opera has bad accessibility support? I would figure it would be close to Chrome
  79. powrsurg
    21:34
    Just referring to a recent tweet
  80. zakim-robot
    21:35
    [Karl Groves, a11y] @powrsurg everything Opera has WRT accessibility support it owes to Blink. Opera gave zero fucks about a11y.
  81. stevefaulkner
    21:35
    In theory for blink clones yes, but opera does not claim any accessibility support or put any resources into support
  82. stevefaulkner
    21:36
    As @karlgroves says but trying to be slightly more diplomatic 😁
  83. powrsurg
    21:37
    Oh, I read that tweet as saying the browser had bad support (as in users of that browser would have a bad experience), not that the company had put little effort into things
  84. stevefaulkner
    21:41
    I don't think that opera's acc it gets for free from blink implementation can be relied upon
  85. stevefaulkner
    21:42
    But happy to be proven wrong
  86. powrsurg
    21:42
    That's fair enough. I just assumed it was on par with Chrome and hadn't been running through tests in screen readers in Opera when I've run through things
  87. powrsurg
    21:43
    guess that is one more thing to test
  88. zakim-robot
    22:03
    [Marcy Sutton, a11y] @aardrian I add alt text manually in HTML for A11ywins, it suuuuucks
  89. aardrian
    22:09
    @marcysutton Yeah, it really does suck. Was hoping someone had found a better way. Or an easier way.