Archive index

A11y Slackers Gitter Channel Archive 15th of March 2016

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    Mar 15 05:54
    [kevinchao89] Why does this treeview fail with VoiceOver, but works fine with NVDA/JAWS? http://accessibleculture.org/articles/2013/02/not-so-simple-aria-tree-views-and-screen-readers/#best
  2. powrsurg
    Mar 15 15:01
    I've got to review 508 compliance of our software for a client. Can I just use the waiver on Karl Groves' site? :)
  3. powrsurg @powrsurg really wishes he finished his media player before be tasked with this
  4. garcialo
    Mar 15 15:15
    What's the expected method for a screen reader user to navigate into panels (in a tab panel) which contain only text?
  5. MichielBijl
    Mar 15 15:20
    modifier keys+arrows probably?
  6. garcialo
    Mar 15 15:21
    That doesn't seem to do anything.
  7. sophieschoice
    Mar 15 15:21
    Ha, I have difficulty navigating through a webpage as well, but that is just starting/newbie issues with VoiceOver and NVDA.
  8. garcialo
    Mar 15 15:22
    Most of the example tab panels I can remember always have a focusable item in them...which I think it kind of cheating =p
  9. MichielBijl
    Mar 15 15:30
    I usually set focus to the tab panel itself.
  10. garcialo
    Mar 15 15:31
    To the panel itself?
  11. MichielBijl
    Mar 15 15:31
    See: http://thepaciellogroup.github.io/UXDev/uxathon-2016/component2/F.html
  12. MichielBijl
    Mar 15 15:31
    Hmm
  13. MichielBijl
    Mar 15 15:31
    Seems I broke that :P
  14. MichielBijl
    Mar 15 15:32
    -.-
  15. garcialo
    Mar 15 15:33
    If you set focus into the panel when the tab is activated, then you're going to have the problem where you have to switch tab, go back to the tabs, switch tab, go back to the tabs, etc.
  16. MichielBijl
    Mar 15 15:34
    No, I mean, next focusable is the tab panel
  17. MichielBijl
    Mar 15 15:34
    That version is borked
  18. MichielBijl
    Mar 15 15:35
    Should be fixed now
  19. garcialo
    Mar 15 15:37
    so, I don't see how I'm supposed to get to the content with a screen reader
  20. garcialo
    Mar 15 15:39
    maybe if there were headings there it would work...I'd have to know there was a heading there...but that seems to me like the most guessable thing to try going to next
  21. garcialo
    Mar 15 15:43
    Maybe it would work better if the panels weren't automatically activated and instead the user had to activate it which then placed focus into the panel?
  22. MichielBijl
    Mar 15 16:12
    We're running a survey: https://www.paciellogroup.com/blog/2016/03/design-pattern-usability-study-and-csun-2016-uxathon/
  23. garcialo
    Mar 15 16:13
    Yeah, I know =p
  24. MichielBijl
    Mar 15 16:13
    So what's with all the questions :P
  25. garcialo
    Mar 15 16:13
    Oh wait, was that a general statement not necessarily targeted as me? Sorry =p
  26. garcialo
    Mar 15 16:14
    Because I have reached something I can't think of a good solution for and was hoping I could get some guidance from others.
  27. MichielBijl
    Mar 15 16:15
    No, I just meant that we're trying to figure out how to solve the issue you describe :) But yeah, maybe someone here has good input!
  28. garcialo
    Mar 15 16:17
    No one does. We will continue hurling through the endlessness of the universe without cause; without hope; without a reliable method of navigating into a content-only panel form its tab.
  29. garcialo
    Mar 15 16:17
    @stevefaulkner You should tweet that.
  30. garcialo
    Mar 15 16:18
    except fix the typo
  31. zakim-robot
    Mar 15 18:17
    [ellyjones] Did someone here mention a Chromium bug a few days ago to do with not drawing alt text when images aren't drawn?
  32. jnurthen
    Mar 15 18:17
    I remember someone mentioning it somewhere - so it was probably here
  33. jnurthen
    Mar 15 18:18
    I think @MichielBijl commented in the bug
  34. MichielBijl
    Mar 15 18:18
    Yeap
  35. MichielBijl
    Mar 15 18:38
    @ellyjones what is the question regarding it?
  36. zakim-robot
    Mar 15 18:46
    [ellyjones] I believe I landed a fix for it today; it should be in tomorrow's chrome canary :)
  37. zakim-robot
    Mar 15 18:46
    [ellyjones] I have images disabled and I've thought it was by design for the longest time that we didn't draw the alt text
  38. zakim-robot
    Mar 15 18:48
    [lliskovoi] hey all, I’m looking at a site that has a table with divs inside of data cells. When I use VO, the table cells don’t get associated with the headers. Is this because of the DIVs? Would this count as WCAG violation?
  39. MichielBijl
    Mar 15 18:52
    @ellyjones: whoop whoop!
  40. MichielBijl
    Mar 15 18:53
    Nice work!
  41. zakim-robot
    Mar 15 18:53
    [ellyjones] heh thanks :) if you don't see the fixed behavior within a couple of days please let me know
  42. zakim-robot
    Mar 15 19:12
    [ellyjones] (also, if there are other chrome a11y problems that are particularly annoying/painful, please let me know and I'll see if I can triage them)
  43. zakim-robot
    Mar 15 19:26
    [skerrvy] @lliskovoi I would say it's very likely because it's using divs instead of a table and there is no relationship defined between the headers and data. I couldn't find anything in the WCAG that explicitly says it would be a failure, as I'm sure you could make it work with aria, although nothing would be as user friendly as table markup itself. https://www.w3.org/TR/WCAG20-TECHS/H51.html
  44. zakim-robot
    Mar 15 20:03
    [lliskovoi] Thanks @skerrvy ! To clarify, the divs are nested with <td>s and there are <th>s in the table, but the divs seem to be the only difference between cells that read out headers and ones that don’t (it’s a calendar, and funny enough, the empty cells seem to work fine).
  45. zakim-robot
    Mar 15 20:11
    [hmig] @MichielBijl - looks like chrome devs fixed the bug where alt text was not displaying: https://bugs.chromium.org/p/chromium/issues/detail?id=443259#c12
  46. zakim-robot
    Mar 15 20:12
    [hmig] :)
  47. MichielBijl
    Mar 15 20:12
    Yeap, see couple lines up ;)
  48. zakim-robot
    Mar 15 20:12
    [hmig] reading is hard ;)
  49. zakim-robot
    Mar 15 20:14
    [ellyjones] hmig: yeah, I got sick of being affected by it so I stole it and fixed it :stuck_out_tongue:
  50. zakim-robot
    Mar 15 20:15
    [hmig] Awesome! TY!
  51. zakim-robot
    Mar 15 20:15
    [hmig] :raised_hands:
  52. zakim-robot
    Mar 15 20:15
    [ellyjones] sure thing :) it's literally my job
  53. zakim-robot
    Mar 15 20:15
    [hmig] lol - when I saw a won’t fix on it, I was pretty upset
  54. zakim-robot
    Mar 15 20:16
    [ellyjones] I always thought it was design behavior that we disobeyed the spec
  55. zakim-robot
    Mar 15 20:16
    [hmig] so, thx
  56. zakim-robot
    Mar 15 20:16
    [ellyjones] until you offhandedly mentioned the bug a few days ago and I was like ohhhh
  57. zakim-robot
    Mar 15 20:17
    [jkiss] @kevinchao89: That’s interesting. It used to work fine enough with VoiceOver (back in 2013). With focus set to and remaining on the tree, arrow keys would happily expand/collapse branches and change which node was selected (using aria-activedescendant). Seems that still works, but now VoiceOver is no longer announcing the name of the node. Not sure why.
  58. zakim-robot
    Mar 15 20:19
    [hmig] @ellyjones: yea - I was pretty upset & needed some advise on what to comment to the "won’t fix” label and @MichielBijl stepped in with a good response
  59. zakim-robot
    Mar 15 20:19
    [hmig] Thx again for fixing it!
  60. zakim-robot
    Mar 15 20:21
    [ellyjones] any time :) if you see other chrome a11y bugs that are annoying/painful/etc please let me know and I can at least triage them
  61. zakim-robot
    Mar 15 20:21
    [hmig] will do! thx :)
  62. powrsurg
    Mar 15 20:39
    "When software is designed to run on a system that has a keyboard, product functions shall be executable from a keyboard where the function itself or the result of performing a function can be discerned textually." - this just means you can do everything through a keyboard, right?
  63. zakim-robot
    Mar 15 20:59
    [deborah_kaplan] That's how I've always interpreted that one.
  64. StommePoes
    Mar 15 21:43
    Is this chromium bug the bug where alt text won't show if it's larger than the dimensions of the unladen image?
  65. zakim-robot
    Mar 15 21:56
    [skerrvy] @StommePoes to my understanding it's a fix for when a user has images disabled in their browser settings and the alt text was not being displayed in its place (?)
  66. StommePoes
    Mar 15 22:25
    Well blink and webkit showed alt text... except if that text was longer than the width of the image
  67. StommePoes
    Mar 15 22:25
    But I don't recall it always blatently not showing alt
  68. StommePoes
    Mar 15 22:25
    meanwhile, Opera/Presto I think was the one that showed it, but let it get cut off if the text was longer (almost as bad)
  69. jonathantneal
    Mar 15 22:25
    Dealing with a CSS issue with <legend> and 456 Berea St recommends giving the legend display:table but will that mess with ATs?
  70. StommePoes
    Mar 15 22:27
    Ah no, just checked, it was other way around: alt gets cut off in blink
  71. StommePoes
    Mar 15 22:31
    I can't even find disable images in the settings, I've always had to use plugins
  72. StommePoes
    Mar 15 22:33
    oh, it's under Privacy. Yeah, I would have definitely guessed that, instead of under Web Content with the fonts and stuff...
  73. StommePoes
    Mar 15 22:34
    hm yup. This is new-- no alts even if image dimensions are large enough. What I get for relying on plugins to block images.
  74. jnurthen
    Mar 15 22:37
    aarrrgghhhh
  75. jnurthen
    Mar 15 22:37
    why did that take me so long to debug
  76. garcialo
    Mar 15 22:37
    no complaining without providing NDA breaking details
  77. jnurthen
    Mar 15 22:37
    was updating text in a live region and FF was reading weird text
  78. jnurthen
    Mar 15 22:38
    i.e. just reading some of it
  79. jnurthen
    Mar 15 22:38
    turns out it was only reading the LETTERS that had changed!
  80. StommePoes
    Mar 15 22:38
    ...and in the end, it was just Colonel Mustard in the library with the candlestick?
  81. garcialo
    Mar 15 22:38
    oh, that's awesome
  82. StommePoes
    Mar 15 22:38
    was that with "additions"?
  83. jnurthen
    Mar 15 22:38
    so it was reading pe ple P L
  84. jnurthen
    Mar 15 22:38
    instead of people page loaded
  85. jnurthen
    Mar 15 22:39
    cause the previous page had most of the letters in common
  86. jnurthen
    Mar 15 22:39
    nope
  87. StommePoes
    Mar 15 22:39
    oh, it was reading differences?
  88. jnurthen
    Mar 15 22:39
    just default aria-live region
  89. jnurthen
    Mar 15 22:39
    had to add aria-atomic but I thought replacing a text node didn't need that
  90. StommePoes
    Mar 15 22:39
    I didn't think so either
  91. jnurthen
    Mar 15 22:39
    I guess the code was really doing a replace not a remove and add (which I normally do - but it wasn't my code)
  92. StommePoes
    Mar 15 22:40
    was that only ff/nvda?
  93. jnurthen
    Mar 15 22:40
    nope JAWS and NVDA with FF
  94. jnurthen
    Mar 15 22:40
    all was fine with IE
  95. garcialo
    Mar 15 22:40
    what about in other browsers?
  96. garcialo
    Mar 15 22:40
    yeah, was going to say it sounds like a FF issue
  97. StommePoes
    Mar 15 22:40
    could be a rendering optimisation of FF?
  98. jnurthen
    Mar 15 22:40
    well - IE doesn't do anything with live regions right - it defers it all to the AT
  99. jnurthen
    Mar 15 22:40
    no idea but I hate bugs like that
  100. stevefaulkner
    Mar 15 23:12
    @jonathantneal display:table used to be an issue in Firefox but was fixed I believe, there is a bug about it somewhere
  101. jonathantneal
    Mar 15 23:40
    Thanks.