Archive index

A11y Slackers Gitter Channel Archive 12th of January 2016

What fresh hell is THIS now? - Patrick Lauke
  1. StommePoes
    12:06
    drag and drop is hard as helkl
  2. StommePoes
    12:06
    it's extremely difficult to implement in the manner we usually see drag and drop
  3. StommePoes
    12:07
    the working accessible versions I have seen (a tic-tac-toe box using grid role and a button, and Gez Lemon's which uses an aria-popup which is unwieldy fitting a lot of text into) do not do what most of us think of as "drag and drop"
  4. StommePoes
    12:07
    this is due to the really crappy support we currently have for SRs
  5. MichielBijl
    12:07
    @StommePoes didn't you build a dnd module?
  6. StommePoes
    12:07
    and I have no idea what Dragon does
  7. StommePoes
    12:07
    I did
  8. StommePoes
    12:07
    Well, ian pouncy wrote one and it only worked with plain keyboard, but not with SRs and keyboard
  9. StommePoes
    12:07
    so we modified it to make everyone some kind of list and list item
  10. StommePoes
    12:08
    then it worked in NVDA, but stil not JAWS
  11. StommePoes
    12:08
    we ended up needing to put role=application to get jaws to work and that necessitated role=document to get the text in teh destinations readable
  12. StommePoes
    12:08
    and going into and out of role="application" really, really, really sucks
  13. StommePoes
    12:08
    sucks hard hairy donkey balls
  14. StommePoes
    12:09
    For non-AT keyboarders, DnD is pretty ok to do
  15. MichielBijl
    12:09
    JAWS, AT keyhoarder
  16. StommePoes
    12:09
    So my best advice to anyone needing a fully accessible DnD is, test the hell out of it, and have some backup design plans for when stuff won't work.
  17. MichielBijl
    12:10
    Might be useful to link to Ian Pouncy and your work (if it's open source that is)
  18. StommePoes
    12:11
    no it's paid-for internal stuff
  19. StommePoes
    12:11
    not public
  20. StommePoes
    12:11
    however you saw it
  21. StommePoes
    12:11
    you saw a version of it
  22. StommePoes
    12:11
    with the cats
  23. StommePoes
    12:12
    I suppose that could be public.
  24. Veyfeyken
    12:28
    Question: is there another way to create a keyboard trap for screenreaders in a modal dialog besides hiding everything but the dialog with aria-hidden?
  25. Veyfeyken
    12:29
    The client can't use aria-hidden like that because the dialog is to far embedded. Angular: 1.000.000.000.000 <div>'s
  26. MichielBijl
    12:29
    No way to place the dialog outside all the other content?
  27. Veyfeyken
    12:30
    Don't know.
  28. Veyfeyken
    12:31
    There's no known alternative that can be used on the specific code of the dialog itself?
  29. MichielBijl
    12:31 
    <main aria-hidden></main> <div role="dialog"></div>
  30. MichielBijl
    12:31
    That would work I guess?
  31. zakim-robot
    12:31
    [rodneyrehm] @veyfeyken: I don’t think there’s another way than using aria-hidden. But why is that not option in your case?
  32. zakim-robot
    12:33
    [rodneyrehm] If you can’t work with @michiel’s suggestion (which is the simplest you’ll find), maybe http://allyjs.io/api/maintain/hidden.html can help
  33. Veyfeyken
    12:35
    Thx. I have a look at that.
  34. Veyfeyken
    12:35
    I'll have...
  35. StommePoes
    12:40
    problem is, any keyboard trap we build with JS traps tabs
  36. StommePoes
    12:40
    but I don't know of a way to trap SR-style navigation.
  37. StommePoes
    12:40
    or, for that matter, regular keyboarder on FF with caret nav on
  38. StommePoes
    12:41
    Rodney's script finding all the sections and aria-hiddening them sounds feasable... writing that yourself does not.
  39. StommePoes
    12:42
    function should be named hideALLTHETHINGS!!!
  40. MichielBijl
    12:43
    hideAllTheThings
  41. StommePoes
    12:43
    Needs more insanity
  42. MichielBijl
    12:43
    hIdEaLlThEtHiNgS?
  43. MichielBijl
    12:44
    h1d3a117h37h1n9s!
  44. MichielBijl @MichielBijl 's brain hurts
  45. StommePoes
    12:44
    https://lh4.googleusercontent.com/-l1jz3VzM9pk/T4m0F53PcOI/AAAAAAAAD6A/hfHqj7LnBSA/s800/meme-all-the-things.jpg
  46. StommePoes
    12:44
    so it needs this sentiment, how to convey
  47. StommePoes
    12:45
    @michielBijl DOCTOR! I NEED A BRAAAAAIN SPECIALIST!
  48. jkva
    12:45
    <gumby />
  49. MichielBijl
    12:45
    @StommePoes are you a doctor?
  50. jkva
    12:45
    She plays one on tv
  51. MichielBijl
    12:45
    No, no, yes! Yes I am!
  52. MichielBijl
    12:45
    NURSE!
  53. StommePoes
    12:47
    https://en.wikipedia.org/wiki/Gumbys "A popular Gumby catchphrase is "My brain hurts!", which was used in the "Brain Surgery" sketch."
  54. StommePoes
    12:47
    ARE YOU.... THE BRAIN SPECIALIST?
  55. MichielBijl
    12:51
    blob
  56. MichielBijl
    12:51
    alt=radio buttons styled as checkboxes
  57. MichielBijl
    12:51
    Who needs conventions anyway
  58. MichielBijl
    12:52
    Or contrast for that matter
  59. MichielBijl
    12:59
    On alt text for images on Gitter: gitterHQ/gitter#1060
  60. stevefaulkner
    13:00
    This message was deleted
  61. MichielBijl
    13:14
    Ah, of course, you can do markdown formatting for images -.-
  62. MichielBijl
    14:22
    Who should I bug about dead links on this Orca help page: https://wiki.gnome.org/Projects/Orca/FrequentlyAskedQuestions/UsingOrca#What_keyboard_commands_do_I_use_to_control_Orca.3F
  63. StommePoes
    14:26
    Joanie would say, join the docs editors and fix :)
  64. StommePoes
    14:26
    @stevefaulkner is this the W3C HTML5 page for headings? https://www.w3.org/TR/html5/sections.html#the-h1,-h2,-h3,-h4,-h5,-and-h6-elements
  65. StommePoes
    14:26
    I thought there were warning notes in the w3c version
  66. StommePoes
    14:27
    @michielBijl all I know is they use a markdowny thing called Mallard.
  67. MichielBijl
    14:27
    Warnings about what?
  68. StommePoes
    14:27
    about that the document outline (see second code example) isn't real
  69. MichielBijl
    14:27
    That is in the outline algorithm section.
  70. StommePoes
    14:28
    Jeez, not handy at all
  71. StommePoes
    14:28
    Developers look up headings to see how to use them, and it shows, right from the W3C official-looking page, tat you can use all h1's
  72. MichielBijl
    14:28
    https://www.w3.org/TR/html5/sections.html#outlines
  73. MichielBijl
    14:28
    Join the docs editors and fix ;)
  74. powrsurg
    14:34
    Isn't there supposed to be a new template going up for WC3 pages? Any chance the changes aren't just CSS and might fix the markup too?
  75. powrsurg
    14:34
    How would <h1 aria-level="4"> work?
  76. MichielBijl
    14:35
    Work?, heading level 4
  77. MichielBijl
    14:36
    Markup changes to TR's is saved for 2017 AFAIK.
  78. MichielBijl
    14:37
    More info: https://github.com/w3c/tr-design
  79. powrsurg
    14:38
    well, it should work. I just thought AT may infer the level from the tag and not realize the override. That example is convoluted ... which is not out of the realm of a strange dev implementing
  80. MichielBijl
    14:39
    Haven't tested it, but any AT that has aria-level support should surely override it. Same as <a role="button">
  81. MichielBijl
    14:39
    E-mail thread on the 2016 design thing: https://lists.w3.org/Archives/Public/spec-prod/2015OctDec/0009.html
  82. StommePoes
    14:41
    I seem to recall some AT would ignore the aria-level unless role="heading" was explicitly added (even if the thing was already a heading)
  83. StommePoes
    14:42
    It might have been a particular version of JAWS though, not sure.
  84. MichielBijl
    14:43
    VoiceOver on OS X does announce as level that is given in aria-level
  85. StommePoes
    14:46
    That's what it should do.
  86. StommePoes
    14:46
    Whoever it was who needed the heading role, that is a bug.
  87. stevefaulkner
    14:56
    @powrsurg https://github.com/ThePacielloGroup/html5-h#html5-h
  88. MichielBijl
    15:01
    Oh my, how CSS has grown; entire CSS level 1 is 68 pages printed out, flexbox level 1 is 66 pages…
  89. MichielBijl
    15:02
    And then people dare complain about WCAG 🙈
  90. powrsurg
    15:08
    @stevefaulkner that would be great if it was supported and got dev traction. Seems like hell for backwards compatibility ...
  91. MichielBijl
    15:16
    @StommePoes isn't orca 3.19 out yet?
  92. MichielBijl
    15:16
    Still says 3.16 on Ubuntu 15.10 =/
  93. StommePoes
    16:33
    I'm running 3.18 however there's a master branch with the latest stable and it's on GH but I never found a version listed on that one (but it's further along than mine)
  94. StommePoes
    16:33
    Also it goes with Gnome versions and freezes
  95. StommePoes
    16:34
    so after a freeze only bug fixes I think get added so there's like a future branch out there for the next stable Gnome that's in /dev
  96. StommePoes
    16:34
    I'm on ubuntu 14.04
  97. StommePoes
    16:35
    so you should be able to be up to 3.18
  98. StommePoes
    16:35
    3.16 I think still had that annoying list-repeat bug
  99. StommePoes
    16:40
    Yesssss the <h> tag, except more complicated!
  100. MichielBijl
    16:45
    We should make a <k> element for kopje
  101. MichielBijl
    16:49
    Ah, but it needs something before the declaration. Hmm, <o-k> =D
  102. powrsurg
    17:13
    has anyone here every heard of a VPAT?
  103. garcialo
    17:20
    Voluntary Product Accessibility Template? Never heard of it.
  104. garcialo @garcialo runs the other way.
  105. StommePoes
    17:22
    bwahahaha
  106. StommePoes
    17:23
    Companies who make products make VPATs which are docs that are supposed to say how accessible their product is
  107. StommePoes
    17:23
    so their customers can see if they can legally use the products
  108. powrsurg
    17:45
    How would one even legally be able to prove that such statements are true?
  109. powrsurg
    17:45
    I can see how one can prove it to be false. Find something wrong
  110. powrsurg
    17:46
    Dunno if that lets you fix things ever or once you have something wrong once you're screwed ....
  111. garcialo
    17:47
    If the company is doing what they should be, there will be some kind of clause in the contract about a11y
  112. garcialo
    17:47
    The VPAT will be referenced; if the product is found to be completely fabricated it would be messy.
  113. garcialo
    17:48
    For really expensive products, we would just run our own assessments and verify stuff on the VPATs.
  114. garcialo
    17:49
    A lot of the time it was more the case that they didn't have an a11y person complete the VPAT than they were deliberately lying.
  115. deborahgu
    17:57
    My general policy when reading a VPAT is if they have a checkmark in every box and no errors listed, they are lying.
  116. deborahgu
    17:57
    When I create them, I almost use them as bug reports. Do some accessibility testing, find everything that ought to be tagged as an accessibility bug, boil that down as appropriate and put it in the VPAT. And when I have read VPATs, the ones that look most like that are the ones that I trust, because that means they have somebody who is actually looking for and filing accessibility bugs.
  117. deborahgu
    17:58
    But from a legal standpoint I am probably not helping.
  118. powrsurg
    18:03
    Heh. If you do your job and actually show what is wrong then you're legally putting yourself in trouble.
  119. zakim-robot
    18:21
    [karlgroves] All this talk about VPATs! Silly people. You can get a Section 508 waiver!
    http://www.karlgroves.com/2011/07/28/get-a-free-section-508-waiver-here/
  120. MichielBijl
    18:28
    What's the latest on the 508 refresh?
  121. MichielBijl
    18:29
    Seeing as the website is still down.
  122. zakim-robot
    19:00
    [johnbhartley] got bumped until October
  123. MichielBijl
    19:02
    Hmm, Eric Wright might actually be right then.
  124. zakim-robot
    19:05
    [johnbhartley] it's set for the great date of 10/00/2016 per http://www.reginfo.gov/public/do/eAgendaViewRule?pubId=201510&RIN=3014-AA37 via Tenon
  125. MichielBijl
    19:27
    Thank you for the info.
  126. MichielBijl @MichielBijl hates those moments where you are half way through writing a detailed bug report and realise it isn't a bug -.-