Archive index

A11y Slackers Gitter Channel Archive 1st of December 2015

What fresh hell is THIS now? - Patrick Lauke
  1. garcialo
    03:52
    @alice excellent apostrophication
  2. MichielBijl
    06:39
    Added a first test example and some resources on the subject to @a11ydoer's and mine repository: https://github.com/MichielBijl/font-awesome
  3. MichielBijl
    09:02
    Added moar tests and results :sparkles:
  4. zakim-robot
    10:25
    [callumacrae] @MichielBijl: I wasn't counting people here :stuck_out_tongue:
  5. zakim-robot
    10:26
    [callumacrae] @garcialo: I was just going to explain VoiceOver. It's for front-end devs, and according to most polls, 95% of us use OS X anyway
  6. zakim-robot
    10:26
    [callumacrae] But like, the theory can be transferred over to other screen readers pretty easily
  7. MichielBijl
    11:27
    Good point :)
  8. MichielBijl
    11:27
    95%? Really? Woah.
  9. zakim-robot
    11:49
    [dylanb] good morning a11ys
  10. dylanb
    11:50
    @MichielBijl your project looks interesting
  11. MichielBijl
    12:02
    Thank you @dylanb. Hope to expand it with the help of some people. I can't run all the tests myself for example; because I don't have all the AT.
  12. MichielBijl
    12:02
    I've been spamming it all over the webs ;)
  13. MichielBijl
    12:07
    There are some interesting comments on it on the WAI WG list. I should link to that in the repo actually.
  14. MichielBijl
    12:11
    Besides, you raised a valid point and it deserves some research—even if it's a dying technique—because we'll be stuck with it for a while.
  15. Veyfeyken
    12:18
    Theoretical question. If I put a role="presentation" on an empty element, for example <aside role="presentation"></aside>, screenreaders will ignore it completely?
  16. MichielBijl
    12:29
    That would render it the same as a div; so theoretically, yes.
  17. MichielBijl
    13:02

    @Veyfeyken http://www.w3.org/TR/html5/dom.html#index-aria-presentation

    “An element whose implicit native role semantics will not be mapped to the accessibility API.”

  18. MichielBijl
    13:03
    And according to the HTML5 spec a role set to presentation is allowed on aside.
  19. Veyfeyken
    13:08
    thanks
  20. MichielBijl
    13:16
    @Veyfeyken just checked with VoiceOver and Safari, doesn't work :/
  21. MichielBijl
    13:17
    Might be my install, but VO still sees it as a “aside, complimentary”
  22. MichielBijl
    13:17
    This is most likely a bug in VO though.
  23. MichielBijl
    13:18
    Because I explicitly linked it to a heading with aria-labelledby
  24. MichielBijl
    13:18
    If I remove that, it does work.
  25. MichielBijl
    13:19
    So it seems that aria-labelledby somehow overwrites role=presentation…
  26. MichielBijl
    13:23
    Testfile here: http://dir.agosto.nl/accessibility/aside.html
  27. MichielBijl
    13:31
    Bug report: https://bugs.webkit.org/show_bug.cgi?id=151703
  28. MichielBijl
    13:33
    Could someone confirm this?
  29. takenspc
    14:01
    If you put role="presentation" on a element, the element will not be exposed to screen readers. So you should not put aria-* on the element.
  30. takenspc
    14:03
    It seems that WebKit intentionally ignore role="presentation" if a element has aria-*
    http://trac.webkit.org/browser/trunk/Source/WebCore/accessibility/AccessibilityNodeObject.cpp#L2002
  31. StommePoes
    14:03
    JAWS on IE and FF: Aside with presentation role (then it says something else I can't understand in Firefox, sounds like "lage ledbie"
  32. StommePoes
    14:03
    @takenspc yeah makes sense, they often have heuristics to get around assumed likely author-errors
  33. StommePoes
    14:04
    Never mind it was the stupid title
  34. StommePoes
    14:05
    NVDA ignores the aside element as well, only announces the h23
  35. StommePoes
    14:05
    h2
  36. StommePoes
    14:05
    so the windows ones here are doing the right thing
  37. StommePoes
    14:07
    Orca/Firefox ignores as well
  38. StommePoes
    14:07
    A better test would be an aside with something other than "aside" in the heading text :P
  39. MichielBijl
    14:08
    @StommePoes who are you calling stupid?!
  40. StommePoes
    14:08
    Did you read "Never mind it was the title, stupid"? : P
  41. MichielBijl
    14:09
    No, I just really liked that title
  42. MichielBijl
    14:09
    Nah
  43. MichielBijl
    14:09
    I should stop to try and make jokes =(
  44. MichielBijl
    14:09
    Anyhow. Changed the text in the aside.
  45. StommePoes
    14:10
    I MAKE JOKE. IS FUNNY. YOU LAUGH.
  46. MichielBijl
    14:10
    YES
  47. StommePoes
    14:10
    But so VO does the right thing when you don't aria-ize it?
  48. MichielBijl
    14:10
    @takenspc ah, makes sense.
  49. MichielBijl
    14:10
    @StommePoes yes.
  50. StommePoes
    14:12
    "Would you like coffee or tea, Mr. Boole?" "Yes." "I'm sorry, you want coffee AND tea?" "No." "You mean you would like coffee?" "No!" "Um, so you want tea?" "Yes!"
  51. StommePoes
    14:13
    The guy seriously thought this would improve and simplify language : )
  52. MichielBijl
    14:13
    Logic is overrated.
  53. MichielBijl
    14:14
    Thank you @takenspc for pointing that out!
  54. powrsurg
    14:54
    @alice (late reply I know, sorry), but from what I've seen that is not grammatically correct according to most writing guides
  55. garcialo
    14:55
    <span class="sr-only"'</span>
  56. garcialo
    14:55
    er...except with proper tagging =p
  57. powrsurg
    15:03
    Is there a site that has recommendations for working around various screen reader bugs?
  58. powrsurg
    15:04
    Like what was just discussed here
  59. garcialo
    15:04
    Not for working around them as far as I know, but @stevefaulkner has one that is attempting to document how they announce...let's see if I can find it
  60. garcialo
    15:05
    http://thepaciellogroup.github.io/AT-browser-tests/
  61. StommePoes
    15:07
    I have a URL somewhere
  62. StommePoes
    15:08
    www.mindoversight.com/demo/Html5ElementTable.php
  63. StommePoes
    15:08
    which is just forms, but forms are a biggie
  64. MichielBijl
    15:08
    Anyone interested in wasting a lot of time on an ambitious project?
  65. StommePoes
    15:12
    I already have 2 potential time-wasting ambitious projects :/
  66. zakim-robot
    15:19

    [jitendra] i dislike working on admin kind of product

    because nobody cares about markup, accessibility, performance

  67. zakim-robot
    15:19
    [jitendra] but these kind of companies pays well
  68. garcialo
    15:24
    @StommePoes You should make an app that can keep track of a person's queue of projects.
  69. StommePoes
    15:25
    I do have to learn Angular so... : P
  70. garcialo
    15:26
    Let me know when you have it set up, I'll be user #1
  71. dylanb
    15:37
    ok, here is an interesting question for the AIRA wonks
  72. StommePoes
    15:58
    ?
  73. zakim-robot
    16:00
    [lliskovoi] anyone know of good demos/examples of accessible autocomplete search fields?
  74. StommePoes
    16:02
    Heydon had one which seemed to test okay, though it had hard-coded choices.
  75. StommePoes
    16:02
    @goetsu had added a fix http://s.codepen.io/goetsu/debug/wKYRvM
  76. StommePoes
    16:03
    here's the tweet with more context https://twitter.com/goetsu/status/664487110758543360
  77. zakim-robot
    16:04
    [lliskovoi] @StommePoes thanks! I’ll check it out
  78. MichielBijl
    16:06
    Anyone, I want to see if I can get some people to work with me on something like Can I Use.
  79. MichielBijl
    16:06
    Well, not exclusively with me, but you get my point.
  80. MichielBijl @MichielBijl registers caniaria.com :')
  81. MichielBijl
    16:09
    Honestly don't care who makes it, just think it would be good to have :)
  82. StommePoes
    16:16
    If the caniuse people already have some funky backend that does lots for them, I wonder if they would share that. Or maybe it's already on github, I dunno. I don't visit that site much
  83. MichielBijl
    16:22
    Good point, will have to investigate!
  84. MichielBijl
    16:22
    To the investigate mobile!
  85. MichielBijl
    16:28
    @StommePoes only the data.json is shared.
  86. MichielBijl
    16:29
    I can still ask of course :)
  87. zakim-robot
    17:50
    [callumacrae] Does anyone have some good stats about the number of people using AT there are?
  88. zakim-robot
    17:51
    [callumacrae] (for context, writing an article about accessibility)
  89. zakim-robot
    18:14
    [jitendra] Indian Prime Minister Set To Launch 'Accessible India Campaign' For Persons With Disabilities on 3rd December
  90. garcialo
    18:14
    awesome =)
  91. zakim-robot
    18:14
    [jitendra] > The campaign will increase accessibility of buildings, transport system, and enhance usability of information and communication services for people with disabilities
  92. zakim-robot
    18:15
    [jitendra] http://www.huffingtonpost.in/2015/11/30/accessible-india_n_8682206.html
  93. MichielBijl
    18:34
    Very good news Jitendra!
  94. MichielBijl
    18:36
    This message was deleted
  95. MichielBijl
    18:37
    Oeh, we might have winner at https://github.com/MichielBijl/font-awesome
  96. zakim-robot
    18:37
    [callumacrae] I'm going by UK stats, managed to find most things
  97. MichielBijl
    18:37
    Test example 4 seems pretty solid.
  98. zakim-robot
    18:37
    [callumacrae] Once I've finished this article, can I post it in here so someone can tell me if I'm being non-PC?
  99. MichielBijl
    18:38
    @callumacrae nice. Sure :D
  100. zakim-robot
    18:38
    [callumacrae] awesome, cheers
  101. garcialo
    18:38
    Definitely.
  102. MichielBijl
    18:38
    I spam here all day, so don't see a problem with one link ;)
  103. MichielBijl
    18:50
    @StommePoes could you test example 4 with JAWS? https://cdn.rawgit.com/MichielBijl/font-awesome/master/tests/index.html#test-4
  104. garcialo
    18:51
    @MichielBijl that's pretty much exactly what I'm doing on a11yNot? ;)
  105. garcialo
    18:52
    er...without the ?
  106. MichielBijl
    18:52
    I haven't seen a11yNot :O
  107. garcialo
    18:52
    check out the code embed http://a11ynot.com/nots/textLooksLikeaHeading.html
  108. garcialo
    18:53
    specifically the "hosted with..."
  109. MichielBijl
    18:53
    Yep
  110. MichielBijl
    18:53
    It is :)
  111. garcialo
    18:53
    I will say it works well
  112. MichielBijl
    18:53
    Same use case really.
  113. MichielBijl
    18:53
    Do you have test results?
  114. garcialo
    18:54
    I manually tested it =p
  115. zakim-robot
    18:54
    [callumacrae] It's reeaally difficult to talk about situational vs permanent disabilities without trivialising permanent disabilities
  116. MichielBijl
    18:54
    Did you document that testing :P
  117. garcialo
    18:54
    no, the icon font a11y was incidental for me
  118. MichielBijl
    18:55
    Ah :(
  119. garcialo
    18:55
    the real point was the test cases I was creating
  120. garcialo
    18:56
    and right now i'm more focused on issues found by automated testing than on those found by manual testing
  121. MichielBijl
    18:56
    Bookmarked a11yNot, very nice resource.
  122. garcialo
    18:56
    well, at least it will be once I have some good data in there
  123. garcialo
    18:56
    still need to fix up all the W3 failure examples
  124. garcialo
    18:56
    there...are a lot of them
  125. MichielBijl
    18:57
    Haven't clicked on all the links to be honest. If you need help, give a shout.
  126. garcialo
    18:57
    Help is always appreciated.
  127. garcialo
    18:58
    I need to create examples where tools usually throw false positives
  128. garcialo
    18:58
    I also need to go through and verify I have tagged the current nots properly with what guidelines apply
  129. deborahgu
    19:16
    NVDA doesn't read before and after pseudo elements. Is that common behavior? Given that web designers will insist on using them (to hold meaningful content. the gods wept) what's best practice to get that content announced?
  130. deborahgu
    19:20
    I suppose best practice in this case would either be to say "don't use ::before" or "if you are going to use ::before, you also have to have a visually hidden offscreen replacement for that text". Unless anyone has better ideas?
  131. MichielBijl
    19:20
    VoiceOver reads generated content (which should be named generated decoration).
  132. deborahgu
    19:22
    I was just seeing an article about that
  133. LjWatson
    19:23
    I wrote about CSS generated content a while back http://tink.uk/accessibility-support-for-css-generated-content/
  134. deborahgu
    19:23
    it seems like if you use it for decorative, you have to know the voiceover will read it. And if you use it for meaningful content, you have to know that other screen readers won't.
  135. MichielBijl
    19:24
    The project I'm working on tries to solve that for icon fonts :)
  136. deborahgu
    19:25
    huh, I wasn't getting to read in Firefox with NVDA. On your chart, Leonie, it looks like it should have worked.
  137. deborahgu
    19:25
    In any case there shouldn't be semantic content in the CSS layer, as you say
  138. jnurthen
    19:50
    @deborahgu Did you ensure you started NVDA before FF?
  139. deborahgu
    19:51
    no, I didn't. I didn't realize that makes a difference.
  140. jnurthen
    19:52
    FF starts up in a different windowing mode when AT is running. Results are always strange if you do things in the wrong order
  141. deborahgu
    19:53
    ah! learn something new every day. and that makes perfect sense; I know that when I have to restart Dragon I often want to restart most of my other applications as well.
  142. deborahgu
    19:53
    thanks
  143. jnurthen
    19:55
    I can't find the bug right now but I have logged this with them in the past
  144. MichielBijl
    20:02
    That's good to know.
  145. MichielBijl
    20:27
    Does anyone have any data on AT support for complex tables? Like in this WebAIM article: http://webaim.org/techniques/tables/data
  146. MichielBijl
    20:28

    It has a note:

    Despite being standard markup for tables for many years, some screen readers still do not fully support complex tables with spanned or multiple levels of row and/or column headers. When possible, try to 'flatten' the table and avoid spanned cells and multiple levels of header cells.

  147. MichielBijl
    20:28
    But doesn't link to any test results.
  148. mejsullivan
    21:28
    Hey Michiel, I know it works alright on JAWS 14 ie8
  149. mejsullivan
    21:28
    Can't help more than that
  150. mejsullivan
    21:29
    jumping in here with a different question
  151. mejsullivan
    21:30
    Does anyone have experience with aria-descriedby attributes with knockout?
  152. MichielBijl
    21:32
    Thanks @mejsullivan.
  153. MichielBijl
    21:32
    You mean example 4?