Archive index

A11y Slackers
Gitter Channel Archive 29th of July 2015

What fresh hell is THIS now? - Patrick Lauke
  1. indrekpaas
    Jul 29 01:09
    Hi! Does anyone know, what category the following ARIA attributes belong based on the W3C taxonomy ( http://www.w3.org/TR/wai-aria-1.1/#state_prop_taxonomy )? aria-current, aria-describedat and aria-roledescription can only be found in the http://www.w3.org/TR/wai-aria-1.1/#global_states list, but are not currently categorised any further.
  2. Scratch2k
    Jul 29 01:30
    @Heydon I know I'm a few hours late but have you seen this one http://freqdec.github.io/datePicker/demo/
  3. zakim-robot
    Jul 29 01:56
    [heidi valles, a11y] ARIA for a button that changes its own button text when clicked (ex. "add" changes to "remove"): aria-live="polite" (would that work?)? role="button" aria-pressed="false" doesn't seem to quite fit as both variations seem press-able. Thoughts?
  4. zakim-robot
    Jul 29 02:00
  5. stevefaulkner
    Jul 29 04:31
    @indrekpaas hi, unclear what you are asking, they are global properties which can be used (when implemented) "The following global states and properties are supported by all roles and by all base markup elements."
  6. indrekpaas
    Jul 29 04:36
    @stevefaulkner I guess I just found it odd that out of 46 properties, these 3 didn't belong to a (sub)category with the rest
  7. stevefaulkner
    Jul 29 04:41
    @indrekpaas ah, I got you now, probably an oversight suggest filing a bug https://www.w3.org/Bugs/Public/enter_bug.cgi?product=ARIA&component=Spec&version=1.1
  8. indrekpaas
    Jul 29 04:42
    @stevefaulkner Cheers, will do
  9. zakim-robot
    Jul 29 08:05
    [estelle, a11y] is listbox an OK role for a text <input>
  10. StommePoes
    Jul 29 08:06
    marco thought it was good for a container of a text box if there was autocomplete lists afterwards https://www.marcozehe.de/2014/03/11/easy-aria-tip-7-use-listbox-and-option-roles-when-constructing-autocomplete-lists/
  11. zakim-robot
    Jul 29 08:06
    [estelle, a11y] scenario is we have a listbox that when somehting is selected, the value goes into the text input. My understanding is that the option list parent is the listbox, and the input is a combo box
  12. zakim-robot
    Jul 29 08:06
    [estelle, a11y] but we can’t actually do data entry on the input
  13. zakim-robot
    Jul 29 08:07
    [estelle, a11y] Was just reading that, but it’s not quite a fit
  14. StommePoes
    Jul 29 08:09
    is this input kinda like the HTML5 one with the lists attribute?
  15. StommePoes
    Jul 29 08:09
    <label for="se">Favourite search engine</label> <input type="url" list="engines" name="se" id="se" placeholder="Select one or enter your own" aria-required="true" required="">
  16. StommePoes
    Jul 29 08:09
    ?
  17. StommePoes
    Jul 29 08:10
    with a datalist element following the input with the combo-box-like list to select?
  18. StommePoes
    Jul 29 08:10
    (that's quite old HMTL btw, I didn't just type it)
  19. zakim-robot
    Jul 29 08:16
    [estelle, a11y] I think the listbox is used incorrectly. I just want to confirm that it is indeed incorrect
  20. zakim-robot
    Jul 29 08:16
    [estelle, a11y] It’s basically a select, but not using <select> for appearance reasons.
  21. zakim-robot
    Jul 29 08:17
    [estelle, a11y] it has to look like all the input fields around it.
  22. StommePoes
    Jul 29 08:17
    you're not doing the div-trick bit? (have a select and overlay a pretty div on top for sighties)
  23. StommePoes
    Jul 29 08:18
    I actually dunno if any of those tricks worked everywhere
  24. zakim-robot
    Jul 29 08:20
    [estelle, a11y] no. It seems to be a correct <div> with listbox with childrend that have role=option. When a selection is made it populates an input, and that input also has listbox
  25. zakim-robot
    Jul 29 08:21
    [estelle, a11y] i think the role of listbox on the input is incorrect. Not sure what it should be though
  26. StommePoes
    Jul 29 08:21
    so you wonder, double listbox roles
  27. StommePoes
    Jul 29 08:21
    maybe no role, remain as a text input?
  28. zakim-robot
    Jul 29 08:21
    [estelle, a11y] yes,
  29. zakim-robot
    Jul 29 08:21
    [estelle, a11y] with the input being an empty element, i don’t see how it can be a listbox
  30. StommePoes
    Jul 29 08:22
    I have some jQuery UI autocomplete bit that adds a role of textbox to some mangled input, Tenon called it out... have any tools said anything about this input?
  31. StommePoes
    Jul 29 08:22
    As in, maybe a tool can reference somewhere stating that it might not be correct.
  32. StommePoes
    Jul 29 08:22
    (prolly not, but who knows)
  33. zakim-robot
    Jul 29 08:22
    [estelle, a11y] just looked at http://mikemaccana.github.io/styleselect/ - issue with styleselect is that the select with “aria-hidden=‘false’” is that it also has display none, so i guess that never gets read
  34. StommePoes
    Jul 29 08:23
    AT-wise, is there different behaviour between the extra role and same thing with no role on the input? (that might matter more than what's correct for now as well?)
  35. StommePoes
    Jul 29 08:23
    Yeah though perhaps they tested in older AT which would sometimes ignore display: none on form controls
  36. StommePoes
    Jul 29 08:23
    and maybe thought the aria-hidden false was good-enough.
  37. zakim-robot
    Jul 29 08:23
    [estelle, a11y] i am thinking the input should be aria-hidden, and the actual listbox should have all the a11y features
  38. StommePoes
    Jul 29 08:24
    until a choice is made?
  39. StommePoes
    Jul 29 08:24
    cause then of course you want to see the input has your choice.
  40. zakim-robot
    Jul 29 08:26
    [estelle, a11y] thanks for the feedback. giving a talk in a few hours so must head to bed.
  41. StommePoes
    Jul 29 08:26
    sorry your sole feedback had to come from me, lawlz
  42. StommePoes
    Jul 29 08:26
    guess others aren't awake yet
  43. stevefaulkner
    Jul 29 08:35

    A11ySlackers chat log 28th of July 2015 - http://w3c.github.io/a11ySlackers/archives/280715.html

  44. stevefaulkner
    Jul 29 08:40
    @estelle had a quick look needs work accessibility tree for the widget is borked
  45. stevefaulkner
    08:58
    aria-hidden="false" will not work on an element that is hidden, in fact aria-hidden=false is same as not having the attribute http://www.w3.org/TR/core-aam-1.1/#ariaHiddenFalse i.e. it is not mapped to anything
  46. Heydon
    09:01
    @Scratch2k Thanks! Pretty robust, but I want to avoid using key bindings that the user will have to guess.
  47. StommePoes
    09:02
    stevefaulkner because it's a boolean attribute
  48. StommePoes
    09:02
    required="false" should also be "required"
  49. StommePoes
    09:02
    kinda sucks in JS tho, removing attributes is different than changing their values
  50. StommePoes
    09:03
    but @stevefaulkner it's still possible whoever built the display: none form control + aria-hidden="lawlz" still did a test with some SR and found the element on the page fine, right?
  51. StommePoes
    09:04
    if that's still a thing nowadays
  52. stevefaulkner
    09:08
    @StommePoes don't believe so as its hidden using display:none, but anythings are possible
  53. StommePoes
    09:09
    @steve I mean the way it was was, at least for JAWS, display: none was an exception with form controls, Gez had some tests showing they could still be accessed. Old stuff though.
  54. StommePoes
    09:09
    I haven't checked in years
  55. stevefaulkner
    09:16
    @StommePoes did the obvious thing and checked demo in JAWS and NVDA with Firefox... No sign of any select element on the page and the fake select is simply ignored
  56. stevefaulkner
    09:21
    "people using screen readers simply experience the regular, unstyled select box." - http://mikemaccana.github.io/styleselect/ does not appear to be the case
  57. StommePoes
    09:22
    is that because of the aria-hidden or because of the display: none?
  58. StommePoes
    09:22
    still, looking at his code, I can't tell why would he put aria-hidden on a control supposedly meant for SRs
  59. stevefaulkner
    09:22
    display: none
  60. stevefaulkner
    09:23
    he has put aria-hidden=false right? in an attempt to expose the display:none content to AT
  61. StommePoes
    09:23
    I think chosen plugin does the display: none bit too
  62. StommePoes
    09:23
    oh aha he thought aria-hidden false would somehow "undo" what display none does
  63. StommePoes
    09:23
    yeah ok, I wouldn't have expected it to
  64. StommePoes
    09:24
    I expect, if aria-hidden acts like a boolean, that you have to remove it. And if it instead does reflect its value, then a developer would only have aria-hidden=false in a case where earlier he had aria-hidden="true" and due to user engagement (click or whatever) this needed to change to "no longer hidden"
  65. StommePoes
    09:24
    but not that it otherwise does anything.
  66. stevefaulkner
    09:25
    yeah, well it was sort of an idea that it would at one point but browser implementers not on board with it https://bugzilla.mozilla.org/show_bug.cgi?id=945194 aria-hidden behaviour in general has been a bone of contention with implementers...
  67. StommePoes
    09:26
    Originally it was "the aria-ish equivalent to the hidden attribute"
  68. StommePoes
    09:26
    where inputs with type=hidden got the same treatement as script, head, link tags
  69. StommePoes
    09:27
    Another reason for having browsers reflect aria and not only to SR users, aria-hidden would be way more handy if it could prevent focus events
  70. stevefaulkner
    09:27
    that's heresy
  71. StommePoes
    09:28
    I'm on 9292.nl trying to get public transit times, and there's this hey-we-use-cookies popup, and I can't get my keyboard to it. Instead, I can get to the whole rest of the page I can't even see :/
  72. StommePoes
    09:28
    native landmark keyboard navigation-- Less Important Than Navigation Transitions
  73. StommePoes
    09:28
    (i didn't even know navigation trasitions were a thing until PPK mentioned them)
  74. stevefaulkner
    09:29
    if you want to stay friends with browser engineers don't mention extending ARIA to general behaviours, landmark nav is one possibility where browser vendors may budge
  75. StommePoes
    09:30
    yeah better to let browsers add popular BS while Javascript developers get to do the same stuff two or three times (add aria to page and then script behaviour that we could have gotten for free afterwards)
  76. StommePoes
    09:31
    you'd think the aria-hidden==hidden thing would consolidate code possible
  77. stevefaulkner
    09:31
    right...
  78. StommePoes
    09:31
    possibly
  79. StommePoes
    09:31
    but I understand how the landmark navigation can actually be lots of work with bugs popping up
  80. StommePoes
    09:32
    ok I'm going to go rant on the twitters at 9292.nl and see if they won't help fix their popup.
  81. stevefaulkner
    09:32
    the issue with aria-hidden=false is that it is complex to expose bits of the tree to acc layer when it is hidden in DOM
  82. stevefaulkner
    09:33
    or so I have been told by browser acc engineers
  83. StommePoes
    09:34
    even it it meant simply the equivalent of removeAttribute?
  84. StommePoes
    09:34
    like, does different stuff happen if a dev with JS simply removes an aria-hidden=true bit from an element with descendents (and let's say it's not hidden with css)
  85. stevefaulkner
    09:37
    @estelle in essence the pseudo select element needs to implement the ARIA listbox design pattern http://www.w3.org/TR/wai-aria-practices-1.1/#Listbox
  86. stevefaulkner
    09:38
    @StommePoes dunno?
  87. StommePoes
    09:39
    @steve and so if it is not implementing these things (the plain non-listy text input) then it is indeed wrong for it to have that role... sounds like the answer to what she was asking.
  88. StommePoes
    09:39
    @stevefaulkner I would hope it did (aria-hidden being removed) since most of us use it in the popup-dialog use case (for the Rest-of-page)
  89. stevefaulkner
    09:40
    @StommePoes am confuzled by what you are asking, code example?
  90. StommePoes
    09:40
    Sure
  91. StommePoes
    09:41
    <div aria-hidden=true>lotsa stuff</div>
  92. StommePoes
    09:41
    <div role=dialog>dialoggie stuff, you only want this available to all users right now</div>
  93. StommePoes
    09:41
    so if this is your initial state. Everyone in the main page (who has the aria-hidden) should be hidden-hidden. Visually maybe we throw another div on top with a sexy grey veil
  94. StommePoes
    09:42
    when the user closes their dialog the dev does a removeAttribute and we should get
  95. StommePoes
    09:42
    <div>lotsa stuff</div>
  96. StommePoes
    09:42
    <div>the old dialog is maybe removed, destroyed, whatever...</div>
  97. StommePoes
    09:42
    it should be now that the tree of the hidden div should now be available.
  98. StommePoes
    09:42
    it no longer has aria-hidden=true on it.
  99. stevefaulkner
    09:42
    yeah and it does i believe
  100. StommePoes
    09:43
    My question was, is it really a problem to have aria-hidden="anything here that's not true" be equivalent to no aria-hidden at all?
  101. stevefaulkner
    09:44
    i don't think so, they are funcitonally equivalent in that aria-hidden=false does not map to anything
  102. StommePoes
    09:44
    devs are doing that with disabled attributes too, some tell others "you can just set the string to false" and others say "just remove the whole attribute" but since that is about form controls, we have differences between properties and attributes. Divs with aria-hidden should just have attrbutes I believe.
  103. StommePoes
    09:44
    OK. that sounds good, actually
  104. stevefaulkner
    09:45
    disabled is one of the attribues that does not have true/false when present it works
  105. StommePoes
    09:45
    if mapping to nothing is equivalent to "random attribute browser ignores because it means nothing" then all devs have to worry about/keep in mind is not to do what that select-plugin guy did and think aria-hidden=false will somehow undo something like display none
  106. StommePoes
    09:45
    yeah it's truely boolean
  107. stevefaulkner
    09:46
    truelean boolean or truely booly
  108. StommePoes
    09:46
    You ought to read the Georg Bool joke from the Lovelace and Babbage comic.
  109. StommePoes
    09:51
  110. stevefaulkner
    09:54
    :smile:
  111. StommePoes
    09:56
    "coffee or tea?" "Yes" "You mean, coffee AND tea??" "No!" "Do you mean coffee?" "No!" "So you mean tea?" "Yes." ... later, the minion comes back and starts to ask about cream or sugar... Boole sweats nervously
  112. stevefaulkner
    10:01
    yeah i presumed boole could only answer yes/no right?
  113. StommePoes
    10:02
    Well, in Sydney's twisted version, he can only answer with boolean logic, yeah. But she does it so funny
  114. StommePoes
    10:03
    the whole book is recommend++,
  115. StommePoes
    10:35
    ah, the SEO Specialist is insisting that "h2's in the footer [and above some filter options in the sidebar] are confusing for Google" and that they should be removed "but the styling can remain the same" (the h2's are styled like headings, of course)
  116. StommePoes
    10:36
    I guess he didn't even check that I had caved in months ago after burning out arguing with them and sending them examples of sites that had no Google problems whatsoever with proper heading structure and even a link to the WCAG Failure example... and had already turned them into <p>'s.
  117. StommePoes
    10:36
    Ug, time to crawl back into bed.
  118. StommePoes
    10:36
    I hate SEO.
  119. StommePoes
    12:58
    yay-- twitter talk with our most-popular public transport website/app states the inaccessible cookie popup problems will be given to the devs and that "Accessibility is a banner issue for us" (i that's a decent translation). :)
  120. MichielBijl
    13:01
    What's a “Banner issue”
  121. MichielBijl
    13:01
    I guess that's good?
  122. StommePoes
    13:05
    "hoog in het vaandel"
  123. StommePoes
    13:05
    important-to-them I suppose...
  124. MichielBijl
    13:05
    Right :)
  125. StommePoes
    13:05
    the "i" above was supposed to be "if"
  126. dylanb
    13:16
    @StommePoes why don't you use role=heading and aria-level=2?
  127. StommePoes
    13:24
    I actually didn't think of that (because usually I can use correct HTML semantics)
  128. techthomas
    13:24
    Does anyone know the best way to log a bug on Freedom Scientific JAWS? Most of the responses we've gotten have been "look at the latest release notes" etc etc
  129. StommePoes
    13:24
    that might be a good idea
  130. stevefaulkner
    13:43

    Does anyone know the best way to log a bug on Freedom Scientific JAWS? Most of the responses we've gotten have been "look at the latest release notes" etc etc

    @LjWatson is a good person to ask

  131. kaleb
    14:35
    When should one use [role=textbox aria-autocomplete=list] vs [role=combobox aria-autocomplete=list]?
  132. techthomas
    14:49
    @stevefaulkner thanks steve! @ljwatson if you have any tips would sincerely appreciate
  133. ckundo
    14:51
    hey folks, we’re hosting an Accessibility Camp on September 26th in Brooklyn, NY.
  134. ckundo
    14:51
    This message was deleted
  135. ckundo
    14:51
    Speaker signup is at http://goo.gl/forms/ReAvqd49hx
  136. ckundo
    14:52
    ohai embedded form
  137. techthomas
    14:54
    Hi @ckundo :) I can't wait for Accessibility Camp in NYC!
  138. ckundo
    14:59
    @techthomas :thumbsup:
  139. zakim-robot
    15:15
    [Karl Groves, a11y] @ckundo w00t
  140. zakim-robot
    15:16
    [Karl Groves, a11y] Karl Groves current status: putting a11y camp NYC on the schedule
  141. StommePoes
    15:59
    ha, I wonder if the googles can read role="heading" aria-level="blah"...
  142. StommePoes
    15:59
    if they can, and the SEOers get wind of it, then even that won't save me
  143. dylanb
    16:36
    I don't think they do
  144. MichielBijl
    16:37
    StommePoes: just add it in with javascript later :P
  145. MichielBijl
    16:37
    That was sarcastic if anyone missed that…
  146. StommePoes
    16:38
    lawlz... there's already the snarky class names live vanbeekdesign.nl (footer), but now they've requested the last h2 gone and also one I've visually hidden because a chunk of content never had a visible heading but one belongs there.
  147. StommePoes
    16:38
    and the last remaining h2 in teh left column when there's a left column, whenever someone pulls those to the live site.
  148. StommePoes
    16:39
    I gave a long list of sites that properly use headings in fat-footers. Sites who clearly have no trouble with "confused googles". Didn't matter.
  149. StommePoes
    16:45
    wehkamp.nl, bol.com, hema.nl, bbc.co.uk, target.com, rabobank.nl, nos.nl
  150. MichielBijl
    16:46
    hema.nl is terrible
  151. MichielBijl
    16:46
    Try to tab through that bad boy…
  152. StommePoes
    16:46
    yeah the menu, god
  153. StommePoes
    16:47
    but these were the list of "look they have headings in their fat-footers and are not dying from lack of google juice"
  154. MichielBijl
    16:47
    I told them repeatedly when I worked on projects for them, but it didn't help
  155. StommePoes
    16:47
    tabbing was also an issue with a certain large blue e-commerce company
  156. StommePoes
    16:48
    no way to even get to the menu, which is as bad as having too much menu-itis.
  157. StommePoes
    16:50
    I was told to imitate another big-blue-ecommerce company's menu (and now these two big blue ecommerce companies have very similar-looking menus) for a Belgian company and while it still has hiccups, I like to think it's much more usable (took me way to long to figure out the JS for it, but was worth it).
  158. StommePoes
    16:51
    I'll send you a tweet, ask you what you think of it (even tho I can't change much by now, it's launched already)
  159. MichielBijl
    16:54
    StommePoes tabbing doesn't seem to bad there
  160. MichielBijl
    16:55
    I'll look into it after the PFWG meeting
  161. MichielBijl
    19:26
    StommePoes: what exactly do you want me to look at?
  162. MichielBijl
    19:26
    Kind of busy with sending out job applications :o
  163. zakim-robot
    20:02
    [jitendra, a11y] I kept changing jobs but unable to find product company in my country which cares about accessibility. In India startups focus in growth not product quality and accessibility and companies are able to make money without having worry about accessibility.
  164. zakim-robot
    20:04
    [jitendra, a11y] If i want to work in accessibility the only thing I can do is to join a service based companies and find and fix accessibility problem on their client’s websites and for that too pay is not good.
  165. zakim-robot
    20:05
    [jitendra, a11y] Opportunities to design and build accessible web product from scratch are not available, atleast in my country
  166. MichielBijl
    20:07
    jitendra: no mature product companies at all?
  167. zakim-robot
    20:10
    [jitendra, a11y] Definition of mature?
  168. MichielBijl
    20:10
    4-5 years old?
  169. MichielBijl
    20:10
    Out of the start-up phase
  170. zakim-robot
    20:11
    [jitendra, a11y] We have many well funded and popular startups but they do not care about Web and Accessibility
  171. zakim-robot
    20:12
    [jitendra, a11y] In India “Android first” is the most preferred approach by startups these days
  172. techthomas
    21:14
    @jitendra do you know the peope at BarrierBreak? They might be good people to ask about local opportunities
  173. zakim-robot
    23:53

    [Marcy Sutton, a11y] Recursive A11y-Slackers plugging http://a11ywins.tumblr.com/post/125378431293/a11y-slackers

    so meta :troll:

  174. indrekpaas
    00:28

    @stevefaulkner

    in fact aria-hidden=false is same as not having the attribute

  175. indrekpaas
    00:29
    But I still would use it in tests for example, where false value is much easier to validate rather than the absence of the entire attribute
  176. zakim-robot
    00:52
    [Tarah, a11y] Started testing VO, these are the results
    http://codepen.io/Tarah-S/live/VLVzXX