Archive index

A11y Slackers Gitter Channel Archive 24th of November 2015

What fresh hell is THIS now? - Patrick Lauke
  1. MichielBijl
    Nov 24 00:03
    Looks cool!
  2. MichielBijl @MichielBijl adds to reading list
  3. MichielBijl
    Nov 24 00:03
    :(
  4. StommePoes
    07:56
    That media query for high contrast is also in IE, has been for several versions now
  5. stevefaulkner
    12:35
    :point_up: November 23, 2015 11:30 PM @MichielBijl of course, it just makes it diffcult to have a discussion when a particpant in the discussion is not able to post to the list, and its a silly rule that is not reflected in other WG lists.
  6. MichielBijl
    13:07
    That is certainly true. Having said that, @rodneyrehm, Florian replied on the thread :+
  7. MichielBijl
    13:07
    It
  8. MichielBijl
    13:14
    It's weird that the a11y list is not accessible unless you're a member. Wouldn't it be possible to change that with the new ARIA/APA groups?
  9. LjWatson
    13:50
    @MichielBijl AFAIK it should be possible for anyone to join the new APA and ARIA lists.
  10. LjWatson
    13:52
    Hmm. I stand corrected. It doesn't seem to be possible.
  11. stevefaulkner
    14:07
    @LjWatson i thought this was one of the things that had been worked out?
  12. LjWatson
    14:08
    @stevefaulkner me too. Have sent email to team-aria@w3.org and public-aria@w3.org
  13. LjWatson
    14:08
    Will do likewise with APA when I get a response on ARIA.
  14. stevefaulkner
    14:10
    @LjWatson @MichielBijl https://lists.w3.org/Archives/Public/public-aria/2015Nov/0000.html
  15. stevefaulkner
    14:10
    i just posted a test mail to public-aria@w3.org seemed to work OK
  16. stevefaulkner
    14:11
    I think @MichielBijl was posting to old pf list
  17. LjWatson
    14:12
    There doesn't seem to be a subscribe link on the email archive page though?
  18. stevefaulkner
    14:12
    oh, that's unhelpful ;-)
  19. StommePoes
    14:13
    Guys, I read somewhere fairly recently about sometimes a non-native-clickable with keystroke listeners (for space and enter for example) might not get those keystrokes passed to them by some screen reader/browser combo's, does anyone know where I might have read this?
  20. dylanb
    14:13
    keeps the list nice and small ;-)
  21. dylanb
    14:14
    @StommePoes http://unobfuscated.blogspot.com/2013/05/event-handlers-and-screen-readers.html
  22. dylanb
    14:15
    it is not comprehensive, but should give you an inkling of what is going on
  23. stevefaulkner
    14:15
    @dylanb yeah its another method to limit public comment
  24. LjWatson
    14:19
    @stevefaulkner think your post to public-aria whent through because your TPG email is associated with your W3C account?
  25. StommePoes
    14:19
    @dylanb that wasn't it BUT this much more useful than what I'd found, thanks : ) I'm trying to prevent someone from adding an application role if we can avoid it
  26. stevefaulkner
    14:20
    @LjWatson thought my gamil one was not tpg but could be wrong
  27. stevefaulkner
    14:21
    @StommePoes can you send test email to public-aria@w3.org assuming you are not a member?
  28. LjWatson
    14:21
    @stevefaulkner would need to check but believe both your addys are. This particular email came from your TPG email though.
  29. stevefaulkner
    14:24
    yeah i sent from tpg as ithought it wasnt connected
  30. garcialo
    14:24
    This is why everyone need at least a dozen email addresses.
  31. LjWatson
    14:24
    @garcialo Ha!
  32. StommePoes
    14:25
    @steveFaulkner sent
  33. stevefaulkner
    14:27
    @StommePoes thanks and it got through!
  34. dylanb
    14:29
    @StommePoes application role FTW!
  35. LjWatson
    14:32
    @StommePoes how did you subscribe to the ARIA list?
  36. StommePoes
    14:33
    Leonie I didn't. Steve asked me to send a mail, as I'm not a member, to test the system.
  37. StommePoes
    14:33
    @dylanb would you recommend it for drag and drop? I figure there's so much other texty stuff in those things, I'd rather not...
  38. LjWatson
    14:33
    Ok, so you're not receiving emails presumably?
  39. StommePoes
    14:34
    Oh I wouldn't know, when I sign up for lists usually I only get the next mail sent, not history.
  40. StommePoes
    14:34
    So, I could get mails
  41. StommePoes
    14:34
    I just sent that mail.
  42. StommePoes
    14:34
    Like, a minute ago.
  43. StommePoes
    14:34
    If I get a mail I'll let you know <grin>
  44. LjWatson
    14:34
    Yes, but you're not subscribed to the list - so you won't receive news emails sent to the list in future?
  45. dylanb
    14:34
    @StommePoes other than when writing, say Google Docs, I would not say 'recommend', I would say 'tolerate'
  46. LjWatson
    14:35
    Thanks :)
  47. dylanb
    14:35
    There are some things that just cannot be done without it
  48. dylanb
    14:36
    and that is the general rule of thumb, if you cannot do something without the application role, then you might be justified in using it...
  49. StommePoes
    14:37
    Well, I was kinda hoping to just add a click listener instead
  50. StommePoes
    14:37
    There are already key listeners for space/enter, and there are mouseup/down listeners for the drag/drop stuff, but since FF sends those through, FF works currently.
  51. StommePoes
    14:37
    And I can get JAWS to work if I manually into forms mode in IE.
  52. StommePoes
    14:38
    I want to see though if I can just pass keystroke onto the click, that oughtta work.
  53. dylanb
    14:38
    what are you trying to drag and drop?
  54. StommePoes
    14:39
    @LjWatson I wasn't subsribed to the list before, if sending that mail subscribes me, that's ok. Mutt's so great an email client I can deal with deluges of mail <grin>
  55. StommePoes
    14:40
    @dylanb inert elements filled with various things. Some of these things are in SVGs which I'm thinking, just forget it for now, others are just inerts (spans, divs, whatever) with text etc in list items.
  56. LjWatson
    14:40
    @StommePoes no it should just mean you've sent the list an email. It should also be possible for non-members to subscribe to the list (and receive all emails), but this isn't possible right now.
  57. StommePoes
    14:40
    I can't see a reasonable way to make these links (they don't go anywhere) or buttons (button really seems to be stretching it)
  58. dylanb
    14:40
    @StommePoes ok, that is a good use case for application role
  59. StommePoes
    14:40
    Hm.
  60. MichielBijl
    14:40
    @stevefaulkner @LjWatson I was unaware that there was a new list.
  61. StommePoes
    14:40
    Unless I can get click working and can avoid it : P
  62. dylanb
    14:41
    you also have to get the arrow keys working - no?
  63. StommePoes
    14:41
    @LjWatson ah, maybe @steveFaulkner's test was kinda useless then : )
  64. LjWatson
    14:41
    StommePoes it was helpful because it proved it's possible for anyone to post to the list, which is an improvement over the old ways.
  65. StommePoes
    14:42
    @dylanb the current setup is simpler than brotherecake's drag and drop. We have instructions to use space (tho enter would work) to select an element, then navigate however you want (tab works, or whatever) to the destination, enter to release it. (I didn't build this, I'm trying to fix it)
  66. LjWatson
    14:42
    @MichielBijl have you transitioned your IE status to the new WGs yet?
  67. StommePoes
    14:42
    @LjWatson cool
  68. MichielBijl
    14:43
    @LjWatson that is up to Mr. Cooper as I'm an invited expert.
  69. LjWatson
    14:43
    @MichielBijl have you contacted him about transitioning though? As IE you should do that proactively.
  70. dylanb
    14:43
    @StommePoes ok, makes sense. Having to enter forms mode manually is terrible usability
  71. MichielBijl
    14:44
    @LjWatson yeah I asked him about it off-list. He send an e-mail to all IEs last week asking which WGs they want to join. I said both. Haven't heard anything since. I'll sent him a reminder.
  72. StommePoes
    14:44
    @dylanb yeah. People should be able to follow the directions whether they've got an SR or not.
  73. StommePoes
    14:45
    (which, it does, in FF. But not in IE : )
  74. LjWatson
    14:45
    @MichielBijl good idea. There are a lot of IE people in PF that will need transitioning. Michael will have a lot to do!
  75. MichielBijl
    15:02
    @LjWatson he's dependent on other busy people (he just replied). So, we'll wait and see.
  76. powrsurg
    15:17
    Should aXe really be complaining about child nodes (<td>s and <th>s) not having appropriate roles when their parent a <tr> is given role="row" by a JavaScript library?
  77. powrsurg
    15:17
    or is there actually something wrong about that?
  78. powrsurg
    15:19
    I'm aware that the role the JS library is adding is unnecessary, but is that more of a false positive by aXe since one would normally expect when you have to add that role you're doing it on <div>s or some other non-tabular element
  79. stevefaulkner
    15:24
    @powrsurg just pinged godfather of ARIA either explicit or implicit roles are acceptable
  80. powrsurg
    15:24
    I thought so, so these are false-positives in aXe
  81. stevefaulkner
    15:25
    so <tr><th role=columnheader is OK, but not recommended
  82. powrsurg
    15:26
    Yeah, I wouldn't do it normally. I'm going to put in a comment to the library author that <tr role="row"> is unnecessary
  83. powrsurg
    15:29
    Oh yeah, further question. We have a button that adds and removes another element from the DOM (not just shows/hides). Should aria-controls be on all the time, or just when the elements are present on the DOM?
  84. stevefaulkner
    15:32
    @powrsurg its only meaningful when the element is present and navigable, but don't know whether it is reported when not present
  85. MichielBijl
    15:35
    Silly me, trying to use aria-expanded on a div…
  86. powrsurg
    15:36
    Right, I'm trying to determine if the DOM manipulation is necessary.
  87. powrsurg
    15:37
    @MichielBijl technically that is valid, but I am in the @stevefaulkner camp where I think it makes more sense when used on a control, rather than the element being controlled
  88. MichielBijl
    15:37
    It's only valid if it has a role: http://www.w3.org/TR/wai-aria-1.1/#aria-expanded
  89. MichielBijl
    15:38
    So putting it on a plain div is invalid.
  90. MichielBijl
    15:38
    Either way, I don't have a clear preference and thus follow the spec.
  91. MichielBijl
    15:39
    It'll be craptastic anyway, because the thing is an accordion on mobile but not on desktop. So hesitant to use details.
  92. MichielBijl
    15:39
    Besides the missing implementations in IE/Edge/Firefox
  93. MichielBijl
    15:39
    Although FF was/is working on it.
  94. MichielBijl
    15:39
    Same for Edge I believe.
  95. powrsurg
    15:40
    okay, technically we only ever had it on divs with roles, fair enough
  96. MichielBijl
    15:40
    .matchMedia to the rescue! Remove all the aria's.
  97. StommePoes
    15:55
    unaria All The Things
  98. stevefaulkner
    15:57

    unaria All The Things

    @StommePoes been to coffee shop again

  99. StommePoes
    15:59
    make me a tshirt steve
  100. StommePoes
    15:59
    (only tourists and kids with no future go to coffeeshops)
  101. StommePoes
    16:00
    i can haz future, I'm sure of it.
  102. MichielBijl
    16:01
    Maybe we should annoy @stevefaulkner till he gives us t-shirts with “Dutch Fuckers” ;)
  103. StommePoes
    16:02
    oooh
  104. MichielBijl
    16:03
    Or maybe we should give him one with “I know some dutch fuckers and all I got was this t-shirt”
  105. MichielBijl
    16:03
    Or something actually original
  106. stevefaulkner
    16:03
    I always go to a coffee shop when in tulip land
  107. StommePoes
    16:04
    ook n mogelijkheid
  108. StommePoes
    16:04
    tourists are allowed : )
  109. LjWatson
    16:05
    @stevefaulkner wants to be an accessibility engineer when he grows up.
  110. stevefaulkner
    16:06
    @LjWatson when i retire
  111. LjWatson @LjWatson grins @stevefaulkner
  112. MichielBijl
    16:08
    Haha
  113. MichielBijl @MichielBijl wants to be @stevefaulkner when I grow up
  114. MichielBijl
    16:09
    With a touch of Dutch
  115. garcialo
    16:09
    @garcialo you think meeting me is overrated, try being me
  116. MichielBijl
    16:10
    Die hard groupies never give up!
  117. MichielBijl
    16:10
    NEVER!
  118. stevefaulkner
    16:11

    With a touch of Dutch

    read that as 'with a touch of douche' thought i have enough of that already

  119. garcialo
    16:12
    hah
  120. MichielBijl
    16:18
    @LjWatson did you even get my e-mail?
  121. MichielBijl
    16:18
    Or did I mess up the e-mailaddress or something…
  122. MichielBijl
    16:20
    @LjWatson Never mind, I messed up.
  123. MichielBijl
    16:28
    Of to DomCode :)
  124. zakim-robot
    19:16
    [tinynow] Can someone tell me whether to use role="alert" or role="alertdialog" on a modal that conveys a warning about say, the loss of network connectivity?
  125. zakim-robot
    19:19
    [marcysutton] A modal window blocking user interaction until they dismiss it?
  126. zakim-robot
    19:20
    [marcysutton] That would be a more likely candidate for role=alertdialog as that is an actual dialog–a pattern also needing focus management and a few other things. role=alert is for ARIA live regions.
  127. zakim-robot
    19:21
    [marcysutton] Here is a dialog article from Marco Zehe: https://www.marcozehe.de/2015/02/05/advanced-aria-tip-2-accessible-modal-dialogs/
  128. zakim-robot
    19:22
    [marcysutton] Here is a guide on the alert role: https://developer.mozilla.org/en-US/docs/Web/Accessibility/ARIA/ARIA_Techniques/Using_the_alert_role
  129. zakim-robot
    19:24
    [tinynow] Thank you!
  130. zakim-robot
    19:25
    [tinynow] We want sighted users to see the warning, even if they are scrolled down the page, so a modal seemed right. Is our thinking wrong?
  131. zakim-robot
    19:30
    [fstorr] @tinynow: @heydon has an example of this here http://heydonworks.com/practical_aria_examples/#offline-alert
  132. zakim-robot
    19:35
    [tinynow] That looks good, we are still worried about a user who is at the bottom of a page if the alert is appearing at the top.
  133. garcialo
    19:37
    @tinynow I think an alert dialog would work for that.
  134. garcialo
    19:38
    Less work if you use an actual alert dialog box. =)
  135. zakim-robot
    19:38
    [tinynow] Thank you all.
  136. garcialo
    19:38
    Good luck!
  137. zakim-robot
    19:54
    [lliskovoi] What would be the best way to deal with tables that have interactive header cells that rearrange content (i.e. by date, name, etc). I'm thinking adding aria-controls to the header cells, and then aria-live="polite" to <tbody>. Is there a better approach?
  138. zakim-robot
    20:06
    [lliskovoi] ah, turns out aria-sort does the trick
  139. MichielBijl
    20:41
    @tinynow: an alert requires no user interaction, where as a alertdialog has to be dismissed.
  140. zakim-robot
    21:00
    [tinynow] Thanks! Now can anyone tell me if the jQuery UI is considered accessible. It seems to meet the keyboard and screen-reader requirements (even though I don't like the markup and heavy library).
  141. zakim-robot
    21:02
    [tinynow] The jqueryUI dialog to be specific http://jqueryui.com/dialog/#modal-message
  142. dylanb
    21:07
    @LjWatson @stevefaulkner Do either of you see any issue with the following example's use of aria-controls http://oaa-accessibility.org/example/34/ ? Specifically, is it ok for aria-controls to point to something that is display:none? I think it is but I want a second opinion.
  143. MichielBijl
    21:07
    I'd love to help out, but I still need to re-read the panelset spec and bug @bkardell about it.
  144. MichielBijl
    21:08
    @dylanb how do you control something that isn't there?
  145. MichielBijl
    21:08
    Like show it maybe?
  146. MichielBijl @MichielBijl has to look at examples before he speaks
  147. dylanb
    21:08
    Well the spec states
  148. dylanb
    21:08
    Identifies the element (or elements) whose contents or presence are controlled by the current element
  149. dylanb
    21:09
    so if presence can be controlled, then being not present must be ok
  150. dylanb
    21:09
    http://www.w3.org/TR/wai-aria/states_and_properties#aria-controls
  151. MichielBijl
    21:10
    That does sound logical
  152. dylanb
    21:10
    @MichielBijl that is what I thought too
  153. stevefaulkner
    21:10
    @dylanb suggest a way to test if it is an issue is by checking in jaws ( which supports aria-controls) and see if it reports it and allows to navigate when target is display:none
  154. dylanb
    21:10
    @stevefaulkner does it need to allow to navigate?
  155. MichielBijl
    21:11
    @tinynow: that doesn't have any aria / a11y info
  156. stevefaulkner
    21:11
    What I mean is does it report the controls relationship?
  157. dylanb
    21:12
    I was hoping to get a theoretical answer (as in - this is how it should behave)
  158. dylanb
    21:13
    but I will try next time I am co-located with my JAWS machine
  159. stevefaulkner
    21:13
    Anyway just an idea - am sitting in a carpark in Surrey countryside :-)
  160. dylanb
    21:13
    hopefully a nice carpark!
  161. stevefaulkner
    21:13
    In Chelsea FC train
  162. dylanb
    21:14
    don't tell me any results
  163. stevefaulkner
    21:14
    Training centre - waiting for daughter to finish training it's night and cold
  164. stevefaulkner
    21:15
    Not my idea of fun
  165. dylanb
    21:16
    its all part of the parenting experience :-)
  166. stevefaulkner
    21:18
    @dylanb but in theory if a button controls the display of another element I guess it's ok
  167. zakim-robot
    21:19
    [tinynow] Perhaps a better question would be, what else does a modal dialog need to be accessible besides: 1)aria-labelledby pointing to descriptive title or label 2)role of dialog or alertdialog 3)assigning/toggling aria-hidden appropriately to a "page-container" and the dialog itself 4)a close button 5)escape key closes it 6)tab key moves among focusable elements inside the dialog only 7)is visible no matter where user has scrolled
  168. stevefaulkner
    21:19
    But really aria-controls is overkill (I think) unless the controller and controlled are disparate in DOM
  169. MichielBijl
    21:32

    @MichielBijl technically that is valid, but I am in the @stevefaulkner camp where I think it makes more sense when used on a control, rather than the element being controlled

    @stevefaulkner what is the reasoning behind putting it on the control? That doesn't add style hooks to the element that is actually expanded. Thus you would need a class or something else.

  170. MichielBijl
    21:33
    I see how it makes sense to have the information there, but that could be easily be obtained by browsers right? So is it an implementation thingy?
  171. MichielBijl
    21:36
    Maybe @powrsurg is willing to tell his reasoning, if he's around.
  172. powrsurg
    21:37
    Only around because of desktop notifications :)
  173. powrsurg
    21:38
    Basically, the reasoning is that AT will announce the state of the expandable when on the control
  174. MichielBijl
    21:38
    Okay, but that information is obtainable via the controls (if that is present)
  175. MichielBijl
    21:39
    aria-controls that is
  176. powrsurg
    21:39
    I guess it wouldn't make sense if the control literally said "Show/Hide", but if it said "Toggle" it would make more sense
  177. MichielBijl
    21:40
    I mean that it shouldn't matter—information wise—whether it's on the button or on the thing that is expanded/collapsed
  178. MichielBijl
    21:40
    But it does for styling reasons.
  179. MichielBijl
    21:40
    You can do [aria-expanded="true"] { display the thing }. You can't do that if it's on the control.
  180. powrsurg
    21:40
    You could
  181. MichielBijl
    21:41
    Then you have to resort to classes or something.
  182. MichielBijl
    21:41
    Only if the control is nearby.
  183. powrsurg
    21:41
    Again, it's valid to do so either way. It just seemed more useful on the controls
  184. MichielBijl
    21:42 
    <button aria-controls="thingy" aria-expanded="true">Toggle!</button> <panel id="thingy">Cats are cute</panel>

    [aria-expanded="true"] ~ panel { display the thing}
  185. powrsurg
    21:43
    aria-controls is only supported in JAWS at the moment, isn't it?
  186. MichielBijl
    21:43
    But how do you make sure that panel is actually the thing you want to expand?
  187. MichielBijl
    21:43
    Not sure about support.
  188. MichielBijl
    21:43
    There should be a “can I use” for aria…
  189. MichielBijl
    21:44
    Free project idea right there ;)
  190. MichielBijl
    21:44
    Get rich with ARIA!
  191. MichielBijl
    21:45
    Anyhow, the thing I don't get is why it is more useful on the button. What is its purpose on the button?
  192. MichielBijl
    21:46
    Besides telling you that the target is expanded.
  193. powrsurg
    21:46
    Well when it's not controlled it's useful there because it announces the state of the expandable then. If you're expanding using display none/block then it's weird to even include aria-expanded since you'll only encounter it when it's expanded (obviously there are other ways of hiding or showing, but you see where I'm going here)
  194. MichielBijl
    21:48
    Okay, but that information can still be obtained via the aria-controls right? You can get at that information regardless. So it could be announced when the button is focused, right? Or am I missing something?
  195. MichielBijl
    21:49
    My only arguments for having it on the expandable thing is that you have something to style it, and that it describes the state of the expandable thing rather than the button.
  196. MichielBijl
    21:53
    3030439.jpg
  197. MichielBijl
    21:53
    alt = the number of test mails on the aria list is too damn high (meme)
  198. powrsurg
    21:54
    Is that information actually obtained?
  199. MichielBijl
    21:54
    I don't know, my question is whether it could.
  200. powrsurg @powrsurg doesn't think he has anything that actually supports aria-controls
  201. powrsurg
    21:54
    just been doing stuff that fits examples
  202. MichielBijl
    21:55
    Only if aria-controls is defined I guess. So that would give broader support if you'd put it on the button.
  203. MichielBijl
    21:55
    I'm not necessarily for/against either; I just hate classes.
  204. powrsurg
    21:55
    From what I've seen online, aria-controls is only supported in Firefox with Jaws. By using aria-expanded on the controls then you're having that useful information announced on more AT
  205. MichielBijl
    21:56
    That is true :)
  206. powrsurg
    21:56
    At least that's the way I've interpretted stuff
  207. MichielBijl
    21:57
    My question is more of a theoretical one.
  208. powrsurg
    21:57
    http://oaa-accessibility.org/example/20/ this example follows what you're saying
  209. powrsurg
    21:58
    http://oaa-accessibility.org/example/22/ this follows the expanded on the controls example
  210. MichielBijl
    21:58
    Like you said, the spec allows both.
  211. MichielBijl
    21:58
    So maybe this is just a moot point.
  212. MichielBijl
    21:58
    I do agree that moar support is good
  213. powrsurg
    21:58
    Right. I was in here a few weeks back COMPLETELY confused about this and that's what I saw
  214. powrsurg
    22:00
    I think the spec allowing both is good, but it's also a little confusing. I think there are a lot of devs that don't think of a11y when they develop initially and think it's just a set of absolutes they need to add on at the end and the fact that it can have arbitrary success criteria catches people off guard
  215. MichielBijl
    22:01
    Catches me off guard too ;)
  216. powrsurg
    22:11
    I am still having issues with groupings and radio buttons/checkboxes. In a lot of areas where we have data tables that also have a column to do some action (enable a user or make them an admin). We've been doing simple checkboxes and and the elements have the same name element (but have aria-labels that are unique). aXe says they need to reference the same element with aria-labelledby, or be in a fieldset or group
  217. MichielBijl
    22:19
    I was thinking about that last couple of days (after the whole discussion Paul J. Adam started).
  218. MichielBijl
    22:19
    Let me put something together. I have an idea! Queue the A-team!
  219. dylanb
    22:20
    and of course aXe is THE reference ;-)
  220. dylanb
    22:22
    @powrsurg JAWS is the only AT that implements aria-controls that I know of and it allows navigation to the controlled item with INSERT+ALT+M
  221. powrsurg
    22:22
    Can you navigate to it if it's not expanded?
  222. MichielBijl
    22:23
    Why would you navigate to something that is not expanded? Surely you want to expand it first.
  223. dylanb
    22:23
    @powrsurg I have not tried that, its on my todo list
  224. powrsurg
    22:23
    and aria-controls takes a list of ids. How does it know which one to go to?
  225. dylanb
    22:23
    good question
  226. MichielBijl
    22:23
    Good point
  227. MichielBijl
    22:23
    The first one :P
  228. MichielBijl
    22:24
    Than how are developers supposed to know which one to expand?
  229. MichielBijl
    22:24
    Ah wait, that's not all controls is good for…
  230. MichielBijl
    22:24
    Silly me.
  231. powrsurg
    22:25
    why are things so hard :(
  232. MichielBijl
    22:25
    It's nature forcing us to work together
  233. MichielBijl
    22:34
    @powrsurg how about this: http://dir.agosto.nl/accessibility/table.html
  234. MichielBijl
    22:35
    For the action thing for a table row
  235. MichielBijl
    22:35
    Problem is that of course if you move to the column with the action, the column header would be spoken twice.
  236. MichielBijl
    22:36
    So, maybe just reference the row heading instead?
  237. MichielBijl
    22:37
    But then you get the name twice if you switch rows.
  238. MichielBijl
    22:37
    Damn you table mode
  239. jnurthen
    22:40
    put aria-labelledby on the input not on the label
  240. MichielBijl
    22:41
    Test updated
  241. jnurthen
    22:41
    (not that will solve the issue you are having) but what you have isn't correct
  242. powrsurg
    22:41 
    <tr>   <th scope="col">Email</th>   <th scope="col">First Name</th>   <th scope="col">Last Name</th>   <th scope="col">Enabled</th> </tr> <tr>   <td>john@example.com</td>   <td>John</td>   <td>Doe</td>   <td><input name="enabled" id="enabled_686" value="1" aria-label="Is john@example.com enabled?" checked="" type="checkbox"></td> </tr> <tr>   <td>jane@example.com</td>   <td>Jane</td>   <td>Doe</td>   <td><input name="enabled" id="enabled_686" value="1" aria-label="Is jane@example.com enabled?" checked="" type="checkbox"></td> </tr>
  243. powrsurg
    22:41
    that last one should have had a different id, but you get my point
  244. MichielBijl
    22:42
    What is the issue you're having with that?
  245. MichielBijl
    22:42
    Because that looks easier and better than my example.
  246. powrsurg
    22:43
    aXe complains that they don't reference the same element with aria-labelledby, or that the elements need to be in a containing fieldset or ARIA group
  247. jnurthen
    22:43
    @MichielBijl I think your example is fine. That is just how table mode & forms mode work
  248. jnurthen
    22:44
    I wish screen readers were smart enough so that they would label form fields in a table with the column and row headers automatically
  249. powrsurg
    22:44
    Or maybe I just found another issue with aXe false positives
  250. MichielBijl
    22:44
    @jnurthen that would be sweet, yes
  251. MichielBijl
    22:45
    The double announcing is what bothers me.
  252. jnurthen
    22:45
    @MichielBijl too much code out there though to do something like that now
  253. MichielBijl
    22:46
    Yeah. Did you see @edds talk on tables and a11y?
  254. jnurthen
    22:46
    I'm not a screen reader user of course but I often get told not to worry too much about the odd thing being spoken twice. At the speed most screen reader users use it is barely noticeable (expecially if it is just when doing table navigation for example)
  255. MichielBijl
    22:47
    Good to know :)
  256. jnurthen
    22:48
    @MichielBijl don't take my word for it though.... ask screen reader users!
  257. MichielBijl
    22:48
    I hope I get to do more user testing in the future, because I feel I miss out on a lot of experience with users.
  258. jnurthen
    22:48
    who is edds? Do you have a link?
  259. MichielBijl
    22:48
    Sure, hang on.
  260. MichielBijl
    22:49
    https://vimeo.com/139062429
  261. MichielBijl
    22:49
    @jnurthen yeah that is what I mean, I don't know a lot of people that do use them. And don't what to ask Léonie or Marco every damn time.
  262. MichielBijl
    22:50
    I could post to the list I guess.
  263. jnurthen
    22:50
    You know Bryan and Matt too...
  264. MichielBijl
    22:50
    True
  265. jnurthen
    22:50
    and you will probably get 4 different answers if you ask all 4
  266. jnurthen
    22:50
    :)
  267. MichielBijl
    22:53
    Haha, I'll post to the list and them; just to have it public.
  268. MichielBijl
    22:53
    Because I don't think I'm the only one wondering.
  269. jnurthen
    22:56
    Honestly - from my experience from screen reader users I know this is not somoething that bothers too badly!
  270. MichielBijl
    22:56
    Not sure if Marco Zehe is a PFWG member though.
  271. jnurthen
    22:56
    don't think so
  272. MichielBijl
    22:57
    Okay, I'll definitely remember that.
  273. jnurthen
    22:57
    just checked. marco is
  274. jnurthen
    22:57
    https://www.w3.org/2000/09/dbwg/details?group=32212
  275. jnurthen
    22:58
    (or the public version for those not in W3C - https://www.w3.org/2000/09/dbwg/details?group=32212&public=1 )
  276. MichielBijl
    23:03
    Can I use rich text on the list for url's? Never tried :/
  277. MichielBijl
    23:16
    I've posted it here: https://lists.w3.org/Archives/Public/public-aria/2015Nov/0017.html
  278. MichielBijl
    23:17
    For anyone interested.
  279. MichielBijl
    23:22
    Thanks @jnurthen!