Archive index

A11y Slackers Gitter Channel Archive 4th of January 2016

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    05:17
    [jitendra] I created this slack group long time ago even before this group I guess. https://accessibility.slack.com/
  2. zakim-robot
    05:17
    [jitendra] I'm going to delete it
  3. MichielBijl
    06:04
    Good old times.
  4. stevefaulkner
    12:26
    @dylanb when is aXe coming back to Firefox? "aXe DevTools could not be verified for use in Firefox and has been disabled."
  5. zakim-robot
    12:29
    [dylanb] @stevef: what version of FF?
  6. stevefaulkner
    12:30
    @dylanb 43
  7. stevefaulkner
    12:31
    @dylanb oh, and top of the year to you!
  8. zakim-robot
    12:31
    [dylanb] top of the year to you too
  9. zakim-robot
    12:31
    [dylanb] it is supposed to be cleared up to version 43
  10. zakim-robot
    12:31
    [dylanb] https://addons.mozilla.org/en-US/firefox/addon/axe-devtools/
  11. zakim-robot
    12:31
    [dylanb] can you try downloading again?
  12. stevefaulkner
    12:32
    sure will do
  13. zakim-robot
    12:37
    [dylanb] I just did a clean install of 43 and installed it and it works fine
  14. stevefaulkner
    12:37
    @dylanb - yep works now :-)
  15. dylanb
    12:37
    that is a little worrying that it disabled it automatically :-(
  16. dylanb
    12:38
    FF is becoming quite militant about extensions
  17. dylanb @dylanb just discovered that the up arrow will edit the last message entered - nice!
  18. StommePoes
    12:40
    I'm trying to find if there's any reason I shouldn't do <hr role="separator"/> within a menu, as opposed to another element with that role.
  19. StommePoes
    12:41
    This page http://www.w3.org/TR/wai-aria/roles#separator links to HTML4.01... I wonder if it was meant to do that, or point to HTML5+ references?
  20. StommePoes
    12:45
    ...and, since I can usually not find the aria page that lists everyone's native roles in a handy table, I often look up elements in the usual W3C pages, like http://www.w3.org/TR/html-markup/h1.html#h1 and wondering... I think it would be useful for this page to state the native role(s) of that element, if any. I know a heading has a role of heading, but I don't know if an hr has a role of "separator" (likely not since the meaning's been changed to "thematic change" or something... not the same as a visual and content separator in a context menu.
  21. StommePoes
    13:15
    btw by menu I mean the role=menu menu-menus, not nav menus.
  22. dylanb
    13:20
    @StommePoes I cannot think of one
  23. stevefaulkner
    13:24
    @StommePoes http://www.w3.org/TR/html-aria/#docconformance has all elements and their default roles (if they have one)
  24. MichielBijl
    14:14
    @StommePoes only reason I can think of is that you shouldn't overwrite default roles.
  25. MichielBijl
    14:15
    Seems hr already maps to separator
  26. MichielBijl
    14:15
    So no need to add it :P
  27. MichielBijl
    14:19
    @StommePoes w3c/aria#129
  28. powrsurg
    14:30
    I saw @stevefaulkner's tweet today about the blog post on inline errors put out on sunday ... do you guys never take time off? :p
  29. MichielBijl
    14:30
    @powrsurg what is time off?
  30. MichielBijl
    14:31
    Is that when the internet is out?
  31. powrsurg
    14:31
    heh
  32. garcialo
    14:31
    @MichielBijl I think that's what happens when you eat.
  33. MichielBijl @MichielBijl eats behind his laptop
  34. garcialo
    14:31
    @MichielBijl Same, but at my PC
  35. powrsurg
    14:32
    I was monitoring a database server that kept on going down for no reason I could deduce ... I blame the stupid VPS host and something on the parent ... but still, being always on doing stuff = you guys are awesome
  36. MichielBijl
    14:33
    :sparkling_heart:
  37. stevefaulkner
    14:34
    @powrsurg the article was written in a leisurely manner
  38. MichielBijl
    14:34
    Anyone doing anything positive for accessibility is awesome =D
  39. powrsurg
    14:40
    That article got me thinking. What should be done, besides displaying an error messages inline, when a form submission fails?
  40. stevefaulkner
    14:46
    @powrsurg nice example by @hanshillen http://files.paciellogroup.com/training/commonsamples/formvalidator_demo/html/FormValidatorDemo2.html
  41. LjWatson
    14:46
    Hello and happy NY A11y Slackers :)
  42. garcialo
    14:46
    Howdy @LjWatson Happy New Year!
  43. MichielBijl @MichielBijl waves at @LjWatson
  44. LjWatson
    14:47
    Is there anyone around with access to a speech recognition tool? Need a 30 second sanity check on something if possible.
  45. LjWatson @LjWatson waves at @garcialo & @MichielBijl
  46. MichielBijl
    15:04
    You know code is bad when it makes you nauseous.
  47. jonathantneal
    16:00
    I have put together 10 rules for accessible tabs at https://gist.github.com/jonathantneal/435478e415bfe4f54d06 — Would you let me know if I have forgotten anything useful, or if I have included anything unnecessary?
  48. MichielBijl
    16:05
    Why do you use a list for the tabs?
  49. MichielBijl
    16:05
    You overwrite the roles anyway. So same as div with roles?
  50. MichielBijl
    16:06
    I haven't read up on the articles, but that seems weird.
  51. LjWatson
    16:06
    @jonathantneal looks good. Might be worth including a note to explain that display:none; should be used to re-enforce the hidden attrib in brosers that do not support hidden?
  52. LjWatson
    16:07
    I also tend to leave attribs like aria-selected and aria-hidden in situ, and simply toggle the values - rather than add/subtract the entire attribute. Think the efficiency gain is negligible, so probably more personal preference than anything :)
  53. LjWatson
    16:08
    @MichielBijl progressive enhancement. The list semantics remain true for people using tech w/out ARIA support.
  54. LjWatson
    16:09
    So with ARIA support you get a collection of tabs in a tablist, otherwise you get a collection of list items in a list. In both cases the AT will tell you what it is, and how many items there are in the collection.
  55. MichielBijl
    16:09
    @LjWatson that makes sense. I don't know when role support kicked in.
  56. LjWatson
    16:10
    @MichielBijl it depends on the browser/AT and the role in question.
  57. jonathantneal
    16:10
    @MichielBijl, you make a really good point, and I would not be 100% sure how to markup the tab list if I did not need to realistically worry about a lack of role support in a much, much older AT (after all, I’m using <section> which would completely fail in much older browsers like Firefox 2).
  58. jonathantneal
    16:13
    Similarly, thank you @LjWatson for the note about [hidden]. Internet Explorer 10 and below require explicit CSS to support this.
  59. MichielBijl
    16:14
    [hidden]{ display: none; }
  60. MichielBijl
    16:15
    Copy + paste ready ;)
  61. jonathantneal
    16:18
    Also, while role may override the implied semantics of the element, there is something to be said for source readability.
  62. jonathantneal
    16:20
    Even if role is well supported by ATs, I might still expect the elements of a tablist to reflect something list-like. Perhaps <nav> with <a> children? Or would that strike you as odd, @MichielBijl, @LjWatson?
  63. LjWatson
    16:20
    @jonathantneal that's an interesting point. It would make things much harder to read if most elements were <div> or <span> or somesuch.
  64. LjWatson
    16:21
    Hmm. Using <nav> as a starting point for a tab set doesn't seem quite right to me. A set of tabs is a collection, which is more in-line with a list.
  65. MichielBijl
    16:21
    Just use panelset ;)
  66. LjWatson
    16:22
    @MichielBijl <panelset> would be nice, but we've got a way to go before it's anything other than a Web Component prototype :(
  67. MichielBijl
    16:22
    In my opinion, a list as a basis is better if you need to support older stuff.
  68. MichielBijl
    16:23
    @LjWatson yes, would make this markup a lot easier.
  69. jonathantneal
    16:23
    So, in the meantime, for the sake of source readability, <ul> with <li> and <a> is the best option, even though all of the list item semantics are overridden?
  70. LjWatson
    16:24
    @jonathantneal yes. The semantics will not be overridden by all browsers/ATs, so doing it this way is just basic progressive enhancement.
  71. jonathantneal
    16:31
    @LjWatson, @MichielBijl, thank you so much for your feedback. I have added notes to reflect your wisdom. I greatly appreciate it. https://gist.github.com/jonathantneal/435478e415bfe4f54d06#notes
  72. MichielBijl
    16:55
    :+1:
  73. zakim-robot
    17:01
    [bmeunier] In a breadcrumb or horizontal navigation bar, is it still important to use a pipe character "|" (or centered dot, or similar visual) to separate each category not displayed by CSS?
  74. garcialo
    17:03
    I don't think so, they're individual links.
  75. garcialo
    17:03
    I mean, it would be nice, but I don't think it would be necessary...what are your concerns?
  76. MichielBijl
    17:09
    I don't understand the question. As far as I know it was never good to use text for visual separation.
  77. zakim-robot
    17:11
    [bmeunier] See http://www.standards-schmandards.com/2004/title-text-separators/ ...
  78. zakim-robot
    17:11
    [bmeunier] «The separator character should be distinct so that users understand that it is a separator. (I.e. it should not appear as part of text items in the title).»
  79. zakim-robot
    17:12
    [gregtarnoff] IMHO @bmeunier separators should be hidden form AT but a label of “breadcrumbs” or similar descriptor should be available.
  80. MichielBijl
    17:13 
    <nav>   <ul>     <li><a href="/home">Home</a></li>     <li><a href="/aboot">Aboot</a></li>   </ul> </nav>
  81. zakim-robot
    17:13
    [gregtarnoff] The separators are there to give context, but “XXXX Slash YYYY” lacks context. “Breadcrumbs list, XXXX, YYYYY, etc” has more relevant context.
  82. MichielBijl
    17:13
    That seems perfectly fine does it not?
  83. MichielBijl
    17:14
    I agree with Greg.
  84. garcialo
    17:14
    @MichielBijl Should add aria-label="breadcrumb" to the nav
  85. garcialo
    17:14
    yeah
  86. zakim-robot
    17:15
    [gregtarnoff] Thats how I build them @michielBijl, I then add separators via CSS content, but probably should change that.
  87. MichielBijl
    17:15
    @garcialo that was meant as a navigation, still could use aria-label="main" or something.
  88. MichielBijl
    17:15
    If that was in fact the main navigation…
  89. MichielBijl
    17:16
    @gregtarnoff that's what CSS generated content is for; decorations.
  90. garcialo
    17:16
    ah, yes...I just figured we were talking about breadcrumbs =p
  91. MichielBijl
    17:16
    Your point is valid either way =P
  92. zakim-robot
    17:16
    [gregtarnoff] @garcialo: I think adding aria-label=breadcrumbs makes sense.
  93. garcialo
    17:16
    well, the argument is really whether the separator is important enough content to not be just decorative
  94. zakim-robot
    17:17
    [gregtarnoff] I don’t think it is.
  95. garcialo
    17:17
    I agree; I don't think it is
  96. zakim-robot
    17:17
    [gregtarnoff] Breadcrumbs are by definition a list of links in order of how to get back up the tree.
  97. garcialo
    17:17
    I think it's nice to have...like having a div with a border around landmarkish areas of the page
  98. zakim-robot
    17:18
    [gregtarnoff] If you indicate they are breadcrumbs, then the separator is decoration.
  99. zakim-robot
    17:19
    [bmeunier] If they are navigational items, I should go back to my perfect accessible horizontal menu bar. (Where did I put this pattern again?)
  100. MichielBijl
    17:20
    There is no APG item on bread crumbs =(
  101. MichielBijl
    17:21
    This is the closest thing: http://www.w3.org/TR/wai-aria-practices-1.1/#Site_Navigator_General
  102. MichielBijl
    17:21
    Would it be helpful if there was an item on bread crumbs?
  103. zakim-robot
    17:22
    [bmeunier] Yes, it would.
  104. MichielBijl
    17:32
    @bmeunier: w3c/aria#130
  105. MichielBijl
    19:28
    @bmeunier suggested it during the APG call, it has support, so I'll work on it =)
  106. zakim-robot
    19:29
    [bmeunier] You’re my hero.
  107. MichielBijl
    19:32
    Don't think I can live up to that status ;)
  108. garcialo
    19:32
    You'll have to get a cape
  109. MichielBijl
    19:32
    Ha
  110. MichielBijl @MichielBijl grabs his towel
  111. jnurthen
    19:55
    @jonathantneal You don't mention the keyboard interaction on your tabs page. If you add tab roles it is vital to change the keyboard interaction too
  112. jonathantneal
    19:57
    @jnurthen, if the keyboard interaction is required and does not happen automatically, I will add this information. Where can you point me to the requirement / expectation?
  113. jnurthen
    19:58
    look at the APG
  114. MichielBijl
    20:00
    @jonathantneal http://www.w3.org/TR/wai-aria-practices-1.1/#tabpanel
  115. jnurthen
    20:01
    using the tab roles without the keyboard interactions will make the page worse for desktop users using AT
  116. powrsurg
    22:47
    from what I recall, most browsers don't support most of the keyboard commands that involve ctrl + [OTHER_KEY] unless that is the only browser tab open
  117. jnurthen
    22:50
    I agree that some of those are not necessary. Many will be dropped and relegated to "optional" keystrokes once this gets fully reviewed. However, the general principle that the set of tabs is a single tab stop, and the arrow key moves from tab to tab will be maintained
  118. jnurthen
    22:50
    this is the same for all aria widgets
  119. MichielBijl
    22:50
    Unfortunately there are a lot of issues with modifier keys =(
  120. MichielBijl
    22:51
    ctrl+up/down arrow have special functions in OS X for example.
  121. MichielBijl
    23:33
    Anyone here have any thoughts on this w3c/aria#116?
  122. garcialo
    23:34
    Everyone knows img
  123. garcialo
    23:36
    I don't think there are any problems created or any confusion from not having role="image"
  124. jnurthen
    23:41
    image would have been better but we would end up having to use img for the next few years even if we replaced it now anyway
  125. jnurthen
    23:41
    same issue with role=presentation and role=none
  126. jnurthen
    23:42
    that one i supported more though due to none being a superior description. img and image are the same just variants and changing it now is not worthwhile
  127. MichielBijl
    23:43
    I agree with role=none
  128. jnurthen
    23:44
    we will still have to use role=presentation until IE goes away though (which could be a LONG time)
  129. MichielBijl
    23:45
    role=image just seems more logical if you look at all the other roles. I get wat @garcialo says, everyone knows img, but still. We don't have a role=fig etc.
  130. jnurthen
    23:45
    no - i agree with you completely. it is widely regarded as a mistake
  131. MichielBijl
    23:45
    Then let's fix it!
  132. jnurthen
    23:46
    just that i'm not sure fixing the mistake now actually helps anyone
  133. jnurthen
    23:46
    its just going to lead to more confusion
  134. MichielBijl
    23:46
    It would have helped me the first time I typed role=image =P
  135. jnurthen
    23:46
    yeah - but now you have learned ;)
  136. MichielBijl
    23:46
    But as Steve pointed out in the comments, only 3 pages out of 160k used role=image.
  137. jnurthen
    23:47
    there you go (although i don't ever trust public web crawlable data)
  138. jnurthen
    23:48
    for instance all of our apps are missing from that data and we produce a tonne of ARIA code
  139. MichielBijl
    23:49
    I'm not sure private webpages are that much better.
  140. MichielBijl
    23:49
    The things I've seen…
  141. MichielBijl @MichielBijl shudders
  142. jnurthen
    23:53
    yeah - but there tends to be more aria behind logins than in front IMO so i think looking at publicly indexable pages always undercounts aria
  143. MichielBijl
    23:54
    There is more aria on my laptop than on most websites.
  144. MichielBijl
    23:54
    Although that might be unfair as I have the ARIA repo here =P
  145. MichielBijl
    23:59
    Hmm, @jnurthen, I just created some tags (ARIA 1.1 and ARIA 2.0) on GitHub. A lot of issues have this in their title. Do you think we should remove it from the title and add the labels?
  146. MichielBijl
    23:59
    Unsure what is better. I like labels, but I know what I like isn't for everyone.
  147. MichielBijl
    23:59
    Trying to clean up the issues a bit.