Archive index

A11y Slackers Gitter Channel Archive 19th of November 2015

What fresh hell is THIS now? - Patrick Lauke
  1. MichielBijl
    06:09
    aria-labelledby on the content/main would allow for a visible heading ( if someone wants to use that). The labelledby would simply function as a fallback.
  2. MichielBijl
    06:10
    @garcialo @cordelia
  3. powrsurg
    14:46
    Does any AT actually do something different when it encounters schema markup?
  4. powrsurg
    14:52
    I mean, if we're marking up content in specific formats so that things like an event or a product or an offer might be easily understood for search engines maybe it would be useful for AT to do the same?
  5. powrsurg
    14:52
    schema, open graph, microdata .... all of those types of markup
  6. StommePoes
    14:57
    I've never heard any difference myself while browsing e-commerce websites which I know are using Schema.org.
  7. StommePoes
    14:58
    I dunno that it would be of any use while browsing, but I could see it being a useful thing if browsers themselves could recognise Schema and allow users to search via those attributes. Or ask for a sort of "quick list" of only marked-up content and the hierarchy.
  8. MichielBijl
    15:25
    Never noticed either.
  9. StommePoes
    15:38
    Playing with fill-in-the-blank & screen readers. It seems all sorts of cool crap can be done. But should it? :P
  10. dylanb
    16:16
    @stevefaulkner I am looking for where it is specified how fieldset maps to the acc name/desc. calculation (or some alternative treatment) and am not finding it
  11. dylanb
    16:16
    help?
  12. garcialo
    16:30
    Just for my own understanding; you're not asking how fieldset contributes to an acc name or description, right? I don't quite get all that lower level stuff yet.
  13. dylanb
    16:37
    I am asking how fieldset+legend contributes to acc name/desc calculation
  14. dylanb
    16:37
    I am also looking in general for where fieldset and legend accessibility API mappings are defined
  15. garcialo
    16:40
    I'll have to look up what acc description actually is. Should the fieldset contribute to the acc name? I always thought of fieldsets as metadata for the field rather than being "part" of the field.
  16. zakim-robot
    16:46
    [dylanb] the fieldset should not, the legend should show up somewhere
  17. zakim-robot
    16:46
    [dylanb] I would expect it to appear as part of the acc name
  18. zakim-robot
    16:46
    [dylanb] because that is what it is semantically
  19. garcialo
    16:47
    Yes, of course. I should have said the legend, but yeah...I agree it should be somewhere. I can see the argument for it being in acc name.
  20. StommePoes
    16:48
    Certainly browsers and AT don't agree when they should announce legends, if there's a standard.
  21. dylanb
    16:49
    to me, it is equivalent to an implied additional aria-labelledby but defining it that way would complicate the algorithm as currently defined IMO
  22. dylanb
    16:50
    @StommePoes that is one problem and I am investigating some reports that CSS positioning of the legend can break this in some cases
  23. StommePoes
    16:50
    IE+JAWS always announces legend all the time no matter what (super annoying). FF+JAWS announces the legend once per getting-to-input (no matter how); NVDA only wants to announce if user gets to fieldset group via focus mode+tab, but not via quick nav.
  24. StommePoes
    16:50
    Do you have which cases/browser/at?
  25. StommePoes
    16:50
    NVDA has filed a bug about their issues, they're not sure what they should do
  26. dylanb
    16:51
    I have a generic report that absolute positioning of legend breaks announcement
  27. dylanb
    16:51
    that is as far as I have got
  28. garcialo
    16:51
    I think I actually prefer the IE+JAWS, but I can see it getting redundant
  29. dylanb
    16:51
    I prefer NVDA
  30. StommePoes
    16:51
    Uh, I prefer FF+JAWS-- I do want to hear the legend at least once since it's often a question when it comes to radios/checks, but I don't want to hear it before every. single. field.
  31. dylanb
    16:52
    except NVDA has some bugs in that it announces the group on links too
  32. dylanb
    16:52
    so when you tab into a fieldset and the first thing you encounter is some peripheral link, it announces the group (and you tend to ignore it because of the context)
  33. garcialo
    16:52
    I imagine I prefer it since my main AT use is in doing assessments. :smile:
  34. StommePoes
    16:52
    Also using "f" to jump to a field in FF can leave the legend off, also troublesome
  35. dylanb
    16:53
    @StommePoes I think that might be related to the issue with the links
  36. StommePoes
    17:00
    NVDA announces a position:absolute legend if you tab into the group (are in forms mode)
  37. StommePoes
    17:00
    in IE
  38. dylanb
    17:03
    I would expect NVDA to work well - it mostly does
  39. dylanb
    17:03
    I am trying to get more details on the AT+browser combo
  40. stevefaulkner
    17:03
    @dylanb http://rawgit.com/w3c/aria/master/html-aam/html-aam.html#fieldset-and-legend-elements
  41. StommePoes
    17:04
    NVDA+FF also reads abso-po legends in FF when group is tabbed into. So, behaviour in NVDA and 2 browsers is not changed with absolute positioning of the legend (though I have merely positioned it. It's still visible too)
  42. stevefaulkner
    17:05
    @dylanb more fieldset - http://rawgit.com/w3c/aria/master/html-aam/html-aam.html#el-fieldset
  43. dylanb
    17:06
    @stevefaulkner thank you!!
  44. StommePoes
    17:07
    IE+JAWS always announcing pos-abso legend regardless of navigation method.
  45. dylanb
    17:08
    which JAWS?
  46. dylanb
    17:09
    @stevefaulkner if I am interpreting this correctly, the fieldset is an element with a role of group
  47. dylanb
    17:10
    the legend is one way of getting the accessible name for that group
  48. dylanb
    17:10
    how that gets announced by the AT is left up to the AT
  49. dylanb
    17:11
    (in terms of when you jump into it, not look at it directly)
  50. StommePoes
    17:12
    JAWS 16
  51. StommePoes
    17:12
    In Firefox, again I can't get legends read out using "f" but abso-po makes no difference.
  52. stevefaulkner
    17:12
    @dylanb yes sounds right
  53. stevefaulkner
    17:13
    how that gets announced by the AT is left up to the AT
  54. stevefaulkner
    17:13
    is how it is for pretty much all aural UI
  55. dylanb
    17:14
    unfortunately
  56. dylanb
    17:14
    or maybe not
  57. dylanb
    17:15
    I think some of it should be specified but some of the differences are ok
  58. stevefaulkner
    17:16
    I would ove to be able to require AT to announce stuff, part of reason I am doing http://thepaciellogroup.github.io/AT-browser-tests/ (too slowly) so can point AT implementers to gaps/interop issues
  59. garcialo
    17:16
    Ah, there it is! I was looking for that very thing the other day and failed miserably. =p
  60. dylanb
    17:16
    nice
  61. stevefaulkner
    17:17
    needs work (PR's welcome)
  62. stevefaulkner
    17:19
    failed miserably. - @garcialo
  63. garcialo
    17:19
    Don't worry @stevefaulkner, once I finish up the 3 or 4 things I'm working on, I'll definitely really want to and hopefully will help you out.
  64. stevefaulkner
    17:20
    :+1: :+1:
  65. StommePoes
    17:25
    Finally, JAWS makes no distinction about legends being abso-po'd in FF either, but I noticed no matter what, if each legend is the same text (different fieldsets, even different forms), they don't get announced again per fieldset. I needed to make text differences to get them all read out. This is again only with tab+forms mode, using "f" navigation never announces legend.
  66. StommePoes
    17:31
    @dylanb so it's possible something else is hurting the abso-pod' legends you heard about, but likely not the abso-po-ing itself.
  67. dylanb
    17:32
    @StommePoes which version of JAWS are you using?
  68. StommePoes
    17:33
    16
  69. StommePoes
    17:33
    @stevefaulkner you have near the top first table of your JAWS test "•Next Visited Link V" did you mean "Nest Visited V"? (instead of link v?)
  70. stevefaulkner
    17:37
    @stommepoes yes that's a code error
  71. StommePoes
    17:40
    hm I see more code errors, I'll see if I have time to do anything with it
  72. stevefaulkner
    17:41
    @StommePoes yeah, yeah heard that before... ;-)
  73. StommePoes
    17:47
    haha
  74. StommePoes
    17:48
    you saw I did manage real work for chaals, things can happen
  75. StommePoes
    17:49
    Tho apparently I might get beer out of that
  76. StommePoes
    17:49
    Would I get any beer for this?
  77. StommePoes @StommePoes thinks, does Will Test For Beer work?
  78. powrsurg
    17:58
    Really stupid question: when you say acc name is "acc" just short for accessibility or is there an acronym I don't know? It seems like you typically mean the former but wondered if the latter was the case
  79. zakim-robot
    18:39
    [cameron] where can I find a reference of broken accessibility, something I can write unit tests against?
  80. garcialo
    18:40
    There is the Before and After Demo site http://www.w3.org/WAI/demos/bad/
  81. garcialo
    18:41
    I'm also working on some discrete accessibility issues at http://a11ynot.com/ in preparation for my CSUN16 talk
  82. garcialo
    18:41
    The latter is VERY work in progress at the moment.
  83. zakim-robot
    18:42
    [cameron] @garcialo: that’s great! Thanks for the tips.
  84. zakim-robot
    18:42
    [cameron] I also was looking at axe-core unit tests: https://github.com/dequelabs/axe-core/blob/master/test/integration/full/skip-link/skip-link-pass.html
  85. zakim-robot
    18:43
    [cameron] for example
  86. garcialo
    18:45
    Ah, I hadn't seen those yet. I'll probably eventually put those on the site. =)
  87. garcialo
    18:45
    Well, the ones that create issues.
  88. zakim-robot
    19:18
    [gregtarnoff] @stefsull: can I DM you?
  89. zakim-robot
    19:51
    [bmeunier] Does a pagination of a search result should to be in it’s own <section> element?
  90. zakim-robot
    19:52
    [gregtarnoff] I can’t think of anything requiring it to be. It seems more like it should be a nav element in the section with the results though to me.
  91. zakim-robot
    19:53
    [bmeunier] <nav>, right. And so it should be title <h1>Pagination</h2> and we could add a screen read off css rule. Right?
  92. zakim-robot
    19:54
    [cmooney] Most popular automated testing tools for accessibility? Not a per page test, but something that works site wide, and for bonus points can do DOM checking as well
  93. zakim-robot
    19:55
    [gregtarnoff] @cmooney: have you looked at Tenn.io?
  94. zakim-robot
    19:56
    [cmooney] that can test a whole site, not a page at a time?
  95. zakim-robot
    19:57
    [gregtarnoff] @bmeunier: depends on your structure - <section><h1>Section title</h1>{{ results }}<nav><h2>Pagination</h2><nav></nav></section> is how I’d approach it. Ideally I never use offscreen text for headings.
  96. zakim-robot
    19:58
    [bmeunier] What would you use @gregtarnoff ?
  97. zakim-robot
    19:58
    [gregtarnoff] @cmooney: I typo’d tenon.io. I think so, it integrates to your build process using grunt or gulp.
  98. zakim-robot
    20:00
    [gregtarnoff] The nav itself doesn’t really need a title, but one certainly helps. 9/10 time pushing something off screen to help screen reader users is actually reducing usability for other people with a11y needs and even those without.
  99. zakim-robot
    20:02
    [cmooney] I am talking about evaluating an already existing site, a fairly large one
  100. zakim-robot
    20:03
    [cmooney] Cart before the horse I know and we are in a redevelopment phase but I’d like to get a site wise snapshot of how bad things are accessibilty wise for what we’ve already got out there
  101. zakim-robot
    20:05
    [gregtarnoff] http://pa11y.org/, SSBart Group’s AMP are a couple. I haven’t used them though. PLans are that AMP might get used fairly soon in a project though.
  102. garcialo
    20:06
    There's also Deque's WorldSpace and OpenAjax's Functional Accessibility Evaluator http://fae20.cita.illinois.edu/
  103. garcialo
    20:08
    Tenon / WAVE API / aXe are other alternatives, but I believe they're all effectively just API options, so you'd need to send to them and build the results into a report.
  104. garcialo
    20:09
    pa11y.org is still a bit wonky; I was able to get it set up, but couldn't get it to scan properly; have never used AMP
  105. garcialo
    20:09
    I like FAE, but their rules are a bit aggressive for my test; to get an idea, you can check out AInspector for Firefox
  106. garcialo
    20:10
    I work with WorldSpace, so I'm biased
  107. garcialo
    20:10
    for my taste*
  108. stevefaulkner
    20:31
    @powrsurg acc name is short for accessible name accname is also the property name used in some APIs
  109. powrsurg
    20:47
    @stevefaulkner that's what I thought. I do see now that I just said "accessibility" when I meant accessibility name .... either way, good to know the correct thing
  110. powrsurg
    20:47
    now how to pronounce a11y!
  111. zakim-robot
    20:52
    [gregtarnoff] @powrsurg I always say "alley".
  112. StommePoes
    20:59
    aye one one why
  113. zakim-robot
    21:29
    [cameron] ehh eleven why
  114. zakim-robot
    21:31
    [cameron] :fonzie:
  115. zakim-robot
    21:32
    [cameron] @jdan: what did you land on? eleventy?
  116. zakim-robot
    21:32
    [jdan] @cameron I’ve been saying eleven-y :simple_smile:
  117. zakim-robot
    21:32
    [cameron] :smile:
  118. zakim-robot
    21:32
    [cameron] cool
  119. powrsurg
    21:56
    how about ali (like muhammad ali)? :p
  120. garcialo
    22:17
    ay-le-ven-dy