Archive index

A11y Slackers
Gitter Channel Archive 24th of July 2015

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    Jul 24 02:12
    [Hans Hillen, a11y] @stevenlambert: I would think that being consistent with the the current OS is most important if we’re talking about an app. If it’s web content that’s more tricky of course, and you wouldn’t want to be inconsistent by changing the order depending on which OS the browser runs in. So pick one and stick to them. Generally the rule is that the primary button can be activated through the enter key from anywhere in the dialog (if that’s what we’re talking about), so not having the primary button first would not be a huge problem. From what I’ve seen people do just explore all controls when they go through a learning phase with a UI, and once they know where to find what they need they’ll use as many ways to skip to those controls as they can.
  2. ewaccess
    Jul 24 02:19
    @stevefaulkner thanks for the hat tip. @StommePoes @deborahgu Windows Speech Recognition (WSR) is great for voice control but bad for dictation. The Show Numbers command is super useful: video
  3. hanshillen
    Jul 24 02:26
    @Scratch2k: I wouldn't consider it fully equivalent. If it was, you would find it acceptable for everybody to use direct text entry as only method. But you chose to use a date picker instead. Why? Because date pickers provide a lot of great features that significantly improve UX and make it easier to come up with dates: navigating by week, month or year, not having to figure out how many days there are in a particular month, quickly navigating backwards / forwards a specific number of days/weeks without having to perform calculations in your head, knowing which day of the week a particular day is, etc. This is all functionality that a date picker provides, and basic text entry pales in comparison. So I would still recommend finding a keyboard accessible version.
  4. Scratch2k
    Jul 24 02:28
    good points, I hadn't considered all that
  5. stevefaulkner
    Jul 24 03:13
    nice to see your face here @ewaccess :-)
  6. sgtphips
    Jul 24 07:03
    Good Evening Everyone. I am really interested in this group!
  7. stevefaulkner
    Jul 24 07:05
    @sgtphips hi, what's your interest?
  8. LjWatson @LjWatson waves @stevefaulkner
  9. stevefaulkner
    Jul 24 07:09
    hey @LjWatson xx
  10. StommePoes @StommePoes waves at room
  11. stevefaulkner @stevefaulkner waves at @StommePoes
  12. StommePoes
    Jul 24 07:29
    Thanks @ewaccess for the video, and yeah the idea isn't so much dictation but navigation/interaction.
  13. LjWatson
    Jul 24 07:29
    Hello @StommePoes and @sgtphips
  14. StommePoes
    Jul 24 07:29
    Hi Leonie, Steve
  15. StommePoes
    08:42
    A small change from "knowing web accessibility is a pre" right under the other "pre's" like SEO and familiarity with Some Obscure CMS
  16. schalkneethling
    08:42
    I like all the attention a11y is getting
  17. schalkneethling
    08:43
    It has taken a long time to get here though
  18. schalkneethling
    08:43
    too long really
  19. MichielBijl
    08:43
    Needs more time though
  20. zakim-robot
    08:44
    [jitendra, a11y] Why “BOT” is showing in front of you people?
  21. MichielBijl
    08:44
    Because we are in the Gitter room
  22. StommePoes
    08:44
    Jitendra you also have bot in front of you
  23. StommePoes
    08:44
    (09:44:12 AM) zakim-robot: [jitendra, a11y] Why “BOT” is showing in front of you people?
  24. schalkneethling
    08:44
    the more people get involved and the more attention is focus on it, the faster tools will evolve to make it easier to add a11y into your development workflow
  25. StommePoes
    08:44
    Do I have bot? I should be just Stomme poes, I'm connected via IRC
  26. schalkneethling
    08:44
    and this is a big win
  27. zakim-robot
    08:45
    [jitendra, a11y] I’m using slack osx app
  28. LjWatson
    08:45
    The two bots manage a pipe between the Gitter and Slack channels, so we can all talk together whichever platform we prefer to use.
  29. MichielBijl
    08:45
  30. StommePoes
    08:45
    If companies actually take it seriously, then maybe. Right now they don't, and changing only at the HR level won't help-- they need to follow up on that
  31. schalkneethling
    08:45
    oh definitely
  32. schalkneethling
    08:46
    it's a start though right?
  33. StommePoes
    08:46
    I interviewed for a largish-and-very-fast-growing e-commerce company. Part of my pitch to them was showing some problems in ZoomText since they ask applicants to point out things they'd change or fix on their sites
  34. StommePoes
    08:46
    later they told me "well but those kinds of people don't use our site" and "you'd probably be fighting the product manager over these things"
  35. StommePoes
    08:46
    um, as if that were a bad thing? :P
  36. schalkneethling
    08:46
    I have seen more and more talk about the fact that web devs should know how to use a screen reader and tutorials and such popping up offering to teach this
  37. schalkneethling
    08:47
    Just because there is some push back or resistance from another department does not mean we have to just leave it at that
  38. schalkneethling
    08:47
    we need to have those conversations
  39. schalkneethling
    08:48
    And I can honestly not believe people still think that "well but those kinds of people don't use our site", holds any truth
  40. StommePoes
    08:48
    Yeah but mostly, I did not blame this company as being weird or different-- they were doing the norm. So it will be nice when that becomes "abnormal" as in "that company isn't run very well"
  41. MichielBijl
    08:48
    More clients should be aware. If the client is aware, they might bring it up.
  42. schalkneethling
    08:48
    yup
  43. MichielBijl
    08:52
    schalkneethling: got reminded of that this week; I was asked to do some a11y testing on an iPad app, because they want to launch it in de US “And they have very strong a11y laws, so now we must check!”. So I wrote out my findings and reported back. The manager asked me about some point that involved SR usage, so I explained why buttons should be labels, and everything should be focusable. He literally asked “Oh, blind people need
  44. MichielBijl
    08:52
    to use it, too?”
  45. MichielBijl
    08:52
    O.o
  46. schalkneethling
    08:53
    Hahahaha, seriously! ;p
  47. StommePoes
    08:53
    well yeah, blind people don't use the internet
  48. StommePoes
    08:54
    they sit at home getting uitkering in darkened rooms waiting for pity visits from family, duh
  49. MichielBijl
    08:54
    So explained that, yes, yes they do :) And showed him how to use an iPad with the screen curtain on. He seemed somewhat interested, but apparently it takes laws for him to include accessibility.
  50. StommePoes @StommePoes boils in a bit of rage, then goes and eats a cooke
  51. MichielBijl @MichielBijl steals a cookie from StommePoes
  52. schalkneethling
    08:55
    yeah, the law thing has been a driving force for some time. The only thing really that got big companies to look at the a11y of their sites
  53. schalkneethling
    08:55
    Kinda weird that people do not just do it because it is the right thing
  54. schalkneethling
    08:55
    and it actually does affect their bottom line
  55. schalkneethling
    08:56
    they just thinnk it does not
  56. MichielBijl
    08:56
    Doing the right thing often doesn't make you money…
  57. MichielBijl
    08:56
    It does actually, but they don't know.
  58. schalkneethling
    08:56
    oh yeah, the old mighty dollar
  59. schalkneethling
    08:56
    exactly, there is a lot of educations that needs to happen of both developers as well as clients
  60. stevefaulkner
    08:58
    @StommePoes did you go for the job at duckduckgo that was advertised recently? seemed a good one for you
  61. StommePoes
    08:58
    @stevefaulkner no, they've moved on beyond my skills, they're using something like knockout and sass and all that stuff I don't give a rats about.
  62. zakim-robot
    08:58
    [jitendra, a11y] Is there any commercial website which is Wcag AAA compatible?
  63. StommePoes
    08:59
    AAA? I doubt it
  64. StommePoes
    08:59
    even lainy feingold's site I think is AA at best...
  65. StommePoes
    08:59
    AAA is pretty hard to do everywhere site-wide
  66. stevefaulkner
    08:59
    AAA is beyon the event horizon
  67. zakim-robot
    08:59
    [jitendra, a11y] No commercial website in world with AAA?
  68. StommePoes
    08:59
    you can do lots of individual stuff at AAA
  69. StommePoes
    08:59
    but whole sites... hm
  70. StommePoes
    08:59
    I can't say there isn't one, but I don't know of one.
  71. zakim-robot
    08:59
    [jitendra, a11y] they what is the point of having AAA
  72. StommePoes
    09:00
    It's a goal to try to reach, and each individual point has its own levels
  73. StommePoes
    09:00
    for example, for some reason heading content with... headings, is AAA as far as requirements (but is easy to implement on A and AA sites of course)
  74. stevefaulkner
    09:00
    @jitendravyas in specific contexts the additional criteria are useful
  75. StommePoes
    09:01
    but I thought that requirement level was because some sites might have good content reasons not to use headings. Or, that's how I read the WCAG pages
  76. StommePoes
    09:01
    cause I wanted to fault/fail a site on not using headings and it turned out for their level it wasn't required-required, was AAA, and that surprised me.
  77. StommePoes
    09:02
    I still recommended headings, but I couldn't "flunk" them on it or call it a "violation". Just "dumb" :)
  78. MichielBijl
    09:02
    Surely using headings should be A
  79. StommePoes
    09:02
    I thought so
  80. zakim-robot
    09:02
    [jitendra, a11y] So If I’m going to make a new website should I aim for A first and publish it and then AA and AAA?
  81. StommePoes
    09:02
    but when I was asked "which things on your list are actual violations" I saw they were AAA
  82. StommePoes
    09:02
    No Jitendra aim for AAA! Whatever you don't make, will be either AA or A.
  83. MichielBijl
    09:02
    Agreed
  84. StommePoes
    09:02
    aim high and whatever doesn't work or make sense ends up a bit lower
  85. StommePoes
    09:03
    I mean, some AAA things are easy, some are difficult for some kinds of content.
  86. MichielBijl
    09:03
    Colour contrast could be a pain. But you can reach that with high contrast mode :)
  87. zakim-robot
    09:03
    [jitendra, a11y] Can anybody confirm if youtube’s embeded player is WCAg compatible?
  88. StommePoes
    09:03
    2.4.10 Section Headings: Section headings are used to organize the content. (Level AAA) <-- this is where I saw AAA and was like, wut?
  89. StommePoes
    09:05
    Jitendra, when I tested a site that had embedded Youtube and embedded Vimeos, I could reach teh Youtube controls with keyboard, all but the main play button in the middle announced a name, and contrast was good (yellow borders around focussed controls). However captions in teh video will matter. YouTube's craptions suck
  90. StommePoes
    09:05
    and embeds don't provide transcripts or subtitles either of course.
  91. stevefaulkner
    09:06
    Understanding Levels of Conformance "The Success Criteria were assigned to one of the three levels of conformance..." http://www.w3.org/TR/UNDERSTANDING-WCAG20/conformance.html#uc-levels-head
  92. stevefaulkner
    09:09
    A and AA does not mean AA is less required than A
  93. MichielBijl
    09:10
    That does mention titles for sections
  94. MichielBijl
    09:10
    But it's pretty vague
  95. StommePoes
    09:11
    No but if someone says they are not aiming for AAA can I claim they are in a WCAG violation of something listed as AAA?
  96. StommePoes
    09:11
    This particular site didn't have a goal but others I've tested explicitly said "we're only going for AA level" (Im' assuming because they mean teh difficult AAA stuff they didn't think they'd be able to do)
  97. zakim-robot
    09:12
    [som, a11y] I read that while sites should consider the guidelines they should also be aware of individual user agent capabilities .. ie. if it’s in the standard but doesn’t work in JAWS or VO then you may be in trouble? Should UA capabilities be considered when aiming for AA?
  98. StommePoes
    09:13
    Well the WCAG only says what a user should be able to do
  99. StommePoes
    09:13
    but not how, the techniques are just suggestions
  100. MichielBijl
    09:14
    @som: just need to work around those issues.
  101. MichielBijl
    09:14
    Same as with other dev issues.
  102. zakim-robot
    09:14
    [som, a11y] I no longer develop for IE8
  103. zakim-robot
    09:14
    [som, a11y] I try not to develop for IE9
  104. zakim-robot
    09:15
    [som, a11y] what is reasonable?
  105. StommePoes
    09:15
    I remember when I was stuck on IE9 because ZoomText did not work on higher versions (it does now)
  106. StommePoes
    09:15
    one thing you might just want to keep in mind is, there are a lot of people stuck on older stuff due to cost
  107. StommePoes
    09:15
    on the other hand as a developer you have a limit somewhere on how much support you will offer
  108. zakim-robot
    09:16
    [som, a11y] so is AA subjective?
  109. StommePoes
    09:17
    Per thingie a user should be able to do, no. How you tackle the problem though and what your solution ends up being can be a sliding judgement. Or, people argue about them a lot
  110. StommePoes
    09:17
    they'll be like "well users can do X if they do it this way" and someone else might be like "well but do users know to do that" or "well but can all your users depend on that if they have to?"
  111. StommePoes
    09:18
    WebAIM mailing list is a good place to hear people present problems and hash out solutions
  112. zakim-robot
    09:26
    [som, a11y] it’s interesting. people do use software in very different ways. We recently had assessors nail us on a number of fronts because while our document was well structured, some buttons weren’t entirely clear when accessed out-of-context through the buttons navigator in JAWS (they did have a describedby label which wasn’t appearing). The assessors suggestion was to create visually hidden text within the button that repeated the section header. However this would be repetitive for a11y users navigating the page normally.
  113. zakim-robot
    09:27
    [som, a11y] One solution creates another problem
  114. StommePoes
    09:27
    by button you mean link?
  115. zakim-robot
    09:27
    [som, a11y] <button>
  116. zakim-robot
    09:27
    [som, a11y] not <anchor>
  117. StommePoes
    09:28
    What did the button do? I assume the visible text or whatever you had stated what it did, just not to which section?
  118. StommePoes
    09:29
    I have often tried to find succint names for sections that aren't the whole section title text, but it depends on the content what you can get away with.
  119. StommePoes
    09:31
    people can skim too-- the moment you hear you're on a button and don't care you can move on-- just getting the beginnings of stuff. For that reason and for sighted skimmers I try to get our copywriters to put teh most useful words in front. But what the element is can also often come first.
  120. StommePoes
    09:31
    When two solutions seem to contradict each other, most people with the resources to do so will say "do user testing". Though for example my company will never, ever ever do user testing, let alone on disabled users.
  121. StommePoes
    09:32
    Lots of companies don't. But if you've got assessors then possibly you can make more of a case for one level of verbosity vs another based on user testing...
  122. zakim-robot
    09:32
    [som, a11y] The button downloaded files based on a checkbox selection. The button text was something along the lines of “Download {x} packages”.
    re. shortening, my reasoning is that if the title can be shortened it should be shortened for everyone.
  123. zakim-robot
    09:33
    [som, a11y] And yes copy is a big concern, why write copy three ways?
  124. StommePoes
    09:33
    If some users only get one instance of a word/phrase, makes sense to have the long complete version. If it's repeated, makes sense to have those shorter
  125. StommePoes
    09:34
    people can also investigate if they come across something and they're not sure, and people can explore context around controls.
  126. StommePoes
    09:34
    We write copy two ways already: mobile is way less marketing bs than the "desktop"
  127. StommePoes
    09:34
    ideally there would be zero bs on any versions but that's not how our company works :/
  128. zakim-robot
    09:34
    [som, a11y] really? you change copy between desktop and mobile?
  129. StommePoes
    09:34
    and this is common
  130. StommePoes
    09:35
    Yeah, there's a lot of crap in the desktop versions
  131. StommePoes
    09:35
    at least they didn't try to stuff it all into the mobile
  132. zakim-robot
    09:35
    [som, a11y] ok. we might remove sections .. very rarely .. but most of the time it’s mobile first. verbose is bad
  133. StommePoes
    09:35
    marketing wants to call everything young and dynamic. If you're looking to buy something from a list of items, you don't care, you want the specs and the price
  134. StommePoes
    09:35
    agreed.
  135. StommePoes
    09:36
    But I'm at the bottom of the totem pole here :)
  136. MichielBijl
    09:36
    “Buy our product! It has young specs and the price is dynamic!”
  137. StommePoes
    09:36
    haha
  138. StommePoes
    09:37
    it's multifunctional!
  139. MichielBijl
    09:37
    That would never sell :p
  140. StommePoes
    09:37
    it slices, it dices, and it also does windows! comes in a clean, fresh plastic frame!
  141. StommePoes
    09:37
    mobile: <h>functions</h><ul><li>slices
  142. StommePoes
    09:37
    </li><li>dices</li></ul
  143. StommePoes
    09:38
    etc
  144. StommePoes
    09:38
    frame dimensions and material.
  145. zakim-robot
    09:38
    [som, a11y] haha .. fair enough!
  146. StommePoes
    09:38
    you could argue one seems more "friendly", the other is more "scanny"
  147. StommePoes
    09:38
    I prefer scanning but you notice sites who sell junk for the home seem to do well with more friendly bs-y text.
  148. StommePoes
    09:38
    so, I guess there's a place for it.
  149. StommePoes
    09:39
    anyway, if you're adding hidden/nonvisual text for users, teh same might apply ig those users also have access to the visible text too
  150. zakim-robot
    09:39
    [som, a11y] for sure. it’s tricky segmenting users based on a device tho
  151. MichielBijl
    09:39
  152. StommePoes
    09:39
    the invisible, if it's a repeat/clarity thing, could potentially be more scan-y and quick
  153. StommePoes
    09:40
    Michiel who knows
  154. MichielBijl
    09:40
    I don't need a website to tell me facts in a friendly way. If I want facts told in a friendly way I'll go to a comedy show.
  155. zakim-robot
    09:41
    [som, a11y] or a doctor
  156. MichielBijl
    09:41
    Hehe
  157. StommePoes
    09:41
    the kind who laughs somewhat inappropriately as he tells you you're about to die... Dr Hibbert!
  158. MichielBijl
    09:41
    NOOO, not Dr Hibbert!
  159. StommePoes
    09:42
    "ahee hee hee"
  160. StommePoes
    09:42
    "you've got cancer, son. Hee hee."
  161. zakim-robot
    09:42
    [som, a11y] dr giggles is also good
  162. zakim-robot
    09:42
    [som, a11y] on that note i’m out for the weekend. thanks for the insight
  163. MichielBijl
    09:42
    Dr. Dre keeps it real, ya digg?
  164. StommePoes
    09:42
    test, test, test!
  165. StommePoes
    09:42
    if you can
  166. zakim-robot
    09:43
    [som, a11y] (i wish i couldn’t)
  167. zakim-robot
    10:35
  168. stevefaulkner
    11:30
    @som can you provide example code for the issue with the buttons, so I can see the context?
  169. stevefaulkner
    11:35
    Controversial oldy but goody - When accessibility is not your problem http://joeclark.org/appearances/atmedia2007/ by Joe Clark (naturally)
  170. JimBobSquarePants
    12:09

    Hey guys, looking for some advice if I may.

    I run an open source responsive framework with which I have attempted to make everything accessible, I've probably missed stuff or made mistakes but I'm really trying my best.

    I am struggling with one particular bit of functionality: A responsive table that emits a vertical format on smaller viewports. I simply cannot get my screenreader (NVDA on Windows) with any browser to associate the cells with the corresponding headers when the table is collapsed. I've tried combinations of id's, aria-hidden + " aria-labelledby but that doesn't work.

    Would you have any pointers you could offer as I am stuck?

    The page with the demo is here. http://responsivebp.com/javascript/tablelist/
    Resize your browser to see it collapse.

  171. StommePoes
    12:20
    One bet I'm making is that somehow with the floats, the table things aren't display: table-thingie anymore.
  172. StommePoes
    12:21
    Not that that should matter, floating list elements means they're not display: list-item anymore and they're still list items
  173. StommePoes
    12:21
    but tables might have weird things going on due to them being used for presentation in the past
  174. StommePoes
    12:21
    and I know JAWS at least used to try to guess with heuristics if a table was really a data table or not.
  175. StommePoes
    12:23
    The moment it's not collapsed anymore, your readers read it out good?
  176. MichielBijl
    12:23
    Seems you need to provide roles (role="grid" etc), because the defaults get removed if you use display: block;
  177. StommePoes
    12:24
    guess this would be the exception to not adding roles that are normally there by default rule... but probably a good use case.
  178. MichielBijl
    12:25
    Yep, I agree.
  179. StommePoes
    12:25
    I wonder how the chris coyier responsive data tables fair... those also get changed to display: block and whatnot
  180. MichielBijl
    12:28
    That has the same issue.
  181. StommePoes
    12:28
    s/fair/fare
  182. MichielBijl
    12:29
    That guy must be fun at parties
  183. StommePoes
    12:30
    awesome avatar tho
  184. MichielBijl
    12:30
    Don't know who/what that is
  185. stevefaulkner
    12:31

    @JimBobSquarePants

    I simply cannot get my screenreader (NVDA on Windows) with any browser to associate the cells with the corresponding headers when the table is collapsed.

    I just tried it with JAWS and NVDA on latest Firefox, seems to read fine i.e. as I arrow through the header and the data cell content are announced

  186. MichielBijl
    12:31
    @stevefaulkner in the narrow view?
  187. MichielBijl
    12:31
    VO didn't recognise the table as such in the narrow view
  188. MichielBijl
    12:32
    Upon closer inspection, its role was reset.
  189. StommePoes
    12:32
    tr:before { content: attr(data-thead); } brilliant, gets rid of manual text in the css
  190. StommePoes
    12:32
    @Michiel to what?
  191. MichielBijl
    12:33
    To nothingness
  192. StommePoes
    12:33
    what did VO call it in squished view
  193. StommePoes
    12:33
    oh\
  194. stevefaulkner
    12:33
    @MichielBijl yes it wasn't announcing the table structure per se but the content was logically ordered and readable
  195. JimBobSquarePants
    12:34

    @MichielBijl @StommePoes Thanks, I gave role=grid a try. Worked well with IE but it caused Chrome to skip the table entirely.

    @stevefaulkner Yeah the cells themselves were logical. I was hoping to replicate what was read on full view as I found the readout confusing without scope

  196. MichielBijl
    12:35
    @StommePoes sorry, set to role of group apparently.
  197. MichielBijl
    12:35
    Weird, said “No exact ARIA role” last time I looked
  198. StommePoes
    12:37
    that Chrome wipes itself correctly keeps surprising me; in my opinion if everyone else does work, I chalk it up to "this browser is so broken, no assistive-needs user would touch it currently.
  199. StommePoes
    12:37
    If Chrome isn't the only one tho then I'd look further
  200. StommePoes
    12:38
    Maybe they spent too much time making their own screen reader that they now need to catch up to working with "real" ones.
  201. stevefaulkner
    12:39
    @JimBobSquarePants If a user is accesing the site via a mobile, they would know about the desktop table structure. I tried in iOS with VO again worked fine, content order was good, I think that the output is acceptable as the association between the header and data is provided via content order (but I am not an end user).
  202. MichielBijl
    12:39
    @StommePoes talking about ChromeVox?
  203. StommePoes
    12:40
    @Michiel and the fact that until recently, chrome didn't do jack with SRs
  204. stevefaulkner
    12:40
    I think that ideally in mobile view would be expsoed as list items, but unsure how practical that is for the code.
  205. StommePoes
    12:40
    it's catching up pretty fast tho
  206. StommePoes
    12:40
    could inject roles I spose
  207. StommePoes
    12:41
    but that seems like over-work
  208. MichielBijl
    12:44
    @steve: wouldn't a dl be more apropriate?
  209. MichielBijl
    12:44
    appropriate*
  210. MichielBijl
    12:45
    Hide the word 'title' and set the name of the movie as the heading
  211. JimBobSquarePants
    12:45

    @StommePoes Testing further both Opera and Chrome skip with the additional role, Must be a Blink thing.

    @stevefaulkner Adding relevant roles/attributes to define the table as a list would be fairly trivial. Rewriting anything else like changing the actual elements would be tricky though.

    There's not really that much JavaScript involved populating the data-attributes. https://github.com/ResponsiveBP/Responsive/blob/45d81146bd23fd7792b5289438e3663c90cb189b/src/js/responsive.tablelist.js

  212. StommePoes
    12:45
    I luv me sum DLs for 2-col tables but the starwars table doesn't fit in there more nicely than some nested list
  213. StommePoes
    12:47
    Maybe the easiest thing (after user-testing to see if users have problems or find it as easy as steve did) is to manually author the roles
  214. MichielBijl
    12:48
    Could be easily automated, if the correct markup is used.
  215. StommePoes
    12:48
    Usually these things are templated... our tables are mostly only 1 tr and 1td with for loops around them...
  216. MichielBijl
    12:50
    True
  217. StommePoes
    12:51
    sigh, Opera has turned (like the pistachio ice cream)
  218. StommePoes
    12:58
    I'm reading Joe Clark's thingie Steve posted, at the fore-background colours section, and I'm expecting him to pull a Godwin.
  219. StommePoes
    13:03
    "If your browser or adaptive technology cannot display or render the title attribute of a link or of anything else, then it’s broken."--Joe Clark
  220. StommePoes
    13:04
    That would be... all of them. Doesn't the majority, even a wrong majority, get to make the rules? Browser democracy is a direct democracy
  221. MichielBijl
    13:05
    So, the fact that acceskey doesn't do anything is not my fault and I shouldn't care AT doesn't doe anything with it, good, load of my mind!
  222. StommePoes
    13:05
    I avoid them mostly
  223. MichielBijl
    13:05
    That was sarcastic…
  224. StommePoes
    13:05
    Tho I know some people who have indeed found them useful
  225. StommePoes
    13:06
    He's sayin what Steve has said for a while about AT... they do need to be called to task for any broken stuff. And the point that users have more control than they know, sure.
  226. StommePoes
    13:06
    But I'll still call a site written in 9px body text a site built by morons, unless it's a pixel-art site dedicated to smallest-glyphs-evar
  227. StommePoes
    13:07
    But, yeah, i can ctrl++ (i do anyway)
  228. StommePoes
    13:08
    Can't tell if he deliberately made the closing </a> bright yellow on white to make a statement or not
  229. JimBobSquarePants
    13:10

    OK, so a combination of id's on the column headers,aria-labelledby, and aria-role=grid works well with Firefox. It reads the titles before each cell which makes it really easy to follow IMO. Yay Firefox! (It does say the table has zero rows and columns though)

    IE will announce the correct number of rows, columns, the position of each cells location in the table and its content in the correct order but not announce the header at any point. Opera and Chrome will announce the content in the correct order but no headers nor position.

    I think setting the position as absolute in the CSS for the header is what throws everything off, I don't see another way of styling the elements though so I feel like I might have got it the best I can for now though with the new roles.

  230. StommePoes
    13:12
    everyone reads out all the content tho, right?
  231. StommePoes
    13:12
    and in a sane order?
  232. JimBobSquarePants
    13:13
    Yeah the cell content in all browsers is read out in the correct order. I feel that not announcing the column headers makes context confusing though.
  233. StommePoes
    13:14
    Do you have any people you can test that on?
  234. StommePoes
    13:14
    and, is there any way for someone to navigate back up and find the headers?
  235. StommePoes
    13:15
    like, when you're in something that does Forms Mode and you feel something's missing, you go out of Forms Mode to sweep for loose <p>s and stuff to see if context was missed... it's not fun, but people do it.
  236. JimBobSquarePants
    13:17

    I can shout out and see if anyone volunteers, I don't get much input normally sadly.

    No. There isn't, still learning this stuff tbh

  237. StommePoes
    13:18
    It would be soooo cool if there were some (paid?) service random web devs could use, like browserStack, but that connects dev sites to volunteer/paid disabled users, sortof like those loop11 style sites but then focussed on users with AT
  238. StommePoes
    13:18
    Our 7-person company doesn't do any user testing either. It's frustrating but more common than not.
  239. MichielBijl
    13:20
    Our 200+ person company doesn't do any user testing either…
  240. JimBobSquarePants
    13:23

    That would be a very useful service yeah. User testing has been something sorely missing from every company I have worked for.

    My framework is a one man show really. I've managed to get quite a few people to use it and it's got quite a healthy following considering it doesn't have a big company name to go along with it but as far as feedback or contribution goes it's very quiet. I'd like to think that is because it "just works" :wink: but I think it's much more likely that people just don't test stuff enough.

  241. StommePoes
    13:24
    And they'll all say, with their mouths, that it's a Good Idea. Both my bosses said it was a Good Idea. But I don't get paid enough to afford taking people to lunch or setting up a lab. I did once stand outside a grocery store and tried to get random people but they were freaked out and ran away
  242. StommePoes
    13:24
    shrug we're limited in what we can do
  243. StommePoes
    13:24
    but with accessibility we have to be a bit more careful
  244. StommePoes
    13:24
    esp when we're playing with aria stuff
  245. StommePoes
    13:26
    and we're probably not too far off guess what random J user does, has trouble with, or wants. But users ofspecialised software is pretty different
  246. dylanb
    13:55
    @JimBobSquarePants looked at your discussion on tables. My conclusion when I researched this recently is that you cannot get really responsive tables to work well across all ATs. I came up with two alternative solutions.
  247. dylanb
    13:56
    This first one, inserts the headers into the cells at the breakpoint http://dylanb.github.io/rwd/part4/responsive.html
  248. dylanb
    13:56
    This second one (which is good for comparisons) uses horizontal scrolling http://dylanb.github.io/rwd/part4/comparison.html
  249. dylanb
    13:57
    which solution you choose (or if you choose something else) is probably more of a UX decision
  250. JimBobSquarePants
    16:20

    @dylanb Thanks for that.

    Yeah I considered adding extra elements to ensure that the headers were read but I feel now that Chrome and Opera really need to do some catchup since I can get both IE and Firefox to read them. That second example I've actually used elsewhere http://responsivebp.com/css/tables/#overflow
    Incidentally there's a bug in windows phone that breaks the scroll functionality on pages rendered right-to-left. Scrolling goes the wrong direction.

  251. zakim-robot
    16:28
    [Jonathan Yung, a11y] Hey all! If I have a button that looks like a link, is it always preferred to use <button> and style it like a link instead of using <a role="button">?
  252. MichielBijl
    16:57
    If it's a button then use <button>
  253. MichielBijl
    16:57
    How come a button looks like a link?
  254. zakim-robot
    16:58
    [Katy Moe, a11y] I guess the question is: does <a role=“button”> have any valid use cases?
  255. zakim-robot
    17:01
    [Ian Pouncey, a11y] <a role=“button”> on its own isn’t keyboard accessible. You need to have a href attribute or specify tabindex=“0” for that. If you use a href then it needs to go somewhere (it will still have a links context menu, so what happens if the user tries to open it in a new tab / window). If you use tabindex you’ve got to make sure the expected keyboard behaviour is implemented. By the time you’ve done all that you might as well use a button.
  256. zakim-robot
    17:03
    [Jonathan Yung, a11y] @Michiel Bijl: Some designs have buttons look like links to put more emphasis on another button (e.g. "Sign up" to look like default button and "Log in" to look like a link")
  257. MichielBijl
    17:04
    Why would you devaluate a Log in button to a sign up button? Don't you want users to log in?
  258. MichielBijl
    17:05
    That would be a question for the designer I guess, but still…
  259. MichielBijl
    17:05
    It's still a button, so <button> :)
  260. zakim-robot
    17:05

    [Jonathan Yung, a11y] @Ian Pouncey: exactly! regarding:

    so what happens if the user tries to open it in a new tab / window
    that can be solved via preventDefault(), but I totally agree that if you spend all the effort to make it behave like a button, use a button!

  261. zakim-robot
    17:06
    [Jonathan Yung, a11y] hahah totally agree. I personally hate sites that devaluate a Log in button especially since I am already a user. But I guess that's for sites that care more about Sign up metrics ¯_(ツ)_/¯
  262. zakim-robot
    17:08
    [Katy Moe, a11y] So there aren’t any valid use cases for <a role=“button”>? I’m guessing the same applies to other elements other than <button> to which you’ve bolted on the button role
  263. MichielBijl
    17:11
    I've written about role="button" on a button a while back, turned out the use case described is bogus: http://www.michielbijl.nl/2014/09/25/buttons-with-a-double-function/
  264. MichielBijl
    17:11
    Describes a button/link that opens a modal
  265. zakim-robot
    18:12
    [Marcy Sutton, a11y] I sure wish the bot posts would hyperlink links
  266. zakim-robot
    18:14
    [Marcy Sutton, a11y] Oh I see @Alice Boxhall already reported it in #slack-a11y-feedback! Hi-Five! :hand:
  267. techthomas
    18:19
    Anyone know where to log issues with Android Talkback functionality? Is the source online anywhere for review? https://code.google.com/p/eyes-free/ is from 2013 Thanks
  268. zakim-robot
    18:37
    [Alice Boxhall, a11y] @techthomas apparently https://groups.google.com/forum/#!forum/eyes-free is the way to go for now.
  269. zakim-robot
    21:07
    [Katy Moe, a11y] @MichielBiji Thanks, interesting article. So the conclusion is that <a role=“button”> is used in this case for progressive enhancement reasons: without JavaScript, the role=“button” isn’t added and the link does in fact navigate somewhere, whereas with JavaScript, the element behaves as a button. Is this correct?
  270. zakim-robot
    21:08
    [Katy Moe, a11y] If so, why not go the whole way and use JS to transform the <a> into a <button> when it takes on that role?
  271. zakim-robot
    23:39
    [shalom, a11y] howdy :simple_smile:
  272. zakim-robot
    23:40
    [Rick Brown, a11y] shalom :simple_smile: