Archive index

A11y Slackers Gitter Channel Archive 2nd of November 2015

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    04:21
    [snehal] Shout-out to @hmig for mentioning this at CSS Dev Conf last week! :+1:
  2. MichielBijl
    08:17
    @stevefaulkner not if you're French or Danish.
  3. MichielBijl
    08:18
    They base 20 I believe.
  4. MichielBijl
    08:53
    They use*
  5. zakim-robot
    09:28
    [katy] What are people’s opinions on the “fields marked with a * are mandatory” message on forms? Should it be compulsory for all forms that use that notation, or are people familiar enough with it not to need it these days?
  6. LjWatson
    09:30
    @Katy the * convention is abstract and may be difficult for people with certain cognitive disabilities to interpret.
  7. LjWatson
    09:30
    The * character may not be read by default with certain screen readers.
  8. LjWatson
    09:30
    So in both cases the statement at the start of the form is helpful.
  9. LjWatson
    09:31
    On Gov.UK we're trying to move away from using * at all, and to use "Required" or "Optional" - depending on whether the form has more of the former or more of the latter.
  10. MichielBijl
    09:37
    @LjWatson so it would be <label>First name (required)</label> or something?
  11. LjWatson
    09:38
    @MichielBijl yes.
  12. LjWatson
    09:38
    If the majority of fields in the form are required, then you only need to call out the optional fields. Or vice versa.
  13. MichielBijl
    09:38
    That looks better indeed
  14. MarcoZehe
    09:38
    Hi there!
  15. LjWatson
    09:38
    It went down well in usability testing we found.
  16. MichielBijl
    09:39
    It's Marco!
  17. MichielBijl
    09:39
    @LjWatson that is actually very smart, thank you for sharing :)
  18. LjWatson @LjWatson waves at MarcoZ
  19. LjWatson
    09:39
    @MichielBijl No problem :)
  20. MarcoZehe @MarcoZehe waves back @LjWatson
  21. MarcoZehe
    09:41
    Hi @MichielBijl :-)
  22. MichielBijl
    09:41
    Hello :D
  23. MichielBijl
    09:43
    Question for screen reader users. Couple of lines back, I mentioned Léonie, directly after that, she used the /me command to say something. To me the only difference to those lines visually is the colour. Is there any difference in what that AT announces?
  24. MichielBijl
    09:43
    Guess you could get it from the wording, but not much else.
  25. MarcoZehe
    09:44
    @MichielBijl: That depends on the client you're using. I'm now, as I just set this up, on Gitter, which is really not that friendly. My normal IRC client, IrcCloud, denotes such lines with a - (dash) sign.
  26. LjWatson
    09:46
    I use Chatzilla which doesn't give any indication AFAIK.
  27. MarcoZehe
    09:46
    Having said that, I really don't understand why we don't use good old IRC instead of these more or less proprietary and centralised services that may not even be there in 2 or 3 years. ;)
  28. LjWatson
    09:46
    I don't use the Gitter UI - just plug it into my IRC client.
  29. LjWatson
    09:46
    @MarcoZehe is IRCCloud a Mac application?
  30. MarcoZehe
    09:47
    I will do that too in a minute or so, I just got this set up. I fiddled with Slack before, but found it very dissatisfying even with the iOS app, so just deleted my account there.
  31. MichielBijl
    09:49
    Thank you both.
  32. MichielBijl
    09:49
    I used to use this via IRC, but my work network blocks port 6667.
  33. MarcoZehe
    09:59
    OK, now with my IRC client.
  34. zakim-robot
    10:09
    [katy] Thanks @ljwatson, that’s an excellent idea
  35. zakim-robot
    10:09
    [katy] We’ve used aria-required to make sure screen readers indicate the fields are required, but you make a good point about cognitive disabilities.
  36. zakim-robot
    10:09
    [katy] I’m also wondering whether the * symbol is interpreted as “required” internationally.
  37. MarcoZehe
    10:24
    Katy: You can always put " denotes required field" above the form to indicate that the is meant exactly that. That helps non-tech people in any language better understand what you're trying to tell them.
  38. zakim-robot
    10:34
    [katy] @MarcoZehe Yes, that is one of the options we have. Currently we have an informational banner saying that, but the designers want to remove it.
  39. MichielBijl
    10:37
    Compose a list of points to keep it, and ask them to do the same (but why they want to remove it).
  40. zakim-robot
    10:42
    [katy] Thanks. I’m not necessarily on one side or the other, though - I’m interested in creative solutions to the problem. Leonie’s was a good example - just don’t use asterisks in the first place.
  41. zakim-robot
    10:43
    [katy] At the very least we’re going to have to replace the fixed banner with some text at the top of the form or similar.
  42. MichielBijl
    10:46
    That would be better. I'm in my morning warrior mode
  43. MichielBijl
    10:47
    http://dir.agosto.nl/dump/300.jpg
  44. MichielBijl
    10:48
    alt = My face photoshopped on a spartan warrior holding a sword (from the movie 300).
  45. MichielBijl
    10:48
    And it is probably not called a sword, but I don't know the proper name for it.
  46. MichielBijl
    10:53
    Working together is always better than creating teams and separating people further.
  47. MarcoZehe
    11:01
    MichielBijl: Yes definitely!
  48. MichielBijl
    11:15
    Is it possible to configure your own modifier key in NVDA?
  49. MichielBijl
    11:16
    Caps lock doesn't seem to work in VirtualBox running on a MacBook Pro. And that doesn't have an insert or extended insert key.
  50. MarcoZehe
    11:16
    @MichielBijl Yes, you can set either the CapsLock and/or Insert Keys as NVDA modifiers. In Preferences/Keyboard settings.
  51. MichielBijl
    11:17
    @MarcoZehe what I meant was like you specify a key to use (say control and shift).
  52. MarcoZehe
    11:20
    @MichielBijl No, that is not possible.
  53. StommePoes @StommePoes walks in, without IRC client tho :(
  54. StommePoes @StommePoes hates emoticons, that was a colon-left-parenthesee
  55. StommePoes
    13:35
    I liked entering this room with IRC, but I do have to admit, having the webby version keep history and even letting me know that I was mentioned or that I have x-number posts since the last time I was logged in is useful. My Pidgin did nothing more than highlight the room tab with a colour change, which isn't enough of a notification to alert me.
  56. MarcoZehe
    13:36
    @StommePoes I use IRCCloud as my IRC client, which also keeps me logged in and provide me with that, and is more accessible than Gitter's interface.
  57. MichielBijl
    13:37
    https://www.irccloud.com
  58. MarcoZehe
    13:37
    Yes that one.
  59. StommePoes
    13:38
    But as a lazy cheap socialist raised on IRC, I haven't considered paying for "IRC with history" :P
  60. MichielBijl
    13:38
    4 euro a month tho.
  61. StommePoes
    13:38
    arg stupid smilies
  62. MarcoZehe
    13:38
    Mozilla runs an enterprise version of that, because we do everything over IRC that requires instant communication. So I just added the a11ySlackers there. ;)
  63. MichielBijl
    13:38
    We need more people in this thread to stop the smiley madness! https://github.com/gitterHQ/gitter/issues/949#issuecomment-152551398
  64. StommePoes
    13:38
    I noticed, even if I typed my text in Pidgin with the smilies off, here on the webby page they'd still be translated to silly stupid yellow people heads
  65. StommePoes
    13:40
    Marco yeah, I read about IRCCloud on that article I think you tweeted, and in the comments it did seem cool and would make sense for companies. It wouldn't surprise me if my company was considering something like that too (tho I doubt there's a lot of old IRCers there, those folks tend to move to stuff like IRCCloud while the non-IRCers go to other stuff).
  66. MarcoZehe
    13:40
    They come through as characters on IRC, so the translation happens in the Gitter web interface with the raw text.
  67. StommePoes
    13:40
    Yeah but they get assigned stupid names
  68. StommePoes
    13:40
    like : some weird text
  69. StommePoes
    13:40
    Back when I did my more regular testing with SRs I was already mentally translating colon-left-parens stuff in my head to faces
  70. StommePoes
    13:41
    Though, the emoticons expand that to, like, a bazillion new options
  71. MarcoZehe
    13:41
    Not here, they came through as colon, then left paren here. The :some text: bit might be pidgin, think I've seen it there.
  72. StommePoes
    13:42
    "the : confounded : ome text: bit might..."
  73. StommePoes
    13:42
    the bigger problem is it'll come right in the middle of words
  74. StommePoes
    13:42
    which, you'd think most of the time, wasn't anyone
  75. StommePoes
    13:42
    anyone's intentions at all
  76. StommePoes
    13:44
    in any case, any emoticon words you see coming from my text, is not my text, but somewhere either an ascii emoticon or, even more likely, I accidentally a colon in my text
  77. MarcoZehe
    13:44
    As for paying 4 euros a month, I would do that and deal with an interface that is consistent, and an iOS app that is accessible, too, instead of trying to fiddle with Adium on Mac, Pidgin on Windows or Linux and some other IRC client on mobile that wasn't. ;) And which would require my machines to be online all the time and awake enough to record all that's
  78. MarcoZehe
    13:44
    coming in. I find IRCCloud a very convenient service.
  79. MichielBijl
    13:44
    Back when I did my more regular testing with SRs I was already mentally translating colon-left-parens stuff in my head to faces
    This, so much this!
  80. StommePoes
    13:44
    pidgin is accessible, on linux
  81. MichielBijl
    13:44
    How the hell do quotes work here…
  82. StommePoes
    13:45
    prolly not on windows since it's some lib-purple port, I dunno, but Orca gets on fabulous with it : )
  83. StommePoes
    13:45
    no idea, I've been quoting people with " symbols
  84. StommePoes
    13:45
    you probably need to add an extra whitelinespacething between
  85. StommePoes
    13:45
    they use markdown down here... we all float down here
  86. MichielBijl
    13:45

    no idea, I've been quoting people with " symbols

    Maybe you need a return after the quote

  87. MichielBijl
    13:45
    Yeah
  88. MarcoZehe
    13:45
    That "inaccessible" part referred to other IRC clients on iOS only. :) Pdgin is accessible on Windows even, so I hear. Or Trilian or what it's called there.
  89. MichielBijl
    13:45
    You need two returns…
  90. StommePoes
    13:46
    yeah, markdown wishes it was python and is whitespace sensitive
  91. MarcoZehe
    13:46
    LOL
  92. StommePoes
    13:46
    I've heard of Trilian, I'll have to go see if it's nice. I have this new job and they want everyone using Windows so I've got a Win Box and miss a lot of my usual stuff. I'm actually still going over to my work machine to do some things still anyway.
  93. StommePoes
    13:47
    markdown's better than that wannabe markdown Python folks use, that nasty sphinx. hateses it I do
  94. StommePoes
    13:47
    I tricked people coming to my Sphynx talk into thinking I was talking about Sphynx ReStructuredText.
  95. StommePoes
    13:47
    They ended up getting a much more interesting topic.
  96. MarcoZehe @MarcoZehe chuckles.
  97. StommePoes
    13:48
    Oh boy, today I have to use WebEx.
  98. MarcoZehe
    13:48
    So @StommePoes, are you still in the Netherlands, or did your new job prompt you to move back to the U.S.?
  99. StommePoes
    13:48
    Yet Another Weird Communication Tool
  100. StommePoes
    13:49
    Nope I'm still home here in NL, so it's remote, but it was a problem for them to set stuff up as, while there is a Pearson Benelux, it's really small and just edumacation-oriented, while I'm in the techy/producty team (which is mostly US and a little UK)
  101. StommePoes
    13:49
    I don't think I'd move to the States for a job. What if I cut my finger? I could remortgage my house to pay for what insurance says it won't cover and end up sleeping in a van or something
  102. MarcoZehe
    13:50
    But they definitely wanted you to work for them, so they managed! Yay! :-)
  103. StommePoes
    13:50
    I don't want to get involved with US healthcase. Also I like our crappy, often-broken public transport, luv
  104. StommePoes
    13:50
    Yeah I was surprised, I heard via via someone that an Apple tech or something of some sort had applied. Surely they know more than I do!
  105. StommePoes
    13:51
    er s/healthcase/healthCare
  106. MarcoZehe
    13:52
    Heh, healthcase (I even misread it as healthchase) is fitting! ;-)
  107. StommePoes
    13:52
    heh
  108. MichielBijl
    13:52
    @StommePoes to quote Steve and Léonie “we have no clue as to what we do”
  109. MichielBijl
    13:52
    Same applies to you and me I guess
  110. StommePoes
    13:52
    Yeah same here except they have much longer experience : P
  111. MarcoZehe
    13:53
    I'll join in that club! I certainly don't know what I'm doing. :D
  112. MichielBijl
    13:53
    More experience in not knowing what the hell we're doing :P
  113. StommePoes
    13:53
    And so I was asked to review some stuff for JS devs and there's this part on Angular and I'm like "I dunno Angular" and then I remember they put words like "Specialist" and "Javascript Developer" in my job title and I cringe
  114. StommePoes
    13:54
    Luckily it was broken enough with minor typo's that I could still kinda comment on it.
  115. MarcoZehe
    13:54
    That's the thing, you know JavaScript, you don't have to rely on Angular or these other "tools" to help people not write JavaScript.
  116. StommePoes
    13:54
    But I cannot look at a recommendation do to "x" in Angular without knowing Angular, and this recommendation was very Angular specific... knowing Javascript wouldn't save my bacon in this case.
  117. StommePoes
    13:55
    And it was an example of "how you could do the manual focus management necessary by the choice of using Angular" but my husband, who's been writing Angular for like 2 years now, was able to also look at it. And then he considered it full of Bad Practice, but I see it was referencing an online article. In any case, I figured I was out of my depth there.
  118. MichielBijl
    13:56
    StommePoes the Full Stack Specialist JavaScript Developer
  119. StommePoes
    13:56
    A full stack of waffles
  120. MichielBijl
    13:56
    That would be better
  121. StommePoes
    13:56
    Full Stack IHOP developer
  122. StommePoes
    13:56
    That means I eat at the laptop while typing
  123. MichielBijl
    13:56
    I'm a full stack slacker
  124. StommePoes
    14:00
    Meh, on the other hand, I know I'll get to learn a lot of cool stuff.
  125. StommePoes
    14:00
    And some day I might get a copy of JAWS that's not breaking the EULA : P
  126. StommePoes
    14:00
    And hopefully Dragon. That will be le cool
  127. MichielBijl
    14:01
    I just installed NVDA today
  128. MichielBijl
    14:01
    And cannot use it because of VIrtualBox and OS X :(
  129. MichielBijl
    14:01
    I'll have to look into it.
  130. MichielBijl
    14:03
    Some people say list-style: none interferes with AT (not announcing list or item count). What are other suggestions to remove default bullets?
  131. MichielBijl
    14:03
    And which AT have trouble with this?
  132. MichielBijl
    14:03
    Because VO handles it fine.
  133. MichielBijl
    14:04
    Whoop, no it doesn't
  134. MichielBijl
    14:05
    It does however announce the number of items in a nav-element :)
  135. MichielBijl
    14:06
    But I don't know if other AT does so too
  136. StommePoes
    14:07
    I've heard a similar issue, but even if it's true, I will continue to use list-style none.
  137. StommePoes
    14:07
    AT bugz is bugz.
  138. StommePoes
    14:08
    What you'd think might honestly mess with ATs tho is when you float li's. People used to wonder, why are the bullets gone if you float them? (actually this depended on browser back then)
  139. StommePoes
    14:08
    And it was, float changed the li's from display: list-item to float. So they no longer had the list-item display state, kinda like changing tables to non table display (like block) is an issue sometimes, or making non-tables display: table.
  140. MichielBijl
    14:09
    I'll use div's! That'll fix everything!
  141. powrsurg
    14:23
    Given how common floating block level (or inline-block level) <li> elements is, I feel like AT should be smart enough to pay attention to the tag and stick with that value
  142. MichielBijl
    14:25
    Or just look at what's in the accessibility tree…
  143. MarcoZehe
    14:25
    We just fixed a bug in Firefox where we no longer expose table semantics for something that isn't a table markup-wise, but only has CSS properties display:table; and friends.
  144. MarcoZehe
    14:26
    If we have a bug in Firefox where we don't display lists or list items as such any more due to some CSS property, let me know. In these cases, I think we should always go with the markup rather than the CSS.
  145. powrsurg
    14:26
    of course, that would explain why I was having issues with ChromeVox recently that just kept on reading my lists simply had an checkbox (which was visually hidden) and a label (which had a different image using CSS generated content for an image -- rather than background so it'd show up in Windows high contrast mode) and started to simply just read each list item as "list item image"
  146. MarcoZehe
    14:27
    ChromeVox, at least the 1.x version, is a story of its own, since it does its own markup interpretation and doesn't use Chrome's accessibility tree. The new version, as far as I know, will do that.
  147. MichielBijl
    14:37
    I'm so in love with the @supports attribute!
  148. MichielBijl
    14:38
    No more JS to check support
  149. StommePoes
    14:40
    Does that deal properly with the times when we'd check for something (like, say, Opera 12 supporting pointer events) and the browser'd be all like "yeah I support that" but it turns out it actually doesn't? (Opera would say yes but then do nothing with it)
  150. powrsurg
    14:46
    will chromevox be going away with the android/chromeOS merge?
  151. MichielBijl
    14:46
    Hmm, dunno
  152. powrsurg
    14:47
    at least with what I was doing, android talkback was doing a better job =/
  153. MichielBijl
    14:48
    @StommePoes http://caniuse.com/#feat=css-featurequeries
  154. MichielBijl
    14:48
    Doesn't say anything about bugs in O12
  155. StommePoes
    14:49
    It did not support pointer events at all. But, it did say that it did, if you asked it in Javascript.
  156. StommePoes
    14:49
    That is, setting pointer-events to whatever did nothing in Opera12.
  157. StommePoes
    14:50
    I learned about this during a talk by Lea Verou, was also the first time I'd heard of pointer-events at all.
  158. StommePoes
    14:50
    But so I mean, browsers sometimes say they support things but that's not always true, just wondered if this @supports thing knew that
  159. StommePoes
    14:51
    or here's another one, IE9 has support (finally) for onInput events, and it mostly supported them but you'd still want to detect it because it manages to completely ignore the delete key : P
  160. StommePoes
    14:52
    Oh man, now I'm being asked to do Skype. Yuck
  161. StommePoes @StommePoes is off to go download it
  162. StommePoes
    14:52
    Some day someone will solve all this communication stuff with One Application And Protocol To Rule Them All
  163. StommePoes
    14:56
    "Your Content may appear in demonstrations or materials that promote the Service. "
  164. MichielBijl
    14:58
    Sounds like a no go for componies
  165. MichielBijl
    14:58
    companies
  166. MichielBijl
    14:58
    I would like to see componies tho
  167. powrsurg
    15:01
    I believe there are some versions of (older) BlackBerry that supported touch events at the software level, but didn't at the hardware level
  168. StommePoes
    15:09
    com ponies...
  169. StommePoes
    15:10
    Well the helpdesk of my com p0ny uses skype and also all the data-stealing googley bits as well.
  170. MarcoZehe
    15:21
    Well, the NSA has been able to read Skype chat logs since late 2011 at least. So if companies really have pricacy concerns with it, they oughta have acted sooner. ;)
  171. MarcoZehe
    15:21
    s/pricacy/privacy/
  172. dylanb
    15:25
    @MichielBijl report the list style issue as a bug with all the AT vendors. That treatment goes against semantic markup.
  173. MarcoZehe
    15:38
    MichielBijl: And if you have a test page accessible somewhere, could you share that here? I'd just like to make sure Firefox creates the accessible tree properly. I know we do it right for list-style:none;, but want to make sure for the others you are having issues with.
  174. StommePoes
    15:42
    Yeah Orca gets lists right in FF
  175. MichielBijl
    16:07
    @MarcoZehe you can find it here, but that is nothing more than a simple ul with list-style: none
  176. MichielBijl
    16:07
    This message was deleted
  177. MichielBijl
    16:07
    @MarcoZehe whoops, forgot the link: http://dir.agosto.nl/accessibility/list.html
  178. MarcoZehe
    16:11
    @MichielBijl Yep, as I suspected, that one works in Firefox and NVDA. A list with 4 items, and each item has text, but no bullet, as expected.
  179. MichielBijl
    16:13
    Correct :)
  180. MichielBijl
    16:13
    Updatet it with an explaination
  181. MichielBijl
    16:14
    Which version of Windows do you run @MarcoZehe?
  182. MarcoZehe
    16:15
    @MichielBijl: Windows 10, but the only thing I really run in it is Firefox and NVDA.
  183. MichielBijl
    16:16
    Thank you; I've updatet the page to include our results.
  184. MarcoZehe
    16:16
    And that combination is pretty Windows-agnostic, e. g. you get same results in 7, 8.1, or 10.
  185. MichielBijl
    16:17
    I'll try with El Capitain tonight
  186. powrsurg
    16:17
    I guess a work-around for any list-style:none issues would be to instead use "list-style-image: url(data_uri_for_transparent_gif)"
  187. MichielBijl
    16:18
    Would list-style-image: linear-gradient(transparent, transparent); be accepted?
  188. powrsurg
    16:22
    a quick test in my firefox list-style-image doesn't seem to support linear-graidents =/
  189. MichielBijl
    16:22
    Safari 9 does :)
  190. powrsurg
    16:26
    Go Safari :p
  191. MichielBijl
    16:33
    Edge has no support for it.
  192. powrsurg
    16:49
    So ... list-style-image with the data URI seems to give you the most support =/
  193. StommePoes
    17:22
    still seems like a hackitty hack we should avoid
  194. MichielBijl
    17:22
    Yep
  195. MichielBijl
    17:23
    So, log all the bugs!
  196. StommePoes
    17:23
    HULK SMASH BUGGS
  197. MichielBijl
    17:24
    Not sure if this is an VO of WebKit bug
  198. MichielBijl
    17:24
    To the accessibility inspector (mobile)!
  199. StommePoes
    17:33
    goddamm... must use Skype for work, and now I get ad banners. Blehhhhh
  200. MichielBijl
    17:34
    No Skype support in Pidgin?
  201. zakim-robot
    18:09
    [jiatyan] Doesn't Skype use proprietary protocol?
  202. powrsurg
    18:35
    Well, MS has been tweaking stuff to make Skype available via web RTC
  203. powrsurg
    18:36
    it requires browsers to support a specific set of the standard (which of course are the ones MS implemented first and neither Chrome nor Firefox completely implement yet)