Archive index

A11y Slackers Gitter Channel Archive 26th of February 2016

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    Feb 26 00:10
    [karlgroves] One talent I definitely never managed to develop was design. I’m soooo terrible at it
  2. zakim-robot
    Feb 26 00:12
    [hmig] lol. I just pr’d to visualregressiontesting.com because there was a pixel sticking out
  3. zakim-robot
    Feb 26 00:12
    [karlgroves] heheh
  4. zakim-robot
    Feb 26 00:13
    [hmig] I did the meanest thing - I pointed it out visualregressiontesting/visualregressiontesting.com#1
  5. zakim-robot
    Feb 26 00:13
    [hmig] once you see it, you can’t unsee it :smile:
  6. zakim-robot
    Feb 26 00:20
    [hmig] question: why is it bad to use aria-label on an <a href>?
  7. zakim-robot
    Feb 26 00:20
    [hmig] was told because it replaces the name of the link
  8. zakim-robot
    Feb 26 00:21
    [hmig] but, I’m not sure why that’s bad
  9. zakim-robot
    Feb 26 00:28
    [karlgroves] Users of voice dictation software will want to be able to reference the link by its visual text.
  10. zakim-robot
    Feb 26 00:29
    [karlgroves] If the voice dictation software supports aria, the accessible name will be different than the visual text
  11. zakim-robot
    Feb 26 00:29
    [hmig] in the instance I’m using it, I want the aria-label to be what is conveyed
  12. zakim-robot
    Feb 26 00:30
    [hmig] I’m working on svg icon patterns
  13. zakim-robot
    Feb 26 00:30
    [hmig] when they are used as links + text
  14. zakim-robot
    Feb 26 00:30
    [hmig] the last example https://jsbin.com/zarifo
  15. zakim-robot
    Feb 26 00:31
    [hmig] note - I know the first example is not working with narrator ...
  16. zakim-robot
    Feb 26 00:31
    [hmig] working on that one still
  17. zakim-robot
    Feb 26 00:32
    [karlgroves] I think the use case there is equivalent to a linked <img>
  18. MichielBijl
    Feb 26 00:33
    Not sure what you're testing, but this might come in handy: https://rawgit.com/MichielBijl/font-awesome/master/tests/svg.html
  19. zakim-robot
    Feb 26 00:34
    [karlgroves] Sorry (and 2nd)
  20. zakim-robot
    Feb 26 00:34
    [hmig] @karlgroves: not sure what you mean there
  21. zakim-robot
    Feb 26 00:34
    [karlgroves] Sorry. Typing before thinking
  22. zakim-robot
    Feb 26 00:35
    [karlgroves] I think what you’re doing is fine given this use case. aria-label is the best approach for a text-based alternative for this and is equivalent to img[alt]
  23. zakim-robot
    Feb 26 00:35
    [hmig] yay!
  24. zakim-robot
    Feb 26 00:36
    [hmig] Yea - I’m trying to ditch the use of the sr-only spans
  25. zakim-robot
    Feb 26 00:36
    [hmig] and hiding the svg
  26. zakim-robot
    Feb 26 00:37
    [hmig] that might be the way I’m forced to go regarding narrator/edge support for the non-linked icon though
  27. zakim-robot
    Feb 26 00:37
    [hmig] time to file bugs :smile:
  28. zakim-robot
    Feb 26 00:38
    [hmig] I wish the screen readers would just read the title in src svg for each symbol
  29. zakim-robot
    Feb 26 00:39
    [karlgroves] aria-labelledby with an ID on the title
  30. zakim-robot
    Feb 26 00:39
    [hmig] doesn’t work with <use>
  31. zakim-robot
    Feb 26 00:39
    [hmig] I tried it
  32. zakim-robot
    Feb 26 00:40
    [hmig] maybe if I nested the titles in svg > title > use
  33. zakim-robot
    Feb 26 00:40
    [hmig] but that’d be redundant to what most generators output for the svg file
  34. zakim-robot
    Feb 26 00:42
    [hmig] @karlgroves: ty for confirming that my approach for the linked examples is ok
  35. MichielBijl
    Feb 26 00:47
    @karlgroves, @hmig, aria-label doesn't always work: https://rawgit.com/MichielBijl/font-awesome/master/tests/svg.html#test-3
  36. MichielBijl
    Feb 26 00:47
    Err
  37. zakim-robot
    Feb 26 00:47
    [hmig] @michiel: interesting …
  38. MichielBijl
    Feb 26 00:48
    Need to test VO though…
  39. zakim-robot
    Feb 26 00:48
    [hmig] I put the aria-label on the link
  40. MichielBijl
    Feb 26 00:49
    Interesting. So <a href="rawr.eu" aria-label="RAWR!"><svg aria-hidden="true"><path id="to-enlightment"></svg></a>?
  41. zakim-robot
    Feb 26 00:49
    [hmig] yes
  42. MichielBijl @MichielBijl should add test number 8…
  43. zakim-robot
    Feb 26 00:49
    [hmig] last two examples
  44. zakim-robot
    Feb 26 00:50
    [hmig] seemed to work well everywhere
  45. zakim-robot
    Feb 26 00:50
    [hmig] need to test mobile next
  46. zakim-robot
    Feb 26 00:50
    [hmig] my first example is the one I still need to fix
  47. jnurthen
    Feb 26 00:50
    @karlgroves Dragon allows you to speak either the aria-label or the visual text on a link.
  48. jnurthen
    Feb 26 00:51
    Their aria implentation doesn't use the APIs but instead uses a plugin to get the info. Most of the time that sucks but in this case maybe it is a good thing
  49. MichielBijl
    Feb 26 00:54
    I should probably get some sleep, big day tomorrow. Night all!
  50. zakim-robot
    Feb 26 00:56
    [hmig] good night! thanks for your input/links!
  51. zakim-robot
    Feb 26 02:27
    [karlgroves] @jnurthen thanks for that info. Is there anywhere I can find out more?
  52. jnurthen
    Feb 26 02:30
    @karlgroves download their extension and examine the code ;)
  53. zakim-robot
    Feb 26 02:31
    [karlgroves] No! I’m too lazy! I just want a list!
  54. stevefaulkner
    Feb 26 06:54

    @hmig related to aria-label

    Testing screen reader label and description announcement on a elements with href http://thepaciellogroup.github.io/AT-browser-tests/acc-name-test/a-href.html

  55. stevefaulkner
    Feb 26 06:57
    labelling mechanisms don't always work as expected with <svg> element as it's not exposed uniformly across acc API's, addition of usually helps
  56. stevefaulkner
    Feb 26 06:57
    addition of a role usually helps
  57. Thanks @stevefaulkner - bookmarked for reference.
  58. Friends, Romans, Countrymen, good morning.
  59. sophieschoice
    Feb 26 09:12
    Morgen :D
  60. Hey @sophieschoice =)
  61. sophieschoice
    Feb 26 09:14
    Nice graphic of reasons why people leave your site: https://twitter.com/AGConsult/status/703142588488888320
  62. I don't like the "nudge" factor, that tends to become things like Booking.com's "There are 5 people looking at this room! BOOK NOW BOOK NOW IT MIGHT BE GONE AAAUUUUGH". Give my my options clearly and honestly, and I will make a decision
  63. stevefaulkner
    Feb 26 09:17
    morning slackers
  64. Hiya @stevefaulkner
  65. sophieschoice
    Feb 26 09:17
    Hi :D!
  66. sophieschoice
    Feb 26 09:19
    @jvka: yeah, I don't like stuff like that. I am already at the site, looking at prices and available dates/rooms/options. Do you really think I need to be persuaded further?
  67. Right.
  68. THERE ARE 5 PEOPLE LOOKING AT THIS EBOOK BUY BUY BUY QUICKLY
  69. sophieschoice
    Feb 26 09:21
    TOO LATE, I BOUGHT THEM ALL!
  70. NOOOOO
  71. Maybe they can print more next year :books:
  72. sophieschoice
    Feb 26 09:25
    Why is it that they ask in web design for programming skills (PHP, jQuery,....) :( I can't code :(
  73. In Web development I'd say having basic JS skills would be an expected requirement. Purely in web design, perhaps not so much. But where is the line dividing those two?
  74. sophieschoice
    Feb 26 09:40
    Exactly
  75. sophieschoice
    Feb 26 09:43
    wow, those are good questions
  76. Very. And tells me exactly where I need to study up =)
  77. sophieschoice
    Feb 26 09:43
    Ha, same here :D
  78. Fairly sarcastically in places, I might add =P
  79. MichielBijl
    Feb 26 10:19
    It's all pretty honest, but some sarcasm might have crept in :P
  80. sophieschoice
    Feb 26 10:20
    Hahaha, it's great :D
  81. sophieschoice
    Feb 26 10:20
    especially this one: What are your favourite image replacement techniques and which do you use when?
    • Image replacement techniques? To quote a teenage girl
      I don’t even.
  82. MichielBijl
    Feb 26 10:21
    =D
  83. StommePoes
    Feb 26 11:24
    I used to do image replacement, haven't done it in a while
  84. StommePoes @StommePoes waves at @hmig
  85. StommePoes
    Feb 26 11:25
    Had a cool chat with Eva Westerhoff and some conference organisers today, getting information on how (here locally) to make the next DomCode conference completely accessible
  86. StommePoes
    Feb 26 11:26
    and not have that crap where the speaker in the wheelchair at the last minute is forbidden on stage because "there's no fence at the edge" as if he were a child
  87. StommePoes
    Feb 26 11:26
    Also, fantasai is no longer on my love list
  88. StommePoes
    Feb 26 11:26
    even though she did still save me and others from a future of Blink being a douchcanoe filles with shitweasels
  89. StommePoes
    Feb 26 11:27
    loves for that, but "well they say a picture is worth a thousand words so why can't alt text be hundreds of characters" shit is just... unhelpful at its most chariable.
  90. ...shitweasels
  91. StommePoes
    Feb 26 11:29
    Guess she just raises hackles at a tool that tests for possible GUIDEline violations and points them out but calls them critical errors. Fine, I'm sure someone can change the text to "issue found, did you mean...?" if non-linting testing tools have to be less straightforward and ego-bruising.
  92. Yeah, unnecessarily harsh
  93. StommePoes
    Feb 26 11:29
    Yeah fantasai basically got someone who knew someone to get that shitweasel bug fixed at Blink where keyboard focus doesn't go where the skip link points to.
  94. StommePoes
    Feb 26 11:30
    So IE is still and issue and of course I haven't gotten any newer chromes/chromiums yet, but zomg that bug is like over a decade old.
  95. sophieschoice
    Feb 26 11:30
    wait, wut? A decade old?
  96. StommePoes
    Feb 26 11:30
    Well it was a KHTML bug
  97. StommePoes
    Feb 26 11:30
    inherited when Webkit budded from it
  98. StommePoes
    Feb 26 11:31
    inherited again when Blink emerged a fresh babe from the womb of Webkit
  99. StommePoes
    Feb 26 11:31
    so, pretty old
  100. sophieschoice
    Feb 26 11:31
    and no-one ever had a lightbulb moment of fixing it? Jeez
  101. StommePoes
    Feb 26 11:31
    And Webkit did some sort of fix but the chromium guys looked at it and called it a half-fix at best.
  102. StommePoes
    Feb 26 11:31
    Sure, there were multiple bug reports
  103. StommePoes
    Feb 26 11:31
    The last-est one was from @rodneyrehm
  104. StommePoes
    Feb 26 11:32
    Way back when someone posted it as an "accessibility" issue in that it was noticed that the skip links didn't work with a screen reader. So they added a hack that made it work with a screen reader
  105. StommePoes
    Feb 26 11:32
    but the real problem was the general one of bad focus management
  106. StommePoes
    Feb 26 11:33
    and for some reason people kept asking "well what about [some complicated possibility that nobody cares about]? So we can't do it" for why all the other browsers followed skip links to non-focusable destinations just fine (except IE which needs some funky layout rendering function to kick in)
  107. zakim-robot
    Feb 26 12:16
    [rodneyrehm] yes master, you summoned me?
  108. What's that sulphury smell?
  109. zakim-robot
    Feb 26 12:17
    [rodneyrehm] my trail blaze, sorry for the inconvenience :wink:
  110. =D
  111. StommePoes
    Feb 26 12:19
    bwahaha
  112. StommePoes
    Feb 26 12:20
    I had noticed it was your name on the bug that heydon twotted with the (finally zomg since forever) Blink skip-link fix.
  113. zakim-robot
    Feb 26 12:20
    [rodneyrehm] I assume you were talking about the Sequential Focus Navigation Starting Point implementation that recently landed in blink?
  114. zakim-robot
    Feb 26 12:20
    [rodneyrehm] "skip link target” would be just one implication of that
  115. StommePoes
    Feb 26 12:20
    aka they now act like every other freaking browser excepting in special cases IE
  116. zakim-robot
    Feb 26 12:20
    [rodneyrehm] what special cases in IE?
  117. StommePoes
    Feb 26 12:21
    They had two skip link bugs
  118. StommePoes
    Feb 26 12:21
    one was based on haslayout specifically
  119. StommePoes
    Feb 26 12:21
    did exactly what pre-bug-fix Blink does now
  120. StommePoes
    Feb 26 12:21
    Haslayout is gone but still another bug has something to do with rendering that still causes this bug
  121. StommePoes
    Feb 26 12:21
    so like setting a height on the destination element fixes it
  122. StommePoes
    Feb 26 12:21
    or, adding tabindex="-1"
  123. StommePoes
    Feb 26 12:22
    since that also worked from Chrome, we used that to fix 2 browsers at once without extra JS
  124. StommePoes
    Feb 26 12:22
    Yeah I don't know where the KHTML issue is
  125. StommePoes
    Feb 26 12:22
    which is the original source
  126. StommePoes
    Feb 26 12:23
    I encountered the first listing of Chrome's bug https://bugs.chromium.org/p/chromium/issues/detail?id=262171 here
  127. StommePoes
    Feb 26 12:24
    Chromium people consider this one, and yours, and another initially only listed for screen readers and skip links as being solved by this new fix
  128. zakim-robot
    Feb 26 12:25
    [rodneyrehm] I think they merged all the bugs into “mine"
  129. zakim-robot
    Feb 26 12:25
    [rodneyrehm] no idea why, though
  130. StommePoes
    Feb 26 12:26
    the one you list references the half-fixed one I listed, so good to see they're looking at fixing the basic problem in Webkit too
  131. zakim-robot
    Feb 26 12:27
    [rodneyrehm] wait, those are different issues
  132. zakim-robot
    Feb 26 12:27
    [rodneyrehm] or, more precisely two different levels of the same problem
  133. StommePoes
    Feb 26 12:28
    yeah
  134. StommePoes
    Feb 26 12:28
    the last
  135. StommePoes
    Feb 26 12:28
    it's why there had been so many bugs
  136. StommePoes
    Feb 26 12:28
    people encountered them in different ways
  137. zakim-robot
    Feb 26 12:29
    [rodneyrehm] when the fragment target changes, the target element should receive focus (if it is focusable) or be set as the “Sequential Focus Navigation Starting Point” (which is what chrome just implemented)
  138. zakim-robot
    Feb 26 12:33
    [rodneyrehm] excuse the sulphur, but I gotta run… bye bye :simple_smile:
  139. StommePoes
    Feb 26 12:34
    haha
  140. bomf
  141. StommePoes
    Feb 26 12:35
    zomg I couldn't help but hear the Red Guy's voice with that
  142. Ooh hellOOOOOOO
  143. StommePoes
    Feb 26 12:37
    "The alt attribute is intended to be a replacement for what the image conveys. Capping it at any amount of text seems counter-productive. (In fact the IMG tag should have worked like OBJECT which can contain arbitrary content, including markup. But we can't fix that at this point.) But let's say you want a cap. Wouldn't 1000 words be the appropriate cap, given that it's common knowledge that a picture is worth 1000 words? 100 characters is way insufficient, even in Chinese."
  144. StommePoes
    Feb 26 12:37
    Yeah, it's useful for everyone to hear your life story on each image you feel is worth a thousand words >:(
  145. @StommePoes: Wait, she sounds reasonable
  146. You don't have to put your lifestory in it
  147. Otoh, isn't that what aria-describedby is for
  148. StommePoes
    Feb 26 12:39
    no
  149. StommePoes
    Feb 26 12:39
    If you need to describe your life story, it's time to consider visible text or a long description
  150. StommePoes
    Feb 26 12:39
    lord knows how old that draft is
  151. That's true, good point.
  152. But, is it our place to decide how many characters someone can use to describe an image? We can only advise to when it might be a bad idea to have lots of text in one.
  153. And that's what Tenon did, yes?
  154. jonathantneal
    Feb 26 12:40
    Good morning! (UGT)
  155. Heya @jonathantneal
  156. You're just in time for the 10 minutes of hate!
  157. jkva @jkva stretches
  158. MichielBijl
    Feb 26 12:41
    @StommePoes what about longdesc?
  159. MichielBijl
    Feb 26 12:41
    Let's make Apple implement that :P
  160. I remember @StommePoes ranting about longdesc at me, years ago. What was that about?
  161. StommePoes
    Feb 26 12:45
    No it was Mia.
  162. StommePoes
    Feb 26 12:45
    Apple has said unequivically that it won't support longdesc as it currently is
  163. StommePoes
    Feb 26 12:45
    They
  164. StommePoes
    Feb 26 12:45
    They'd rather devs used other things.
  165. StommePoes
    Feb 26 12:46
    But in any case, yeah if it's Ginormous Chart of Doom or something, it's good to consider if that alt text ought to be visible/article text and let the alt stick to "Providing short text alternative" as the WCAG says.
  166. StommePoes
    Feb 26 12:47
    I mean, I don't personally use longdesc either.
  167. StommePoes
    Feb 26 12:47
    I always use an actual link to content.
  168. StommePoes
    Feb 26 12:47
    But that's because current implementation, like Aria, is currently only offered to a select group of users, screen reader users.
  169. jonathantneal
    Feb 26 13:00
    I’m working with the author of okay-nav on a vanilla JavaScript version (it currently relies on jQuery), and I’d also like to make it accessible. https://github.com/VPenkov/okayNav
  170. jonathantneal
    Feb 26 13:01
    Here’s the version I started last night. Any feedback is appreciated. So far, I’d like to include a rule that if focus leaves the navigation while the overflow is open then it automatically closes. http://s.codepen.io/jonneal/debug/wGwjZN
  171. StommePoes
    Feb 26 13:03
    That's a good rule
  172. Ok. So I am interested in good UX. I read UX books. I make mockups. Try to think of the user's perspective for everything I build.
    I make sure I talk to the user. I advise from my technical perspective. In the end, I work with the user to get what the user needs.
  173. "Me: You might want X." "User : No I want Y".
    /me builds Y
    "User: No, make it like X"
  174. (╯°□°)╯︵ ┻━┻)
  175. Clearly, I should have advised more about why X would be better.
  176. But! "Don't assume you know what the user wants! You are not the user!"
  177. All I can do is advise, I guess.
  178. So, my revised internal ruleset

    1) You are not the user.
    2) The user knows only partially what they want, and expect you to work with them to get to a good end result.
    3) The user's ideas will change as their information grows.
    4) Embrace change.

  179. jonathantneal
    Feb 26 13:21
    Okay, I’ve added the rule to close the navigation overflow when the navigation loses focus. If anyone else has found any issues, please let me know. The current version is not accessible.
  180. MichielBijl
    Feb 26 13:25
    @jonathantneal I'll have a look when I'm home.
  181. MichielBijl @MichielBijl is having lunch with whisky.
  182. jonathantneal
    Feb 26 13:26
    @MichielBijl, well done.
  183. MichielBijl
    Feb 26 13:27
    Days off have their advantages.
  184. StommePoes
    Feb 26 13:29
    Could... I ask someone(s) with VO on Mac and iOS to do a quick test for me on figures?
  185. @StommePoes I have VO on Mac, but no iOS nearby
  186. StommePoes
    Feb 26 13:29
    I told @thierryk I'd post some current tests w3c/tr-design#61
  187. StommePoes
    Feb 26 13:30
    But I don't have mac/VO
  188. I can make a screen recording if you like
  189. StommePoes
    Feb 26 13:31
    I just need to know: are the images ignored? and if not, what do you get?
  190. StommePoes
    Feb 26 13:31
    Orca so far is the only one who acknowledges there are images
  191. StommePoes
    Feb 26 13:31
    both JAWS and NVDA are ignoring them on both IE and FF.
  192. @StommePoes Where do I need to look? fantasai's site?
  193. StommePoes
    Feb 26 13:32
    This is bad because we have a figure with no... figure. Just a seemingly random figcaption. It's good that we aren't getting some long-ass URL read out.
  194. StommePoes
    Feb 26 13:32
    not styled yet
  195. jkva @jkva turns of jethro tull
  196. hang on, lemme check
  197. MichielBijl
    Feb 26 13:35
    iOS reports alt I think (can't see source) for first example.
  198. StommePoes
    Feb 26 13:35
    there's no alt
  199. MichielBijl
    Feb 26 13:35
    But doesn't announce the figure group.
  200. StommePoes
    Feb 26 13:35
    these are
  201. MichielBijl
    Feb 26 13:35
    I'll check when I'm home.
  202. StommePoes
    Feb 26 13:35
    <figure><img src="blah"> <figcaption>blah blah blah</figcaption></figure>
  203. MichielBijl
    Feb 26 13:36
    Ah, then it's the source.
  204. StommePoes
    Feb 26 13:36
    I haven't gotten any figure anything's announced, but I didn't think anyone did that yet anyway
  205. MichielBijl
    Feb 26 13:36
    Something something.jpg
  206. StommePoes
    Feb 26 13:36
    You hear a source? aha
  207. StommePoes
    Feb 26 13:36
    interesting
  208. StommePoes
    Feb 26 13:36
    what happens when you deliberately go looknig for graphics?
  209. StommePoes
    Feb 26 13:36
    eg "g" or whatever VO's shortcut is?
  210. StommePoes
    Feb 26 13:36
    JAWS & NVDA both report no images found.
  211. MichielBijl
    Feb 26 13:36
    I think stevefaulkner said something about this couple days back.
  212. StommePoes
    Feb 26 13:37
    He had old results from tests he had done saying shit still sucked support-wise
  213. StommePoes
    Feb 26 13:37
    Thierry just wondered if anything was different
  214. VO on OSX reads entire alts to me
  215. exporting to mp4, hang on
  216. StommePoes
    Feb 26 13:37
    there are no alts
  217. figcaption
  218. StommePoes
    Feb 26 13:38
    yeah those are just text, they ought to be read out indeed.
  219. MichielBijl
    Feb 26 13:38
    With web rotor, it finds two images on iOS.
  220. StommePoes
    Feb 26 13:39
    Question is what the SRs /browsers are doing with the images, which do not have alt attributes and are in the special placement of being inside figures, with no other content inside, and combined with figcaptions.
  221. MichielBijl
    Feb 26 13:39
    Latest version mind you.
  222. StommePoes
    Feb 26 13:39
    interesting
  223. StommePoes
    Feb 26 13:39
    which is?
  224. StommePoes
    Feb 26 13:39
    I might as well write it down
  225. MichielBijl
    Feb 26 13:39
    9.2.something.
  226. StommePoes
    Feb 26 13:40
    on iOS?
  227. StommePoes
    Feb 26 13:40
    which
  228. MichielBijl
    Feb 26 13:40
    Alt=screenshot of about page shows 9.2.1 as version of os for iPhone. https://usercontent.irccloud-cdn.com/file/yAmgSgH3/IMG_0854.PNG
  229. StommePoes
    Feb 26 13:41
    which iOS version?
  230. StommePoes
    Feb 26 13:41
    11?
  231. StommePoes
    Feb 26 13:41
    also: thanks
  232. MichielBijl
    Feb 26 13:41
    9.2.1
  233. StommePoes
    Feb 26 13:42
    Wait... oh there's only a iOS version not a separate VO version Michiel?
  234. MichielBijl
    Feb 26 13:44
    Yes. Same on OS X.
  235. MichielBijl
    Feb 26 13:44
    They're tied to your OS. You cannot upgrade VO separately as far as I know.
  236. StommePoes
    Feb 26 13:45
    Ah ok. @jkva which version of MacOS do you have?
  237. Yosemite
  238. MichielBijl
    Feb 26 13:46
    10.10.something then.
  239. @MichielBijl 10.10.5 yes
  240. StommePoes
    Feb 26 13:46
    and so did I hear you also searching for images jkva?
  241. I did not searhc
  242. I just read the page
  243. StommePoes
    Feb 26 13:46
    Michiel found them using the web rotor
  244. MichielBijl
    Feb 26 13:47
    What if you do VO+U?
  245. StommePoes
    Feb 26 13:47
    what's u a shortcut for
  246. MichielBijl
    Feb 26 13:47
    Web rotor on Mac
  247. MichielBijl
    Feb 26 13:47
    Or item chooser thing.
  248. VO+u = links menu
  249. StommePoes
    Feb 26 13:47
    You should be able to cycle
  250. StommePoes
    Feb 26 13:47
    links, headings, etc
  251. MichielBijl
    Feb 26 13:48
    You can explore more if you arrow through it.
  252. MichielBijl
    Feb 26 13:48
    Yeap
  253. MichielBijl
    Feb 26 13:48
    You can adjust what's in it in settings.app
  254. MichielBijl
    Feb 26 13:48
    Not sure images are in there by default.
  255. Headings, links, forms
  256. Images not be default irt seems
  257. sophieschoice
    Feb 26 13:55
    @StommePoes Mac with El Capitan doesn’t announce the pictures
  258. sophieschoice
    Feb 26 13:56
    Funny :P a Dutch voice trying to speak English :P
  259. Other way around isn't much better :P
  260. sophieschoice
    Feb 26 13:56
    LOL :D
  261. StommePoes
    Feb 26 13:57
    Which version is el capitan, 11.x?
  262. They never get the "G" right...
  263. sophieschoice
    Feb 26 13:57
    10.11.3
  264. StommePoes
    Feb 26 13:58
    and if you actively search for images?
  265. sophieschoice
    Feb 26 13:58
    noob question: how do you do that?
  266. StommePoes
    Feb 26 13:58
    I used to have to try to test our dutch-language sites with English JAWS, which came with Finnish and 2 voices for Spanish but no Dutch...
  267. StommePoes
    Feb 26 13:58
    I dunno, I've never owned a Mac. Michiel knows though/.
  268. StommePoes
    Feb 26 14:05
    updated http://stommepoes.nl/work/figures.html if the info looks correct
  269. jkva @jkva adds "Get better at VO" to his TODO list
  270. #todo { overflow: scroll } :-/
  271. MichielBijl
    Feb 26 14:09
    Good luck ;)
  272. MichielBijl
    Feb 26 14:11
    For what it's worth, both images are shown on VO+Safari on OS X 10.11.3 El Capitain.
  273. StommePoes
    Feb 26 14:13
    Hm, now I can haz conflicts
  274. StommePoes
    Feb 26 14:14
    Does zee kepten say there are images or no? And if so, what does zee kepten say about them/
  275. StommePoes
    Feb 26 14:14
    ?
  276. MichielBijl
    Feb 26 14:15
    ^ that's what the captain says
  277. Those british shorthairs are way too soft for their own good
  278. MichielBijl @MichielBijl loves his shorthairs.
  279. sophieschoice
    Feb 26 14:18
    @MichielBijl What sorcery is that :P ? I go to @StommePoes ’s webpage in Safari and then I use ‘wijzig’ -> start spraakfunctie. That’s not the correct way?
  280. I start VO with cmd+f5
  281. @MichielBijl what's your screen recording software?
  282. StommePoes
    Feb 26 14:19
    oh the browser may matter here
  283. StommePoes
    Feb 26 14:19
    that is, safari may act different than FF possibly
  284. StommePoes
    Feb 26 14:20
    I'm assuming you're all using saffy/default browser
  285. MichielBijl
    Feb 26 14:20
    @sophieschoice, that's not VO, that's just the speech module of VO.
  286. MichielBijl
    Feb 26 14:20
    And browser can matter, yes :)
  287. MichielBijl
    Feb 26 14:20
    VO is started by cmd+F5 as Job points out.
  288. MichielBijl
    Feb 26 14:20
    And I use QuickTime for my screen recorder needs (ctrl+cmd+n)
  289. Ah, nice
  290. TIL
  291. StommePoes
    Feb 26 14:23
    ok, wonder if I can send this to thierry
  292. StommePoes
    Feb 26 14:23
    so we're not sure if the iOS version gives the filenames as well
  293. StommePoes
    Feb 26 14:23
    though... it probably does
  294. StommePoes
    Feb 26 14:23
    if these two share code
  295. powrsurg
    Feb 26 14:24
    @MichielBijl (late reply I know), but that was a developer module that prints out a table of all database queries. It's actually a data table that should be a table. But for supposed performance reasons they made it all div based
  296. MichielBijl
    Feb 26 14:25
    Aha :/
  297. MichielBijl
    Feb 26 14:25
    @StommePoes, I can record that too, hang on.
  298. powrsurg
    Feb 26 14:25
    I guess that's a good 180 from where we were about 10-15 years ago where everything was table-based when they shouldn't
  299. sophieschoice
    Feb 26 14:27
    aha, so I got the real Voice-Over now. thanks @MichielBijl .
  300. sophieschoice
    Feb 26 14:27
    @StommePoes I can confirm that VoiceOver indeed says it as in the video of Michiel.
  301. StommePoes
    Feb 26 14:28
    <MrBurns>Eeeeexcellent</MrBurns>
  302. StommePoes
    Feb 26 14:28
    powrsurg Now we have Drupal with divitis
  303. sophieschoice
    Feb 26 14:28
    it was just me being a noob :P
  304. MichielBijl
    Feb 26 14:33
    If someone wants to help with the subs, that'd be great.
  305. StommePoes
    Feb 26 14:38
    "grey_poes.jpg" and "text-flow-vectors-tb.png"
  306. StommePoes
    Feb 26 14:39
    I noticed once when you went back it said image not found, but then continued to find them.
  307. StommePoes
    Feb 26 14:39
    so, page updated and modified, sent to thierry
  308. StommePoes
    Feb 26 14:39
    Thanks all
  309. MichielBijl
    Feb 26 14:40
    Now with subs
  310. MichielBijl
    Feb 26 14:40
    Yeah, that's if you try to navigate beyond the first or last image.
  311. zakim-robot
    Feb 26 15:34
    [ccwilcox] Following up on the problem I was seeing in Firefox yesterday where the target of a "Skip to" link does not get focus, it seems to be a long-standing bug. I think this is the most relevant Bugzilla entry, with the most recent comment from @rodneyrehm https://bugzilla.mozilla.org/show_bug.cgi?id=308064
  312. MichielBijl
    Feb 26 15:50
    Hmm, we need more TPG people in here, we only have Steve, Léonie, and Karl…
  313. MichielBijl
    Feb 26 15:51
    Guess this is as good a time as any to announce that I have joined the Paciello Group then ;)
  314. zakim-robot
    Feb 26 15:53
    [deborah_kaplan] well cool!
  315. Slowly the Dutch influence grows
  316. sophieschoice
    Feb 26 15:55
    cool :D
  317. MichielBijl
    Feb 26 15:56
    I'll be moving into Hans' clog soon ;)
  318. stevefaulkner
    Feb 26 15:57
    that's one huge clog to fill
  319. MichielBijl
    Feb 26 15:57
    Yeah, but I need room for my cats, so that's okay.
  320. Wear thick socks
  321. MichielBijl
    Feb 26 15:57
    ^ and that
  322. MichielBijl
    Feb 26 15:57
    I'll just have to eat a bit more.
  323. You don't want to clog up the clog though
  324. :sparkles:
  325. zakim-robot
    Feb 26 16:02
    [hmig] @stevef: Thanks for the link! The best way, that I found, for the svg icon (where <use> is utilized), is to add <title id=“whatevs”> and aria-labelledby to the svg ...
  326. zakim-robot
    Feb 26 16:02
    [hmig] Oh, and role=“img” to the svg
  327. zakim-robot
    Feb 26 16:03
    [hmig] to me, it’s better than hiding the svg and adding a span with offscreen text, but def not ideal
  328. StommePoes
    Feb 26 16:04
    I still use the offscreen text bit for the robustness, but yeah would be nice if the defaults worked as advertised. Which I assume eventually they will. If we whine hard enough.
  329. StommePoes
    Feb 26 16:04
    I'm good at whining, but apparently not at the right people.
  330. :neutral_face:
  331. "Michiel Bijl? Oh yeah, yeah I knew him before he worked at TPG. Those were different times."
  332. StommePoes
    Feb 26 16:07
    It means Michiel is now going to end up in those weird films they make
  333. StommePoes
    Feb 26 16:08
    and those photoshopped toilet-related photos we keep seeing on Twitter
  334. TPG makes weird films?
  335. StommePoes
    Feb 26 16:08
    the members do
  336. zakim-robot
    Feb 26 16:08
    [hmig] hi @StommePoes! :simple_smile:
  337. StommePoes
    Feb 26 16:08
    You saw at the role=drinks-- @ljWatson feeling a wall for braille, comes across @stevefaulkner's chest, exclaims it says "C cup"
  338. Ah right =D
  339. StommePoes
    Feb 26 16:09
    Hi @hmig!
  340. StommePoes
    Feb 26 16:09
    ANOTHER HEATHER
  341. StommePoes
    Feb 26 16:09
    Heather everywhere but no Timothy or Myrtle
  342. Hmm, Myrtle
  343. "Myrica", not a bad name
  344. "Yeah we named our daughter Myrica Gale"
    "That's beautiful"
    "Yeah it's Bog Myrtle"
  345. StommePoes
    Feb 26 16:11
    Myrtle is kind of an old-lady name now, as it was popular decades ago. But a plant's a plant.
  346. StommePoes
    Feb 26 16:11
    Name all your kids after plants.
  347. "Hemlock, suppertime!" =)
  348. Bog Myrtle is great beer flavouring though
  349. mmmm
  350. Makes the best tripel
  351. zakim-robot
    Feb 26 16:24
    [hmig] @StommePoes - what are your thoughts on adding mobile tests to the a11y svg testing?
  352. zakim-robot
    Feb 26 16:25
    [hmig] I’m thinking we def should, but not sure of the array to cover
  353. zakim-robot
    Feb 26 16:25
    [hmig] yikes … brb …. appt
  354. StommePoes
    Feb 26 16:29
    Well, from the Twitters I see that Jitendra has been pestering Sara Souidan about Opera Mini doing this, that and the other wrong or weird with SVGs, so actually yeah, it does sound like a good idea.
  355. StommePoes
    Feb 26 16:30
    Actually I have some kind of Opera on my Windows phone, but I'm not sure which I added.
  356. StommePoes
    Feb 26 16:31
    It says Opera 9.0.0.211
  357. StommePoes
    Feb 26 16:31
    but I still don't know if that's mini or mobile.
  358. StommePoes
    Feb 26 16:31
    ah Mini.
  359. StommePoes
    Feb 26 16:31
    Anyway if we add those tests Jitendra would also be good for testing them, as it seems he's already busy with that
  360. MichielBijl
    Feb 26 16:32
    Is Sara at Opera?!
  361. zakim-robot
    Feb 26 16:45
    [ccwilcox] @StommePoes on Windows phone, it's Opera Mini. http://www.opera.com/mobile
  362. zakim-robot
    Feb 26 16:53
    [marcysutton] @michiel: Sara is a freelancer, and the world’s resident SVG expert :smile:
  363. StommePoes
    Feb 26 17:07
    Michiel not that I know of, but she does the SVG things on mobile
  364. StommePoes
    Feb 26 17:08
    But Jitendra uses Opera Mini so he's been asking her about weird stuff he's running into sometimes with SVGs
  365. StommePoes
    Feb 26 17:08
    like the lazy-loading SVGs, the issue he found was that Opera Mini doens't have a scroll event. The lazyloads must have been listening for that to trigger them, meaning nada for Mini users.
  366. StommePoes
    Feb 26 17:09
    I guess an easy fix there is first check for that event existing, and if not, dumb all the things on the browser at once. Esp mini will deal with that anyway. (or the Windows phone, which also limits donwloads)
  367. StommePoes
    Feb 26 17:11
    Sometimes they're scaling weird.
  368. MichielBijl
    Feb 26 17:30
    @Marcy Sutton, yeah that's what I thought, so I didn't get StommePoes' remarks as to why Jitendra would be pestering Sara with Opera stuff.
  369. MichielBijl
    Feb 26 17:31
    Ah, but thanks for the explanation Mall
  370. MichielBijl
    Feb 26 17:36
    I just bought some more Lego to celebrate :D
  371. sophieschoice
    Feb 26 17:58
    which set :D?
  372. StommePoes
    Feb 26 18:02
    dino's
  373. jpdevries
    Feb 26 18:16
    Speaking of Opera, if there is anyone here from Opera, who can I send cupcakes too for whoever majorly fixed flexbox in Opera 35? One of my flexbox layouts when from totally borked to absolutely perfect without changing any code!
  374. MichielBijl
    Feb 26 18:21
    @sophieschoice 60117 and 60081
  375. MichielBijl
    Feb 26 18:22
    And got 31039 yesterday 🙈
  376. sophieschoice
    Feb 26 18:23
    cool :D
  377. sophieschoice
    Feb 26 18:25
    I got back in Lego last year (blushes) and bought two modulars.
  378. MichielBijl
    Feb 26 18:25
    I love lego, I buy it whenever I get nervous
  379. MichielBijl
    Feb 26 18:27
    This is pretty much me (in the role of Denmark): http://satwcomic.com/so-close-yet-so-far
  380. sophieschoice
    Feb 26 18:27
    :D:D
  381. MichielBijl
    Feb 26 18:29
    Description for that image: Denmark makes a remark toward Norway that he really likes “Let the Right One In” and states that the Norwegians are really good at horror. Sweden says to Denmark that that is a Swedish film. Norway asks Denmark if he can tell Norwegian and Swedish apart. Denmark panics and throws lego bricks in the air while yelling “POCKET
  382. MichielBijl
    Feb 26 18:29
    LEGO”. Sweden and Norway are left wondering whether Denmark always carries lego around.
  383. MichielBijl
    Feb 26 18:29
    Now that's a good example of a longdesc :P
  384. MichielBijl
    Feb 26 18:30
    Or rather, something you might find in a details/summary
  385. sophieschoice
    Feb 26 18:31
    I still need to reconcile myself a bit with the fact I got back into Lego however (ie: need starting to create/build with things with it more)....
  386. MichielBijl
    Feb 26 18:32
    Yeah I've been building sets, but need to collect a bit more before I can tear it all down and build something else :P
  387. MichielBijl
    Feb 26 18:32
    Which is probably bullshit, but it gives me a good excuse to collect more Lego…
  388. jpdevries
    Feb 26 18:36
    My Grandfather is suffering from dementia. I’d like to take this as a opportunity to research and explore assistive tech in this area. a11y has been a fascination of mine for some time, but this is a fairly new arena for me. Any resources or tips would be greatly appreciated.
  389. jnurthen
    Feb 26 18:38
    @jpdevries Have you looked at the work that the cognitive TF are doing - https://www.w3.org/WAI/PF/cognitive-a11y-tf/ . Someone who is a member there would probably know what is available.
  390. MichielBijl
    Feb 26 18:40
    (think there is a book that accompanies that)
  391. jnurthen
    Feb 26 18:41
    @jpdevries specifically this section of the current draft - https://w3c.github.io/wcag/coga/gap-analysis.html#aging-and-dementia
  392. jnurthen
    Feb 26 18:41
    It is mostly a gap analysis at the moment but it may help.
  393. sophieschoice
    Feb 26 19:20
    if you implement structured data tags from schema.org, does that affect assistive technology in a bad way?
  394. MichielBijl
    Feb 26 19:21
    That's just classes etc right?
  395. sophieschoice
    Feb 26 19:21
    yeah, like <ol itemscope itemtype="http://schema.org/BreadcrumbList">
  396. StommePoes
    Feb 26 19:24
    Mostly they're invisible to AT
  397. StommePoes
    Feb 26 19:25
    The old Google breadcrumbs were weird, I can't remember, but now the schema breadcrumbs are okay with Google so they're normal lists as well.
  398. sophieschoice
    Feb 26 19:29
    ah OK
  399. sophieschoice
    Feb 26 19:29
    but 'mostly they're invisible' doesn't sound to good :P
  400. jpdevries
    Feb 26 19:41
    @jnurthen @MichielBijl Thanks for sharing! I’ll take a look
  401. StommePoes
    Feb 26 20:28
    schema.org is a google thing
  402. StommePoes
    Feb 26 20:28
    for google, by google
  403. StommePoes
    Feb 26 20:28
    it's not made to affect AT
  404. StommePoes
    Feb 26 20:28
    it's so google bots can scrape stuff people search for
  405. StommePoes
    Feb 26 20:28
    Makes sense AT would ignore most of it. I can see a few things being useful
  406. StommePoes
    Feb 26 20:28
    but a lot of it would be clutter.
  407. StommePoes
    Feb 26 20:28
    We used schema on some of our webshops
  408. StommePoes
    Feb 26 20:29
    damn near doubled the code with all the meta tags
  409. zakim-robot
    Feb 26 20:29
    [hmig] does that affect performance atal?
  410. zakim-robot
    Feb 26 20:29
    [hmig] at all
  411. sophieschoice
    Feb 26 20:31
    well, the leaner your document and assets are, the better :)
  412. sophieschoice
    Feb 26 20:32
    especially for mobile and crappy wifi/internet connections
  413. MichielBijl @MichielBijl never works with schema
  414. MichielBijl
    Feb 26 20:37
    Haven't read up on it, but this seems weird: w3c/html#102
  415. MichielBijl
    Feb 26 20:37
    Oh never mind…
  416. StommePoes
    Feb 26 20:40
    Since it was pretty much all markup, rather than scripts or images, it wasn't necessarily a huge bloat weight wise. Plus, it technically counts as content.
  417. MichielBijl
    Feb 26 20:40
    That initial title isn't as clear as it could be.
  418. StommePoes
    Feb 26 20:41
    Ha!
  419. StommePoes
    Feb 26 20:41
    I wish the CSS squirrel guy would make a comic of this
  420. StommePoes
    Feb 26 20:42
    Chaals vs the Hixie
  421. Squeee!
  422. MichielBijl
    Feb 26 21:46
    We should have like the celebrity wrestling/boxing thing with clay but with web people
  423. garcialo
    Feb 26 21:46
    @MichielBijl Celebrity Deathmatch?
  424. MichielBijl
    Feb 26 21:47
    Yeap, that's the one
  425. MichielBijl
    Feb 26 21:47
    You and Karl can do the commentary.
  426. garcialo
    Feb 26 21:48
    I think it would be more appropriate for it to be Billy and Karl.
  427. StommePoes
    Feb 26 21:54
    Billy and Karl need to be IN the deathmatch
  428. StommePoes
    Feb 26 21:54
    I wonder, does Kyle still do the occasional squirrel?
  429. MichielBijl
    Feb 26 21:54
    That too
  430. MichielBijl
    Feb 26 21:55
    Hell, let them comment their own fight
  431. stevefaulkner
    Feb 26 22:12
    karl and billy - cage fight https://youtu.be/nhGfh0W8t80
  432. StommePoes
    Feb 26 22:18
    haha
  433. StommePoes
    Feb 26 22:18
    Well, I'm glad I'm not the only one put off by "that's stupid!" comments in Github. I notice it makes my comments also go down to my natural level as well, hard to fight.
  434. stevefaulkner
    Feb 26 22:31
    @StommePoes which stupid GitHub comments?
  435. MichielBijl
    Feb 26 22:36
    Ha, gay.
  436. MichielBijl
    Feb 26 22:36
    They should do that for real.
  437. MichielBijl
    Feb 26 22:39
    What is the best way to markup a dialog?
  438. MichielBijl
    Feb 26 22:39
    Between two people that is.
  439. MichielBijl
    Feb 26 22:39
    So say a chat log or something
  440. StommePoes
    Feb 26 22:39
    @stevefaulkner not the ones where the chromium bug was called stupid (zomg I loved that), but the ones where a testing tool made to point out stuff usefully to developers is called stupid.
    Though I guess the biggest gripe was just the doom-like ubervillain names of the errors.
  441. StommePoes
    Feb 26 22:40
    I do p's, actually. Headings can be better but only if it's short
  442. StommePoes
    Feb 26 22:40
    I used to do dl's, but to me those became key-value pairs many years ago.
  443. MichielBijl
    Feb 26 22:40
    Yeah I was thinking dl's
  444. StommePoes
    Feb 26 22:40
    I just do <p><span to bold maybe>Name: </span>blah blah...</p>
  445. MichielBijl
    Feb 26 22:40
    I think that's the example in the html spec looks at stevefaulkner
  446. StommePoes
    Feb 26 22:41
    You could do dl's.
  447. StommePoes
    Feb 26 22:41
    It was back when dialog was being discussed
  448. MichielBijl
    Feb 26 22:41
    I'll make something up
  449. StommePoes
    Feb 26 22:41
    At some point the spec said "for dialogs" but I dunno if that's still a thing, that was a long time ago.
  450. StommePoes
    Feb 26 22:41
    Michiel: that's always awesome
  451. MichielBijl
    Feb 26 22:42
  452. stevefaulkner
    Feb 26 22:44
    Who was calling tenon names?
  453. StommePoes
    Feb 26 22:48
    w3c/tr-design#61
  454. MichielBijl
    Feb 26 22:48
  455. StommePoes
    Feb 26 22:48
    Michiel I think the orginal dialoggy example for dl got removed after some point.
  456. MichielBijl
    Feb 26 22:49
    “Wouldn't 1000 words be the appropriate cap, given that it's common knowledge that a picture is worth 1000 words? 100 characters is way insufficient, even in Chinese.” woah
  457. StommePoes
    Feb 26 22:49
    Yeah that one kinda pissed me off.
  458. MichielBijl
    Feb 26 22:49
    What happened to using longdesc :P
  459. StommePoes
    Feb 26 22:49
    However I did find an earlier back and forth between steve and a Vlad and I kinda agreed with Vlad at least on the practicality
  460. MichielBijl
    Feb 26 22:50
    I have respect for fantasai, but she has some quirks (don't we all)
  461. StommePoes
    Feb 26 22:50
    However I see Tenon's 100-char limit as, not a "critical error astronauts will die" but definitely "human! Look at this!"
  462. StommePoes
    Feb 26 22:50
    Sometimes alt might really have to be long, BUT ideally if it's horribly long, it's a sign you should try to move it elsewhere.
  463. StommePoes
    Feb 26 22:51
    Michiel about the dl, now that I think more about it... do you see a conversation as a list?
  464. StommePoes
    Feb 26 22:51
    I mean, a paragraph is not a list of words, and I personally don't feel a form is a list of questions (but a sequence)... it might similarly be a weird idea that an interview is a list. Might only be a sequence.
  465. StommePoes
    Feb 26 22:52
    Otherwise, at least dt and dd gives you more styling hooks. Visually and textually it needs to be obvious who said what.
  466. MichielBijl
    Feb 26 22:56
    Could be a list of quotes…
  467. MichielBijl
    Feb 26 22:57
    Headings+p's sounds horrible (litter web rotors etc)
  468. MichielBijl
    Feb 26 22:57
    Maybe list of p's is the best way to go…
  469. MichielBijl
    Feb 26 22:57
    But again, that's why I asked :P
  470. StommePoes
    Feb 26 23:05
    In the past, I've just done p's.
  471. StommePoes
    Feb 26 23:05
    If you can group subjects inside the interview though, you can intersperse it with headings.
  472. jnurthen
    Feb 26 23:26
    I have real issues with Tenon reporting things as Errors when they are only potential errors
  473. MichielBijl
    Feb 26 23:28
    Send them to talktous@tenon.io
  474. jnurthen
    Feb 26 23:28
    Every tool that does that results in our customers running the tool against out products and sending us an error report (book) . Someone then has to go through the false-positive errors which takes away from engineering time
  475. jnurthen
    Feb 26 23:28
    Karl knows my views on that
  476. zakim-robot
    Feb 26 23:28

    [karlgroves] >Send them to talktous@tenon.io

    +100.

    My phone number and email address is on literally every page of the site

  477. jnurthen
    Feb 26 23:29
    I haven't had error reports from tenon yet but it is only a matter of time
  478. zakim-robot
    Feb 26 23:29
    [karlgroves] Then that would seem counter to this statement: “
    jnurthen BOT [6:26 PM]
    I have real issues with Tenon reporting things as Errors when they are only potential errors"
  479. StommePoes
    Feb 26 23:29
    I do think the stuff that internally is based on a sort of thumb-rule could change what they're called
  480. StommePoes
    Feb 26 23:30
    like "super crazy long alt text: this might be excessively horrid. Recheck it and see if putting that amount of text elsewhere is less irritating" or something.
  481. StommePoes
    Feb 26 23:30
    Leave the <a href="foo"><img src="bar"></a> for the astronauts-are-going-to-die critical errors
  482. StommePoes
    Feb 26 23:31
    And this is all beside the developer being able to set critical and priority levels
  483. StommePoes
    Feb 26 23:31
    I mean even when you set things, the stuff that's known-subjective could get different text.
  484. jnurthen
    Feb 26 23:31
    @karlgroves I have an issue with ANY tool reporting false-positives not just tenon. It results in these 100+ page conformance reports we get which have the same false-positive errors on every page. You end up drowning out the real errors with the chaff
  485. StommePoes
    Feb 26 23:31
    Some errors are like, always absolutely errors-- validation and linting stuff are pretty much error-errors. Others...
  486. StommePoes
    Feb 26 23:32
    @junurthen although you can change your settings for that...
  487. StommePoes
    Feb 26 23:32
    what settings are you using?
  488. StommePoes
    Feb 26 23:32
    or are they coming through no matter what?
  489. jnurthen
    Feb 26 23:32
    @StommePoes I don't use tenon
  490. StommePoes
    Feb 26 23:32
    Oh, well Tenon lets you decide how much astronaut-killing level you want
  491. jnurthen
    Feb 26 23:32
    I saw your report on the CSS thing which had things labelled as error when they are not
  492. StommePoes
    Feb 26 23:32
    That was as anon account
  493. StommePoes
    Feb 26 23:32
    so I could sent it as a public link
  494. StommePoes
    Feb 26 23:32
    however
  495. StommePoes
    Feb 26 23:33
    When you're logged in and are using your key (as you would as a using developer)
  496. StommePoes
    Feb 26 23:33
    you will set levels of how much whining you want
  497. StommePoes
    Feb 26 23:33
    so for example if you're using Tenon as a check on a build system
  498. StommePoes
    Feb 26 23:33
    you set your levels to 100% certainty
  499. StommePoes
    Feb 26 23:33
    which leaves out a lot of the "astronauts may die: alt text kinda long"
  500. StommePoes
    Feb 26 23:33
    but holds to teh "link has no text" stuff
  501. MichielBijl
    Feb 26 23:33
    Ohh Karl, can has Tenon account?
  502. MichielBijl
    Feb 26 23:33
    TEST ALL THE THINGS!
  503. StommePoes
    Feb 26 23:34
    You don't want a build to fail on maybes, but the stuff even a machine can't get wrong
  504. jnurthen
    Feb 26 23:34
    so " Error / priority 100%
    Preformatted text found" would be picked up?
  505. StommePoes
    Feb 26 23:34
    Michiel: BUY ONE
  506. zakim-robot
    Feb 26 23:34

    [karlgroves] > I don't use tenon

    I saw your report on the CSS thing which had things labelled as error when they are not

    Did you read my response? Can you offer a different argument or evidence where the tool is wrong? Say, the 100 character thing. Do you think its too short? Let’s talk about that.

    Right now - I mean literally as we speak - I’m working on improving Tenon’s tests based on feedback from Alastair Duggin’s tests of various tools.

  507. MichielBijl
    Feb 26 23:34
    @StommePoes, you should probably join #role-drinks on freenode
  508. StommePoes
    Feb 26 23:34
    I need to get back to restarting pidgin... is there a logger somewhere?
  509. StommePoes
    Feb 26 23:34
    That;s the only thign I miss with irc/pidgin
  510. MichielBijl
    Feb 26 23:34
    irccloud.com
  511. StommePoes
    Feb 26 23:35
    Hm, maybe
  512. StommePoes
    Feb 26 23:35
    Esp if I join a w3c group
  513. MichielBijl
    Feb 26 23:35
    Comes in very handy with w3c stuff
  514. zakim-robot
    Feb 26 23:35

    [karlgroves] I mean, honestly @jnurthen - do you think for a second that any other tool vendor gives a shit about this sort of feedback as I do?

    Don’t like something? Tell me. Again, my own personal mobile number is on every page of the site.

  515. StommePoes
    Feb 26 23:36
    Karl: I will give you this: for the anon-account tests, like the one I sent whats-her-name, I do kinda agree with her that it sounds false-positivey for the stuff where the machine wants to flag something (because a human SHOULD look at it) but can't tell if it's an error. Like the pre thing and the alt thing.
  516. StommePoes
    Feb 26 23:36
    And that's a pretty easy change, I would think.
  517. StommePoes
    Feb 26 23:36
    Maybe it's because we can send these anon reports to people who haven't seen a webinaar and don't know Tenon's idea's of "critical"
  518. StommePoes
    Feb 26 23:36
    critical sounds like your spaceship is running out of air kinda thing
  519. StommePoes
    Feb 26 23:36
    in normal human speak.
  520. jnurthen
    Feb 26 23:36
    @karlgroves The problem is I can't use tenon for any work stuff. The only way I can test it is to use it in my own time. As I said I haven't come across a customer report using it yet. Once I do I will be able to give more input.
  521. StommePoes
    Feb 26 23:36
    And also in dev speak, I'd say.
  522. MichielBijl
    Feb 26 23:37
    Why can't you use it for work stuff James?
  523. StommePoes
    Feb 26 23:37
    @jnurthen I'm sorta using it now (not building enough components at the moment) and offering my key to our 1 other developer, but I think only next month will we start building real things again
  524. jnurthen
    Feb 26 23:37
    although unless you have solved the logon issue it is unlikely to actually be able to test anything
  525. zakim-robot
    Feb 26 23:37
    [karlgroves] Yeah, the <pre> thing is on my list to improve. The alt thing I want more feedback on. Changing it is easy, but I don’t want to say “OK, 200 is the new limit.”
  526. StommePoes
    Feb 26 23:37
    No Karl my point isn't really about tweaking the test
  527. jnurthen
    Feb 26 23:38
    @MichielBijl not allowed to submit code to 3rd party servers
  528. zakim-robot
    Feb 26 23:38

    [karlgroves] > although unless you have solved the logon issue it is unlikely to actually be able to test anything

    Yes, buy an on-prem license

  529. StommePoes
    Feb 26 23:38
    I think 100 is a fine place to tell people "take another look"
  530. MichielBijl
    Feb 26 23:38
    Ah, makes sense.
  531. StommePoes
    Feb 26 23:38
    It's how it's named that bugs people.
  532. jnurthen
    Feb 26 23:38
    @karlgroves not likely I can do that
  533. zakim-robot
    Feb 26 23:38
    [karlgroves] Not likely you can use a reverse tunnel either though, is it?
  534. StommePoes
    Feb 26 23:38
    The alt thing can't be called a "critical error" when a human may very well determine that a particular image really does need (and is ok with) a 140 word alt. It should be called... alert or warning or somthing.
  535. zakim-robot
    Feb 26 23:39
    [karlgroves] We don’t call anything a critical error
  536. jnurthen
    Feb 26 23:39
    @karlgroves that is against security policy
  537. zakim-robot
    Feb 26 23:39
    [karlgroves] I figured.
  538. StommePoes
    Feb 26 23:39
    Karl is it possible to add login creds as params and send, or is that not an option because people don't want their creds plaintext in URLs?
  539. zakim-robot
    Feb 26 23:41

    [karlgroves] Well, everything is done over HTTPS (TSL) but the thing with logins is that it is an insane target to hit. I know that it has (and continues to be) a thorn in the side of all other tool vendors, no matter waht they want to say publicly. Realm authentication and things like that are easy. Passing a cookie is easy. But then lets add in other sensible security restrictions, like measures to thwart session hijacking and the like and it gets to be a huge PITA

    I recommend people use Selenium for that stuff. It kicks the shit out of anything any tool can do

  540. StommePoes
    Feb 26 23:44
    @karlgroves hm, that's true. (not critical error)
  541. jnurthen
    Feb 26 23:44
    but Error is enough to be "critical" to QA teams
  542. jnurthen
    Feb 26 23:44
    they will log a bug for it
  543. StommePoes
    Feb 26 23:44
    yes.
  544. jnurthen
    Feb 26 23:44
    which is noise
  545. StommePoes
    Feb 26 23:44
    Again, with an account you can set your noise tolerance
  546. jnurthen
    Feb 26 23:44
    There is enough real noise without having fake noise ;)
  547. StommePoes
    Feb 26 23:45
    if you have a bunch of newbies, you'd want them to hear All The Things. People who know stuff and just want typos caught etc, set it to 100% certainty.
  548. jnurthen
    Feb 26 23:46
    no. I want newbies to have 100% certaintly. With experience you can discount the other errors easily but having them reported to new folks makes things worse
  549. zakim-robot
    Feb 26 23:47

    [karlgroves] >but Error is enough to be "critical" to QA teams
    Then the QA teams need training before they use a tool they do not understand. You don’t give someone a welder and expect them to build a bridge, do you?

    The UI says anything with certainty > 80% is an error. It means we’re 100% positive that the test has failed its criteria.
    The priority score should also be used to determine how to treat the issue reported. priority takes into account a number of factors, but we’ve realized that we need to separate out severity from the priority calc

  550. jnurthen
    Feb 26 23:48
    the PRE one disputes this. It is 100% in the UI I saw.
  551. StommePoes
    Feb 26 23:49
    "no. I want newbies to have 100% certaintly." Oh, we don't. I want everyone who's never heard of alt text to come to me with each of those errors so I know what to train them on, so they can learn, so they can be better devs. and so they can do some a11y checking before it comes to me.
  552. StommePoes
    Feb 26 23:49
    jnurthen there's also a bug for that one
  553. zakim-robot
    Feb 26 23:49
    [karlgroves] We’re 100% confident that the test is a failure. However as I’ve said, I want that test to be more robust
  554. jnurthen
    Feb 26 23:49
    fair enought. i understand bugs
  555. zakim-robot
    Feb 26 23:50
    [karlgroves] I currently have 10 important test bugs I hope to address - hopefully before CSUN
  556. jnurthen
    Feb 26 23:50
    good to hear
  557. zakim-robot
    Feb 26 23:50
    [karlgroves] Color contrast is the focus this weekend. Its long overdue
  558. jnurthen
    Feb 26 23:50
    are you going to do background images and gradients
  559. MichielBijl
    Feb 26 23:51
    Oeps
  560. MichielBijl
    Feb 26 23:51
    Wrong window
  561. StommePoes
    Feb 26 23:51
    That would need something like some of the screenshotting tools like in selenium
  562. StommePoes
    Feb 26 23:51
    makes sense to use those
  563. zakim-robot
    Feb 26 23:51
    [karlgroves] Actually, that’s what’s been taking so long. I’ve been experimenting with that. That type of test isn’t ready to go yet, but that’s my target
  564. StommePoes
    Feb 26 23:51
    for background images and gradients
  565. jnurthen
    Feb 26 23:51
    yeah - i have that working in my FF plugin
  566. jnurthen
    Feb 26 23:52
    it seems everything has a background image or gradient these days so without it things are impossible
  567. StommePoes
    Feb 26 23:52
    That tool from North carolina that does contrast is an extremely slow screenshotting tool, takes like half a day to take a 1 page screenshot and then measures pixels :P
  568. jnurthen
    Feb 26 23:52
    heh - mine is better :)
  569. garcialo
    Feb 26 23:52
    Now don't you get people all excited!
  570. StommePoes
    Feb 26 23:52
    I just don't bother wht the tools anymore. I jsut measure CSS values and throw them into Lea Verou's thing
  571. jnurthen
    Feb 26 23:53
    picking up colours from backgrounds also allows you to test text in SVG which is cool :)
  572. jnurthen
    Feb 26 23:53
    @garcialo I've teased you with that before right ;)
  573. garcialo
    Feb 26 23:54
    Can't wait for you to share it!
  574. zakim-robot
    Feb 26 23:54
    [karlgroves] The biggest challenge making a testing tool is dealing with the crazy stuff people do on their sites.
  575. jnurthen
    Feb 26 23:54
    heh I hear that
  576. jnurthen
    Feb 26 23:55
    I have 1000+ products my tool has to work with written with a myriad of frameworks
  577. jpdevries
    Feb 26 23:56
    @MichielBijl thanks again for the alzheimers video. The part about multiple choice questions was particularly helpful.
  578. jnurthen
    Feb 26 23:56
    thanks to 100+ acquisitions in the last x years (not including the companies those companies had acquired)
  579. jnurthen
    Feb 26 23:57
    one of the products used to have frames nested 6 levels deep before you actually got to the application.
  580. jnurthen
    Feb 26 23:57
    its amazing how many testing tools don't deal with frames at all
  581. zakim-robot
    Feb 26 23:58

    [karlgroves] I had a guy email recently saying “TENON CAN’T TEST MY SITE!!!” and it was giving him 522 responses.

    Come to find out they were loading like 70 huge images. Turns out it took 44 seconds on my FiOS connection to load the page.

  582. jnurthen
    Feb 26 23:59
    I was going to say that tenon probably doesn't need to load all the images but then they can impact the layout, and colour contrast and could even be positioned over text etc. so it could need them after all