Archive index

A11y Slackers Gitter Channel Archive 3rd of October 2017

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    @zakim-robot
    Oct 03 01:11

    [eric] I really like the concepts of modularity/extensibility being able to write back to the browser, but these aspects are also a cause of concern for me. I’ve seen enough “in the trenches” dev that I worry people will be slathering things with ARIA and then advertising it as accessible without testing actual support (especially considering ARIA isn’t fully or consistently supported in all browsers).

    Add the force multiplier of modularization into the mix, and you have a very real concern of a sexy-looking landing page enticing developers to adopt and potentially never update if issues are uncovered (see that 5k starred drag and drop lib posted earlier using the term “accessible” as a very timely example).

    The native semantics tied to W3C specs give you a whole lot for free, especially if your browser isn’t one of the big four. I have another concern around that, especially where emerging markets/lower powered devices are concerned.

  2. zakim-robot
    @zakim-robot
    Oct 03 01:46

    [svinkle] Doesn't sound much different than issues seen today. Native, frameworks, aria; always lots of human issues surrounding implementation. There's always going to be human error, but I think this has potential to eliviate some of those issues.

    If done well, the abstraction the comes with using a well written, accessible component will help in the long run.

  3. zakim-robot
    @zakim-robot
    Oct 03 02:16
    [marcysutton] Having tools that make it easier to get it right than they do today are good, in my opinion.
  4. zakim-robot
    @zakim-robot
    Oct 03 11:19
    [jv] If I have table in design where last row show total amount
  5. [jv] What should be the code of last row?
  6. zakim-robot
    @zakim-robot
    Oct 03 11:48
    [tommyfeldt] Is anyone familiar with coding native iOS forms? I’m trying to figure out if it’s possible to mark an input field as required, but can’t find anything
  7. zakim-robot
    @zakim-robot
    Oct 03 13:44
    [kivi] @jv I would use the <tfoot> tag: see https://www.w3schools.com/TagS/tag_tfoot.asp
  8. [jv] @kivi doesn't it mean less important? while total amount is most important
  9. zakim-robot
    @zakim-robot
    Oct 03 13:55
    [kivi] @jv it has to do with semantic structure, not importance. <tfoot> is perfect for totals. Here's another reference: https://html.com/tables/
  10. [kivi] @tommyfeldt some interesting ideas at https://stackoverflow.com/questions/40486047/form-validation-in-objective-c
  11. zakim-robot
    @zakim-robot
    Oct 03 14:05
    [joao.garin] Hello all, does anyone have some info on font size switcher vs zoom approach for accesibility? should both be used? are they both relevant? some read information on this would be highly appreciated tks!
  12. zakim-robot
    @zakim-robot
    Oct 03 14:15
    [svinkle] @jv @kivi <tfoot> thumbsup emoji
  13. zakim-robot
    @zakim-robot
    Oct 03 15:16
    [vavroom] @joao.garin I personally feel like font size switcher are generally overkill, except perhaps where sites have a target visitor of elderly people. Make sure the site's text resizes well, without breaking layout. Think text resize as opposed to browser zoom.
  14. [joao.garin] yeah the site does behave well in both scenarios..just I don’t know which one would be preferable to add. ts for the tip that makes sense
  15. [joao.garin] thumbsup emoji
  16. zakim-robot
    @zakim-robot
    Oct 03 16:39
    [robdodson] @eric I didn't get into it much in the post but AOM phase 4 allows computing of the accessible tree. That's something we've never had before. It would for the first time allow us to do cross platform automated testing to ensure the things we build express the correct semantics in different browsers
  17. [robdodson] I'm not sure I follow the emerging markets argument. I think UC Browser is based on Chromium these days. Possibly other ones are as well. In theory it would be really easy for those folks to pull in the AOM patches when they land.
  18. Michael Fairchild
    @mfairchild365
    Oct 03 16:42
    computing of the accessible tree would be amazing, for the exact reason you just described.
  19. zakim-robot
    @zakim-robot
    Oct 03 16:42
    [robdodson] yeah. I think a lot of folks talk about automated screen reader testing as a kind of holy grail. This would be pretty darn close.

  20. [eric] @robdodson Hey, thanks for taking the time to answer. I just wanted to apologize for the criticism, my current project has been an unending slog of antipatterns and I’m definitely looking for a fight in the wrong place—sometimes it’s hard to see the forest from the trees.

    I really am excited for the potential that post represents, it’s hard to shake the skeptical mindset sometimes. Automated testing has been the only way we’ve gotten this project on track, so anything that improves on that would be wonderful.

  21. zakim-robot
    @zakim-robot
    Oct 03 16:50
    [robdodson] @eric no worries. I know web components frankly have a pretty bad track record when it comes to accessibility :)
  22. zakim-robot
    @zakim-robot
    Oct 03 18:05
    [svinkle] Hey @bethany.curley! :wave:
  23. zakim-robot
    @zakim-robot
    Oct 03 18:25
    [bethany.curley] Hey @svinkle! Thanks for the welcome.
  24. [robdodson] just saw this and thought it was interesting: https://twitter.com/rick_viscomi/status/915266192792842241
  25. [robdodson] using httparchive we're able to run lighthouse across the top 500k sites
  26. [svinkle] Wow. That’s both impressive that you can do that, and sad about the poor results.
  27. [robdodson] yeah there's a breakdown of scores in a separate thread where we're discussing improving the way scoring is surfaced in lighthouse: https://github.com/GoogleChrome/lighthouse/issues/3444#issuecomment-333266658
  28. [robdodson] a lot of the usual suspects at the top. Missing alt text, unlabeled form elements.
  29. [robdodson] In the near term I'm planning to redo the scoring so those common issues affect the score more
  30. [robdodson] longer term there's discussion around if we should do scoring at all or something else
  31. zakim-robot
    @zakim-robot
    Oct 03 18:39
    [svinkle] That’s an interesting thought. How does scoring work for “low hanging fruit” like a missing alt attribute, vs something more complex like focus management on a modal window? Perhaps I should just check it out myself.
  32. [svinkle] I’d think small things would appear on top as, “this is an easy fix, but the impact is huge, therefore it’ scored higher.”
  33. [robdodson] right now the scoring is not great. Basically every test has the same weight
  34. [robdodson] and we just tally up how many tests you failed versus how many you passed
  35. [robdodson] the first step is to factor criticality into the tests. Some failures are minor, some are critical. That should affect the score
  36. [svinkle] Right,
  37. [robdodson] the next step would be to not count non-applicable tests
  38. [robdodson] i think that's something axe is working on
  39. [robdodson] returning an array of tests that didn't apply
  40. [robdodson] but to your original question... it's hard to test something like focus managmeent in a modal window
  41. [robdodson] So I don't know if aXe even has a test for that because it would be tricky to do without false positives (I could be mistaken)
  42. [robdodson] For that reason folks are generally opposed to the scoring cuz you get a "100%" but really we were only able to test low hanging fruit
  43. [svinkle] True. I guess I was thinking of at least looking for typical attributes and reporting if anything’s missing? Hmm, actually yeah that’d be tough to even know that it’s “supposed” to be a modal to begin with.
  44. [robdodson] same reason why when you run axe on the command line it always leaves a message saying it can only really test like 20-50% of stuff
  45. [svinkle] yep, very true.
  46. [robdodson] yeah if a div is meant to be a modal but doesn't at least include a role="" then we have no way of knowing what that div is
  47. [svinkle] “Don’t forget to manually test all the things!”
  48. [robdodson] alice refers to this as the "ARIA bootstrap problem"
  49. [robdodson] a div is just a div till you at least give the system some kind of hint as to your intention
  50. [svinkle] Yep.
  51. [robdodson] once we have role="" aXe can test that it has required attributes, and that stuff isn't misspelled
  52. zakim-robot
    @zakim-robot
    Oct 03 18:45
    [marcysutton] I always advocate for testing focus management in your own app. There's only so much we can guess about intent from the framework level
  53. zakim-robot
    @zakim-robot
    Oct 03 19:05
    [blind3y3design] my 2 cents: leave scoring. A bad score always super valuable when trying to convince stakeholders that this is something that must be fixed.
  54. [robdodson] yeah i'm wondering if there is a sweet spot of saying "we can't test everything, but among the stuff we can test, here's how you're doing"
  55. [robdodson] but really stress that even getting 100 on that is just a small first step
  56. [svinkle] With the exception of colors, I’m kind of apposed to the idea that you should need to convince stakeholders of anything. This is your site or app that you’re working on. Does it fail the tests? If so, it’s your responsibility to fix it, not to ask permission first. My 5 cents CDN.
  57. zakim-robot
    @zakim-robot
    Oct 03 19:11
    [robdodson] I hear what you're saying. I don't think every team fits the same mold so there are some scenarios where giving folks a score might be motivating.
  58. [robdodson] I've worked at big enterprise places before which had their own scoring system. So teams needed to get at least an B+ or something to launch their app
  59. [robdodson] It's not perfect, but it at least gets something on their radar
  60. [robdodson] cuz otherwise those teams would ignore the topic entirely :\

  61. [blind3y3design] @svinkle I'd love to not have to convince people that we should be spending extra time fixing issues that produce no noticeable change for "the average user." Even in Ed-Tech where we're required by law to meet WCAG 2.0 and Section-508 it's like pulling teeth to make a11y a priority.

    Lighthouse has been a godsend. I can tell our other devs "Run these audits and validate that you're passing," and it's simple for them now.

  62. [robdodson] Lighthouse + aXe. Without aXe we would not have anywhere near the test coverage we do
  63. [svinkle] @blind3y3design Valid point; didn’t think about it at that angle, where you’re trying to get other devs on board. That’s gotta be rough.
  64. [eric] I’ve been singing Lighthouse’s praises for tying a11y with perf, such a smart decision. That usually opens the door to discussions about more thorough testing if the client is motivated by these kinds of metrics.
  65. [robdodson] y'all are totally making my day right now :)
  66. zakim-robot
    @zakim-robot
    Oct 03 19:16
    [blind3y3design] <3 at the very least I've gotten most of our devs to repeat "just use button" which is all thanks to you @robdodson!
  67. [svinkle] A daily scrum mantra: “Just use a button. Just use a button. Just use a button…”
  68. [robdodson] i felt the button pain a lot working on the Google I/O 2016 site...
  69. [robdodson] i came in late to that project and like everything was clickable... none of em were buttons
  70. [joshbruning] I can't seem to find the aXe Chrome Extension open source project - am I blind, or is it not open-sourced?
  71. [robdodson] all custom elements or icons or something else
  72. [svinkle] brutal.
  73. [robdodson] @joshbruning are you talking about lighthouse or the prior accessibility devtools extension?
  74. [robdodson] lighthouse is here: https://github.com/GoogleChrome/lighthouse
  75. [robdodson] that does the auditing
  76. [joshbruning] https://chrome.google.com/webstore/detail/axe/lhdoppojpmngadmnindnejefpokejbdd?hl=en-US
  77. [joshbruning] Oh, is the guidance to use lighthouse, then...
  78. [robdodson] oh that's aXe
  79. [robdodson] I use both tbh
  80. [robdodson] Lighthouse uses aXe's tests under the hood. So they run similar audits
  81. [robdodson] But because Lighthouse depends on aXe, it means we trail behind a bit
  82. [robdodson] all the axe stuff lives here: https://github.com/dequelabs
  83. zakim-robot
    @zakim-robot
    Oct 03 19:21
    [robdodson] i'm not sure if the extension source is on github or not... @marcysutton do you know?
  84. [marcysutton] Our extensions aren't open source, but the JavaScript engine is.
  85. [joshbruning] Is it possible to open-source the extension?
  86. [marcysutton] May I ask why? We aren't planning to do it.
  87. [joshbruning] aXe has the permission "Read and change all your data on the websites you visit" which is a security risk
  88. [robdodson] yeah it has to read the page to actually do the audit so that makes sense
  89. zakim-robot
    @zakim-robot
    Oct 03 19:27
    [marcysutton] I don't see how any accessibility auditing tool would be useful if it couldn't read the websites you visit. The change portion is how we make the highlighter work.
  90. [joshbruning] We would just like to perform a security review on the code.
  91. [marcysutton] Can you email me at marcy.sutton@deque.com and we can have a chat about it
  92. [joshbruning] Sure - I believe Rob will follow up. :)
  93. zakim-robot
    @zakim-robot
    Oct 03 19:48
    [thanks4allthefish] https://dropbox.design/
  94. [thanks4allthefish] oh dear god

  95. [blind3y3design] Quoting a tweet I saw earlier:

    Only tested in Chrome, FF and Safari scrolling is shite

  96. [joshbruning] Wow, I completely missed the scroll bar. I attempted to click the dropbox logo several times. (sweat emoji)
  97. zakim-robot
    @zakim-robot
    Oct 03 19:57
    [thanks4allthefish] the scroll bar doesn't move if I drag and move it, this is the worst site I have been to in recent time
  98. [joshbruning] It sometimes can be grabbed and sometimes not - depends on the state of the page
  99. [joshbruning] (can be grabbed if the right part is scrolling, but not the left)
  100. [thanks4allthefish] yeah, if it is in the area where the panes move, then scroll doesn't move
  101. zakim-robot
    @zakim-robot
    Oct 03 20:16
    [wenlib] My organization is going to upgrade to Windows 2010 in the next while. I'm wondering if the presence of narrator and the magnifier on everyone's desk top has changed the way you do accessibility testing in your org.
  102. zakim-robot
    @zakim-robot
    Oct 03 22:15
    [svinkle] Hey @jenn! :wave:
  103. [jenn] hello! =D :wave:
  104. [jenn] i'm excited to be here! i'm in the midst of writing a really long a11y strategy doc for eventbrite... trying to subversively push an a11y agenda
  105. [svinkle] Excellent!
  106. [jenn] big kudos to marcy for the encouragement!
  107. [jenn] and to cordelia and rob for coming to the office recently :)
  108. [svinkle] Awesome! Sounds like a good time to me. :)