Archive index

A11y Slackers Gitter Channel Archive 8th of June 2017

What fresh hell is THIS now? - Patrick Lauke
  1. zakim-robot
    @zakim-robot
    Jun 08 01:11
    [caesar] Can I get an opinion on something: when there are two consecutive links in a paragraph, NVDA in browse mode will simply plough through both without stopping. I believe in JAWS, VO and others, an explicit keypress is required to continue reading once it hits a link.
  2. [caesar] First, is that right? And second, in your opinion is this behaviour desirable?
  3. [caesar] (the NVDA one described, I mean)
  4. [caesar] I believe this is something to do with the interpretation of "phrasing content" in HTML?
  5. James Teh
    @jcsteh
    Jun 08 01:17
    caesar: That depends on how you're reading. If you're reading continuously (say all), then all of them will continue to and past a link. If you're reading line by line (or element by element in the case of VO), then the behaviour is different.
    NVDA defaults to "screen layout", which means that if two elements are on the same line visually, they will be rendered on the same line in browse mode. The advantage of this is that the flow of text isn't broken, which is particularly useful for articles with a lot of links; think Wikipedia, where every second word is a link. Without screen layout, the user has their flow broken for every one of those links. It also gives the user a better feel for something like date entry where the dd, mm and yyyy fields are all separate.
    Note that users can turn off screen layout if they wish.
    JAWS also has screen layout, but it is disabled by default.
  6. caesar: Speaking technically, screen layout only breaks lines for block elements. With screen layout off, lines are also broken for inline elements.
  7. caesar: disclosure: I'm one of the NVDA lead devs.
  8. zakim-robot
    @zakim-robot
    Jun 08 01:20
    [caesar] @jcsteh - Hi James. Yes, I know of you. Big fan of your work :)
  9. James Teh
    @jcsteh
    Jun 08 01:21
    Side note: Those of you talking via zakim-robot... what are you using? IRC or something else? I'm using Gitter.
  10. zakim-robot
    @zakim-robot
    Jun 08 01:21
    [caesar] @jcsteh: the reason I ask is because I've come across the situation many times where designers have used two <a> in a row to implement different things onscreen interactive elements, and positioned them using CSS
  11. James Teh
    @jcsteh
    Jun 08 01:23
    [caesar] Are they positioned horizontally or vertically? If horizontally, rendering them on the same line may or may not make sense, depending. If vertically, ideally, they wouldn't be rendered on the same line... and I'm surprised they would be, but I guess it's not impossible if somehow the elements got a display style of inline.
  12. zakim-robot
    @zakim-robot
    Jun 08 01:23
    [caesar] Just trying to figure out whether I should be recommending that they use block elements, or if this is even a problem.
  13. [caesar] @jcsteh - vertically.
  14. James Teh
    @jcsteh
    Jun 08 01:24
    [caesar] I'd be interested to see a test case for that behaviour if you have one.
  15. zakim-robot
    @zakim-robot
    Jun 08 01:24
    [caesar] @jcsteh - most of the bot users I know are on Gitter, e.g. the Deque guys
  16. [caesar] @jcsteh let me PM you something
  17. [caesar] Oh wait. Not sure how that works in zakim...
  18. James Teh
    @jcsteh
    Jun 08 01:25
    So if they're on Gitter, why do they come up via the bot? Different Gitter channel?
  19. zakim-robot
    @zakim-robot
    Jun 08 01:29
    [caesar] You mean Slack? I believe this chat is native to Slack, and Gitter users are accessing it through Zakimbot... if I'm understanding your question corrrectly.
  20. James Teh
    @jcsteh
    Jun 08 01:30
    Oh! I thought this was native Gitter. I'd probably prefer Slack actually.
  21. zakim-robot
    @zakim-robot
    Jun 08 01:31
    [caesar] I could be wrong. It's just the impression I got
  22. [caesar] There are definitely a whole bunch more channels in web-a11y.slack.com
  23. zakim-robot
    @zakim-robot
    Jun 08 04:43
    [marcysutton] You can get into this Slack via web-a11y.herokuapp.com.
  24. Sean Elliott
    @seanus1138
    Jun 08 07:28

    Hey guys, im developing a keyboard accessible main menu script - the main menu can have simple drop downs or mega menu drop downs.

    For desktop the main menu will be horizontal and on mobile it will be vertical.

    I have a desktop version of the script working correctly as per the interactions described here https://www.w3.org/TR/wai-aria-practices/#menu.
    What im looking for is some advice for handling the keyboard interactions when the main menu is in a mobile viewport (https://codepen.io/cbracco/full/kvbgz).

    When your in the top level menu on mobile because the menu is vertical the up and down arrows moves the focus to the prev and next item. Space, enter, left and right open the sub menu and shift the focus to the first item in the sub menu. This is as per the keyboard interactions.

    What im not clear on for mobile is the interaction when in the sub menu. On desktop using the left and right key, the interaction is to close the current sub menu. Move focus to next top level menu item. Open the sub menu of the next top level menu item with focus.

    Would it be advisable to keep that interaction on mobile?
    Hopefully this makes sense.

  25. zakim-robot
    @zakim-robot
    Jun 08 11:34
    [sophieschoice_slack] question: for a familytree of a royal family, I currently use nested numbered lists (<ol>), though I can imagine if a screenreader-using visitor reaches the level of grandchildren, it would be difficult to "see" to which branch of the tree or to which parents/grandparents the current person in the list belongs to. Or isn´t it a problem because it is a nested list? What would you suggest for presenting a family tree?
  26. [sophieschoice_slack] in my newbie mind numbered lists would be the way to go, but that is just me...
  27. zakim-robot
    @zakim-robot
    Jun 08 12:44
    [geekf] I would suggest nested unordered list <ul> unless you are sure about how to to order them in each generation.
  28. [geekf] I am not sure how the context for previous generation can be provided once you have moved down the list but from the markup point of view that sounds logical.
  29. zakim-robot
    @zakim-robot
    Jun 08 14:03
    [quidkid] Would it be possible to make a menu like this accessible? why or why not: http://webredox.net/demo/wp/anema/demo1/
  30. powrsurg
    @powrsurg
    Jun 08 14:03
    Does a link with the CSS property "pointer-events: none;" change the accessibility of it? You can tab to it, but not click on it?
  31. zakim-robot
    @zakim-robot
    Jun 08 14:32
    [scottohara] pointer events none just makes it so you can’t click. you can still tab to that element and activate it with whatever it’s activation keys are
  32. zakim-robot
    @zakim-robot
    Jun 08 15:10
    [spell] Hi gang. Quick question. Are there any concerns with vertical text like this? https://www.screencast.com/t/huUZTbWYEN
  33. zakim-robot
    @zakim-robot
    Jun 08 15:52
    [ricky_haror] You have to focus more to read it. People don't read vertically. I personally would not recommend it.
  34. zakim-robot
    @zakim-robot
    Jun 08 15:58
    [tyronem] would any users with cognitive issues have problems reading it?
  35. zakim-robot
    @zakim-robot
    Jun 08 17:50
    [spell] @tyronem Yeah, that is what I am asking here, really. Would love to know. I'm leaning against it, but would love more insight.
  36. James Nurthen
    @jnurthen
    Jun 08 19:33
    Anyone have an opinions on how to announce colspans and rowspans in a grid. The grid we have is self voicing as ARIA doesn't support enough of the things we need so there is a lot of flexibility. I was thinking of the following
  37. cell value - Merged Cell - 3 wide, 3 high (or something like that)
  38. Before this we would already have the row/column headers if they have changed from the previously visited cell.
  39. It is similar to what Excel does but excel announces c3 through d4 - but announcing all the row/column headers for the different cells seems too verbose to me. They can get away with it due to their headers being guaranteed to be short.
  40. zakim-robot
    @zakim-robot
    Jun 08 19:49
    [kivi] Until someone decides to merge cells XFC1048576 and XFD1048576 that is. ;)
  41. James Nurthen
    @jnurthen
    Jun 08 19:58
    :)