fluid-work IRC Logs-2008-05-02

Owned by Fluid IRC Log Bot
Last updated: May 02, 2008Version comment
6 min read

[09:30:41 EDT(-0400)] * colinclark (n=colin@bas1-toronto09-1279336219.dsl.bell.ca) has joined #fluid-work
[09:46:45 EDT(-0400)] * apetro-_ (n=apetro@12.20.36.2) has joined #fluid-work
[09:56:25 EDT(-0400)] * theclown (n=theclown@142.150.154.101) has joined #fluid-work
[10:44:05 EDT(-0400)] * apetro-_ (n=apetro@12.20.36.2) has joined #fluid-work
[11:10:54 EDT(-0400)] * apetro-_ (n=apetro@12.20.36.2) has joined #fluid-work
[11:40:24 EDT(-0400)] * apetro-_ (n=apetro@205.215.222.135) has joined #fluid-work
[11:43:36 EDT(-0400)] * EricDalquist (n=EricDalq@205.215.222.135) has joined #fluid-work
[12:43:55 EDT(-0400)] * anastasiac (n=team@142.150.154.105) has joined #fluid-work
[13:06:10 EDT(-0400)] * aaronlev (n=chatzill@pD9E4E5B5.dip.t-dialin.net) has joined #fluid-work
[13:29:14 EDT(-0400)] * aaronlev (n=chatzill@pD9E4E5B5.dip.t-dialin.net) has joined #fluid-work
[13:30:41 EDT(-0400)] <theclown> it's quiet in here. so:
[13:30:44 EDT(-0400)] <theclown> http://yro.slashdot.org/yro/08/05/02/0049225.shtml
[14:57:02 EDT(-0400)] * aaronlev_ (n=chatzill@pD9E4DF68.dip.t-dialin.net) has joined #fluid-work
[15:01:47 EDT(-0400)] * anastasiac (n=team@142.150.154.105) has joined #fluid-work
[15:14:53 EDT(-0400)] <anastasiac> theclown: this is good news!
[15:16:03 EDT(-0400)] * davidb (n=davidb@bas4-toronto06-1279309985.dsl.bell.ca) has joined #fluid-work
[15:16:06 EDT(-0400)] <theclown> anastasiac, yup
[15:17:08 EDT(-0400)] * theclown anastasiac, davidb and I were discussing the active descendent aria property, and how it works.
[15:17:34 EDT(-0400)] * davidb goes to the kettle
[15:18:25 EDT(-0400)] <anastasiac> ah - ok, any questions/comments I should know about?
[15:18:31 EDT(-0400)] <davidb> http://www.w3.org/TR/wai-aria-practices/#focus_activedescendant
[15:18:35 EDT(-0400)] * theclown at one point, the Reorderer set the ad value on the container node as focus moved about the container. e.g, a thumbnail gets focus ==> the ad property of the lightbox is updated. Is that how it works>
[15:18:50 EDT(-0400)] * theclown davidb is stealing my thunder.
[15:18:57 EDT(-0400)] * anastasiac looks at the code...
[15:19:09 EDT(-0400)] <theclown> actually, the correct url to look at is:
[15:19:13 EDT(-0400)] <davidb> theclown: (smile)
[15:19:33 EDT(-0400)] <theclown> http://www.w3.org/WAI/PF/Group/aria/#activedescendant
[15:19:42 EDT(-0400)] <theclown> whoops. that's the private link. hang on...
[15:19:56 EDT(-0400)] <anastasiac> yes, that's how it works: as focus moves around, the ad val on the container is updated to reference the newly focused thing
[15:20:16 EDT(-0400)] * davidb waits for theclown's thunder
[15:20:37 EDT(-0400)] <theclown> http://www.w3.org/TR/wai-aria/#activedescendant
[15:20:40 EDT(-0400)] <theclown> boom?
[15:21:30 EDT(-0400)] * theclown it says that this the ancestor maintains keyboard focus, and updates the ad to tell the AT what is "active". It's for a case where focus does not move to the descendant.
[15:22:12 EDT(-0400)] <anastasiac> ah, ok, I see the distinction you're making - is actual focus still on the container, or on the actual descendent?
[15:23:09 EDT(-0400)] <theclown> for the active descendant case, it's on the container.
[15:23:23 EDT(-0400)] <anastasiac> for the reorderer case, it's the descendent. hm...
[15:23:40 EDT(-0400)] <anastasiac> does that mean that activedescendent is inappropriate for us?
[15:24:00 EDT(-0400)] <theclown> the sense i get is if you are not going to move focus to the descendant, but keep it on the container, then somehow one has to inform the AT what is active within that container
[15:24:14 EDT(-0400)] <theclown> re: does that mean that activedescendent is inappropriate for us? I'm beginning to think so.
[15:24:21 EDT(-0400)] <anastasiac> right - so we don't need it, because we are moving focus?
[15:24:26 EDT(-0400)] <theclown> exactly
[15:24:26 EDT(-0400)] <anastasiac> interesting...
[15:25:30 EDT(-0400)] <theclown> if the UA implements aria, and they use ad as described, what happens when (1) the thumbnail has focus ,and (2) the lightbox's ad property is updated. Does the AT get told what is active twice? Dunno, but that's the worry.
[15:26:33 EDT(-0400)] <anastasiac> good question - might be nice to try to find out, either through use of an actual AT, or by asking the AT developers who know exactly what's done internally with this attribute
[15:27:26 EDT(-0400)] <davidb> or just ask aaronlev_ he knows everthing
[15:27:38 EDT(-0400)] <aaronlev_> i know almost nothing
[15:27:41 EDT(-0400)] <theclown> ah, and he is in the room.
[15:27:53 EDT(-0400)] <davidb> aaronlev_: shhh, i've built you up man.
[15:28:03 EDT(-0400)] <aaronlev_> heh
[15:28:06 EDT(-0400)] <theclown> davidb, i can hear you.
[15:28:17 EDT(-0400)] * davidb plugs his ears
[15:28:19 EDT(-0400)] <davidb> now you can't!
[15:28:23 EDT(-0400)] <davidb> oh darn
[15:28:29 EDT(-0400)] <aaronlev_> if you move focus and set ad (or vice versa), all in the same event handler, it ends up as 1 focus event
[15:28:51 EDT(-0400)] <aaronlev_> we set a timer of 0 when focus or ad of changes
[15:28:51 EDT(-0400)] <davidb> aaronlev_: cool so FF sends the focus event to platform a11y?
[15:28:57 EDT(-0400)] <aaronlev_> yes
[15:29:02 EDT(-0400)] <davidb> makes sense
[15:29:08 EDT(-0400)] <aaronlev_> and after our tiny timer we fire the focus event
[15:29:16 EDT(-0400)] <aaronlev_> so we know where focus really is and can avoid dupes
[15:29:18 EDT(-0400)] <aaronlev_> etc.
[15:29:36 EDT(-0400)] <davidb> aaronlev_: is there any difference in user experience (screen reader) between using AD, vs giving children focus?
[15:29:40 EDT(-0400)] <aaronlev_> hmm seems like we should talk about that in the ua implers guide
[15:29:46 EDT(-0400)] <aaronlev_> davidb: no
[15:29:50 EDT(-0400)] <davidb> aaronlev_: ok nice.
[15:29:57 EDT(-0400)] <aaronlev_> maybe in IE (smile)
[15:30:03 EDT(-0400)] * davidb vomits
[15:30:15 EDT(-0400)] <davidb> aaronlev_: are they talking to you as they impl?
[15:30:31 EDT(-0400)] <aaronlev_> a little
[15:30:36 EDT(-0400)] <aaronlev_> they really appreciate the guide
[15:30:44 EDT(-0400)] <davidb> good to hear
[15:31:06 EDT(-0400)] <davidb> the UA BP i assume
[15:31:22 EDT(-0400)] <anastasiac> aaronlev_, we set the activedescendent in the focus event handler
[15:31:29 EDT(-0400)] <davidb> anastasiac: so sounds like you can use AD, or give children focus for same user experienct
[15:31:48 EDT(-0400)] <davidb> anastasiac: focus event handler for the container?
[15:32:00 EDT(-0400)] <aaronlev_> anastasiac: i think that should be fine
[15:32:05 EDT(-0400)] <anastasiac> no, in the the focus handler for the descendent, we set the property in the container
[15:32:12 EDT(-0400)] <davidb> oh i see
[15:32:27 EDT(-0400)] <aaronlev_> why would the descendant get focus at all?
[15:32:39 EDT(-0400)] <aaronlev_> the descendant shouldn't have tabindex="-1" or anything
[15:32:42 EDT(-0400)] <aaronlev_> when yoiu use ad
[15:32:56 EDT(-0400)] <davidb> aaronlev_: but in this case i think it does
[15:33:02 EDT(-0400)] <davidb> shouldn't but does
[15:33:11 EDT(-0400)] <aaronlev_> i guess technically it's allowed
[15:33:19 EDT(-0400)] <aaronlev_> but then focus would need to be rerouted to the parent
[15:33:24 EDT(-0400)] <aaronlev_> or ancestor anyway
[15:33:28 EDT(-0400)] <aaronlev_> that is the containing widget
[15:33:33 EDT(-0400)] <aaronlev_> where activedescendant is set
[15:33:53 EDT(-0400)] <davidb> anastasiac, might be talking about key event handling too though...
[15:33:56 EDT(-0400)] <davidb> which bubbles
[15:34:11 EDT(-0400)] <aaronlev_> what is the advantage of making the descendants actually focusable if you are using ad
[15:34:40 EDT(-0400)] <theclown> anastasiac, one thing I seem to recall is that the lightbox used active descendant to maintain which thumbnail should get focus when switching between windows.
[15:35:19 EDT(-0400)] <anastasiac> putting focus on the descendent is more natural for activation of the link within the descendent
[15:35:22 EDT(-0400)] <davidb> anastasiac, theclown, do you ever change tabindex to "0" for a thumbnail?
[15:35:37 EDT(-0400)] <anastasiac> davidb, no, we don't
[15:35:38 EDT(-0400)] <davidb> anastasiac: maybe don't want AD then.
[15:35:44 EDT(-0400)] <davidb> anastasiac: hmm. ok
[15:36:13 EDT(-0400)] <anastasiac> it does sound like activedescendent is designed to accomplish a purpose that we don't need accomplished, because we're doing it ourselves, essentially
[15:36:17 EDT(-0400)] <davidb> i'm going to blog about this i think... i was waffling about it before but this might be enough.
[15:37:20 EDT(-0400)] <davidb> anastasiac: sort of. the best practices does suggest for containers... one uses the activedescendant 'pattern'.
[15:37:46 EDT(-0400)] <anastasiac> I think I need to take a closer look at the spec and the best practices
[15:38:06 EDT(-0400)] <davidb> anastasiac: i'm happy to sit with you monday... maybe theclown too?
[15:38:19 EDT(-0400)] <theclown> how does the Reorderer remember which thumnail has focus when you switch away by moving to another window (or frames)?
[15:38:25 EDT(-0400)] <anastasiac> maybe - let's talk then
[15:38:32 EDT(-0400)] <davidb> k
[15:38:33 EDT(-0400)] <anastasiac> theclown, I'll double check...
[15:38:38 EDT(-0400)] * anastasiac looks...
[15:39:13 EDT(-0400)] <anastasiac> hm. we actually don't use the activedescendent, we keep track of it ourselves
[15:39:37 EDT(-0400)] <theclown> then, active descendant: begone!
[15:39:41 EDT(-0400)] <anastasiac> I think the history is that the activedescendent stuff was added on after
[15:39:55 EDT(-0400)] <theclown> or, wait until we fully understand the spec and best practices
[15:39:57 EDT(-0400)] <anastasiac> perhaps begone, yes...
[15:40:20 EDT(-0400)] <anastasiac> perhaps understanding is the better approach than banishment
[15:40:43 EDT(-0400)] * anastasiac is appalled at her own grammar
[15:40:50 EDT(-0400)] <davidb> anastasiac: and we might inform the spec with our understanding (smile)
[15:41:03 EDT(-0400)] <anastasiac> maybe...
[15:41:28 EDT(-0400)] <anastasiac> though it sounds like the ad stuff serves a useful purpose for the case that it is intended for, we just don't happen to be one of those cases
[15:42:06 EDT(-0400)] <theclown> anastasiac, yes, that was my take earlier.
[15:42:12 EDT(-0400)] <davidb> anastasiac: right. i think lightbox might be a good case for a pattern we use in dojo tabcontainer
[15:42:16 EDT(-0400)] <theclown> that's why I started this discussion
[15:43:16 EDT(-0400)] <anastasiac> davidb: was this pattern what you were going to blog about?
[15:43:29 EDT(-0400)] <davidb> anastasiac: yeah.
[15:43:53 EDT(-0400)] <anastasiac> ok, then I don't need to ask you to tell me about the pattern, I can just read your blog (smile)
[15:44:06 EDT(-0400)] <davidb> anastasiac: ideally i would put it up on the DHTML a11y wiki we need to create.
[15:44:15 EDT(-0400)] <davidb> anastasiac: yeah.
[15:47:28 EDT(-0400)] <davidb> theclown: good catch btw.
[15:47:45 EDT(-0400)] <theclown> davidb, thanks
[15:53:05 EDT(-0400)] <anastasiac> for the record, I've created a JIRA for looking into this: http://issues.fluidproject.org/browse/FLUID-565
[17:41:50 EDT(-0400)] * theclown_ (n=theclown@142.150.154.101) has joined #fluid-work
[17:50:44 EDT(-0400)] * aaronlev_ (n=chatzill@pD9E4DF68.dip.t-dialin.net) has joined #fluid-work
[18:03:02 EDT(-0400)] * theclown (n=theclown@142.150.154.101) has joined #fluid-work
[18:03:42 EDT(-0400)] * theclown (n=theclown@142.150.154.101) has left #fluid-work
[21:50:31 EDT(-0400)] * EricDalquist (n=EricDalq@ppp-70-226-161-218.dsl.mdsnwi.ameritech.net) has joined #fluid-work