IRC log of UA on 2004-04-16

Timestamps are in UTC.

00:04:30 [RRSAgent]
RRSAgent has joined #UA
00:05:32 [MattSEA]
zakim, aacc is Susan_Crane
00:05:32 [Zakim]
+Susan_Crane; got it
00:05:55 [MattSEA]
zakim, Rich is Mary_Hoetzel
00:05:55 [Zakim]
+Mary_Hoetzel; got it
00:05:59 [MattSEA]
zakim, aaaa is Rich
00:05:59 [Zakim]
+Rich; got it
00:06:05 [MattSEA]
zakim, Earl is Earl_Johnson
00:06:05 [Zakim]
+Earl_Johnson; got it
00:08:56 [MattSEA]
zakim, Susan is Susan_Crayne
00:08:56 [Zakim]
+Susan_Crayne; got it
00:12:05 [MattSEA]
zakim, who's making noise?
00:12:16 [Zakim]
MattSEA, listening for 10 seconds I heard sound from the following: ??P0 (49%), Susan_Crayne (66%), Peter_Korn (32%)
00:12:49 [MattSEA]
rrsagent, make log global
00:12:53 [MattSEA]
rrsagent, make log world
00:13:10 [MattSEA]
jg: Roadmap progress?
00:13:31 [MattSEA]
rs: Work on role in RDF, and accelerators.
00:14:34 [MattSEA]
rs: Some people want to store accesskey info in an external file. May be too much for authors, but looking into it.
00:14:59 [MattSEA]
al: RDF has issues with smaller browsers/parsers.
00:15:30 [MattSEA]
rs: Want to have standardized role types and use them for pervasive devices. Wouldn't want to create a lot of custom widgets and extend them.
00:17:05 [MattSEA]
jg: Extension work on IE and Mozilla.
00:17:22 [Zakim]
+ +1.512.431.aadd
00:17:41 [MattSEA]
zakim, aadd is Michael_Lew
00:17:41 [Zakim]
+Michael_Lew; got it
00:18:37 [MattSEA]
jg: Provides standard features for screen readers, etc., a common format people can use instead of constantly reinventing.
00:18:59 [Zakim]
00:19:21 [MattSEA]
zakim, ??P8 is Jessie_Li
00:19:21 [Zakim]
+Jessie_Li; got it
00:21:13 [MattSEA]
jg: How would caret in Mozilla work?
00:21:19 [MattSEA]
pk: That's critical.
00:22:24 [MattSEA]
al: uses the same functionality as edit, but has same problems.
00:22:36 [MattSEA]
al: I turned things on that had been turned off, and let the user enter into that mode.
00:23:24 [MattSEA]
jg: This is in Firefox?
00:23:32 [MattSEA]
al: Yes, but need to set a preference.
00:23:43 [MattSEA]
al: It uses the layout to decide what is the next block, not the DOM tree.
00:24:28 [MattSEA]
zakim, who's making noise?
00:24:38 [Zakim]
MattSEA, listening for 10 seconds I heard sound from the following: Mary_Hoetzel (28%), Rich (71%), Susan_Crayne (46%), Peter_Korn (4%)
00:25:12 [Zakim]
00:25:39 [MattSEA]
al: It might not be bad to fix, but would need to look at it.
00:25:43 [Zakim]
00:25:52 [MattSEA]
pk: One of the higher priority problems with Mozilla on Unix, I would say.
00:26:03 [MattSEA]
zakim, ??P8 is Jessie_Li.
00:26:03 [Zakim]
+Jessie_Li.; got it
00:26:52 [MattSEA]
al: Daniel Glazman is working on Nvu, the composer component.
00:27:17 [MattSEA]
rs: Should ask him, or get him to point to where to fix it in edit mode.
00:27:52 [Zakim]
00:28:29 [MattSEA]
pk: To do what you can do with a mouse, you need to be able to select text, which includes every character in the doc. Not possible right now. And you'd want to do that in document order.
00:28:51 [MattSEA]
zakim, ??p8 is Kyle_Yuan
00:28:51 [Zakim]
I already had ??P8 as Jessie_Li., MattSEA
00:28:59 [MattSEA]
zakim, ??p9 is Kyle_Yuan
00:28:59 [Zakim]
+Kyle_Yuan; got it
00:29:10 [MattSEA]
zakim, Jessie_Li. is Jessie_Li
00:29:10 [Zakim]
+Jessie_Li; got it
00:30:08 [MattSEA]
pk: Currently the only mode in Gnopernicus is where the caret is. No rerendering for layout purposes.
00:31:38 [MattSEA]
al: Might be a good idea to have the caret work from the DOM when in the browser. Or if a screen reader is in use.
00:33:17 [MattSEA]
al: With XBL, which is like #define, you can turn any tag into any other tag, but that would be changing the DOM. Not sure why there's dual layout in DOM arch.
00:33:20 [MattSEA]
rs: IE has the same thing.
00:34:14 [MattSEA]
al: It's a question of at what point we update the DOM. Only with screen readers? All users? Only in browser, not editor?
00:34:43 [MattSEA]
al: I'll put it on my list to figure out.
00:36:06 [MattSEA]
al: Maybe DOM order will be so much better, or introduce new bugs. So maybe add a hidden pref to turn it on and off, and we can experiment.
00:37:29 [MattSEA]
al: Rendered order is what it does right now.
00:37:36 [MattSEA]
pk: No, because there are things it will never reach.
00:37:40 [MattSEA]
al: Those are bugs.
00:37:58 [MattSEA]
pk: All I care about is an easy to select option.
00:38:10 [MattSEA]
zakim, who's making noise?
00:38:23 [Zakim]
MattSEA, listening for 10 seconds I heard sound from the following: Mary_Hoetzel (4%), Rich (20%), Susan_Crayne (58%), Peter_Korn (69%), Jessie_Li (14%), Kyle_Yuan (10%)
00:38:31 [MattSEA]
zakim, mute Susan
00:38:31 [Zakim]
Susan_Crayne should now be muted
00:40:27 [MattSEA]
jg: Any way to map between the DOM and the rendering engine?
00:40:32 [MattSEA]
al: They map to each other.
00:40:39 [MattSEA]
jg: So you would have an idea about how they're rendered.
00:41:22 [MattSEA]
pk: I think those bugs have been filed.
00:44:41 [MattSEA]
zakim, who's making noise?
00:44:52 [Zakim]
MattSEA, listening for 10 seconds I heard sound from the following: Mary_Hoetzel (4%), Rich (95%), Peter_Korn (4%), Jessie_Li (4%)
00:45:25 [MattSEA]
al: We use rendering information for bounds.
00:47:32 [MattSEA]
al: On Windows, our table nav was based on one table element with lots of children.
00:49:49 [MattSEA]
zakim, who's talking?
00:49:59 [Zakim]
MattSEA, listening for 10 seconds I heard sound from the following: Mary_Hoetzel (29%), Kyle_Yuan (4%)
00:50:20 [MattSEA]
zakim, mute Mary temporarily
00:50:20 [Zakim]
Mary_Hoetzel should now be muted
00:50:35 [MattSEA]
zakim, unmute Mary
00:50:35 [Zakim]
Mary_Hoetzel should no longer be muted
00:50:36 [Zakim]
Mary_Hoetzel should now be unmuted again
00:52:21 [Zakim]
00:53:37 [MattSEA]
al: There's more to be fixed on this than caret browsing.
00:55:00 [MattSEA]
pk: Whether screen reader should just suck out the data and do what it needs to is still an unanswered question.
00:55:15 [MattSEA]
al: What if the screen reader takes over the arrow keys and handles the selection.
00:55:19 [MattSEA]
pk: How should it know?
00:55:36 [MattSEA]
al: It has ATK.
00:55:51 [MattSEA]
rs: Interface isn't tailored for that.
00:55:57 [MattSEA]
pk: May be possible to build something like that.
00:56:48 [MattSEA]
pk: Mozilla is in the best position to provide the kind of navigation we need, over the screen reader.
00:57:17 [MattSEA]
al: I don't think there's a model for keyboard interaction under Mozilla.
01:00:39 [MattSEA]
pk: IBM has done a lot of work on HPR. Shows they've got a lot to share on keyboard.
01:02:13 [MattSEA]
jg: One problem we had with the extension was using the caret to move keyboard focus to the header. Maybe that's a bug.
01:03:37 [MattSEA]
jg: Gnopernicus can read list boxes, so helps get around the DOM needs.
01:05:32 [MattSEA]
pk: Gnopernicus is not the only AT that will need to do this. There are 3 ATs on GNOME, and more coming.
01:06:45 [Zakim]
01:07:06 [MattSEA]
pk: If you had the logical navigation of the caret, all consumers of that would get the benefit.
01:11:46 [kyle]
kyle has joined #ua
01:18:38 [MattSEA]
next meeting: Monday, May 10 (Tuesday, May 11 Beijing)
01:23:08 [Zakim]
01:23:17 [Zakim]
01:23:18 [Zakim]
01:23:18 [Zakim]
01:23:19 [Zakim]
01:23:20 [Zakim]
01:23:22 [Zakim]
01:23:23 [Zakim]
01:23:26 [Zakim]
01:23:27 [Zakim]
WAI_UAWG()8:00PM has ended
01:23:28 [Zakim]
Attendees were Matt, +1.781.316.aaaa, +1.512.527.aabb, +1.914.328.aacc, Peter_Korn, Susan_Crane, Mary_Hoetzel, Rich, Earl_Johnson, Susan_Crayne, +1.512.431.aadd, Michael_Lew,
01:23:30 [Zakim]
... Jessie_Li, Kyle_Yuan
01:23:39 [JRG]
Matt could you reserve the bridge?
01:23:54 [JRG]
For April 10th?
01:24:17 [MattSEA]
01:24:20 [MattSEA]
rrsagent, bye
01:24:20 [RRSAgent]
I see no action items