Canvas Accessibility Sub-Group Teleconference

13 Jan 2014

See also: IRC log


Rich Schwerdtfeger, Mark Sadecki, Jatinder, Rik Cabanier, Philippe, Paul Cotton
Mark Sadecki
Mark Sadecki


<trackbot> Date: 13 January 2014

<scribe> scribe: MarkS


Review Action items taken up at last meeting

MS: I will follow up with Jay Munro regarding the two action items on him for last week.

Revisit Bug 23980 RE: Naming of focusRing methods http://j.mp/L3I6tf

-> http://www.w3.org/WAI/PF/HTML/wiki/Summary_of_focusRing_method_naming#Summary Summary

MS: drawFocusIfNeeded seems to be a favorite at this point.

JMann: I am OK with that

RS: : I am OK with that

JMann: drawFocusIfNeeded seems to solve all the issues I originally raised.
... talked about splitting the methods?

RS: Ian originally wanted to have both together to make it easier on developers. I still think that is a good idea.

RC: If we change the name and clarify purpose in spec, there should be no reason to split them up

JMann: seems like if you are using the 2D context, this is the way to do it. Spoke with Graphic designers that understand this concept

RESOLUTION: to change drawSystemFocusRing to drawFocusIfNeeded

Continue Bug Walkthrough (next up: 23984-23987)

<plh> https://www.w3.org/Bugs/Public/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&component=CR%20HTML%20Canvas%202D%20Context&list_id=30985&order=Importance&product=HTML%20WG&query_format=advanced

-> https://www.w3.org/Bugs/Public/show_bug.cgi?id=23984 23984 Is it possible to implement drawCustomFocusRing? (JM. Moved to Level 2)

-> https://www.w3.org/Bugs/Public/show_bug.cgi?id=23985 23985 drawCustomFocusRing and color scheme (JM. Moved to Level 2)

-> https://www.w3.org/Bugs/Public/show_bug.cgi?id=23986 23986 Current default path

-> https://www.w3.org/Bugs/Public/show_bug.cgi?id=23987 23987 Empty path and drawSystemFocusRing

PLH: some of these are still assigned to CR component

RS: 23984 should not be in the list because it is being moved to L2
... 23985 same thing

JMann: 23986, which path actually gets drawn?

PLH: I think you are talking about 23987

JMann: if you call it immediately, what happens?

RC: depends on what you have done previously

PLH: Assuming you have not done anything

RC: whatever path is in the canvas at that point, will be the focus area

PLH: The spec states that there is always a current default path, what is it?

<richardschwerdtfeger> http://www.w3.org/html/wg/drafts/2dcontext/html5_canvas_CR/#current-default-path

JMann: scenario 1, just created canvas, call drawFocusIfNeeded, what happens, and just drew a path and called drawFocusIfNeeded what happens

PLH: I will amend the bug

JMann: what do we think should happen?

PLH: I don't think anything should happen, there is no default path when the canvas element has been created

JMann: if you call drawFocusIfNeeded before than you have a region associated with that path

RS: can always have the path be outside the canvas area, nothing gets drawn
... it should have a default path of some sort

RC: by default a fallback element has not region associated with it.

JMann: I think it should get the bounds of the entire canvas element

RS: Makes sense

RC: that is not in the spec right now right?

RS: no, should be

RC: should that be in the HTML spec? That is where we talk about fallback content

RS: What does context.fill do ?


RC: if you bring focus to fallback content, but the associated path has not even been drawn, nothing should happen

JMann: agree
... so if I do the beginPath forces UA to cache the last path always.

<richardschwerdtfeger> For CanvasRenderingContext2D objects, the points and lines added to current default path by these methods must be transformed according to the current transformation matrix before they are added to the path.

RS: 0,0 is path, doesn't go anywhere?

JMann: i think we should just change algorithm to abort

RC: you need to update the region

RS: If you are adding to the path, then yes

RC: updated a11y api with the bounding box of the element

RS: we could put a note that says to make sure that you have created a path

RC: there is always a path

<JatinderMann> The beginPath() method must empty the list of subpaths in the context's current path so that the it once again has zero subpaths.

JMann: the moment you call beginPath, you have no path until you build on that path . it's empty
... so we draw the bounding of the canvas, or draw the last path, or we draw 0,0 or we ??

RC: If the current default path is empty, then don't draw anything but set the accessible region back to the original value

JMann: do we all agree that it is an error case?

RC: I think it something an author might do
... a way to remove the association with the fallback content

JMann: ah to reset a11y dom
... that is a use case
... means we shouldn't do nothing
... so 0,0 or the entire canvas. I like entire canvas, which is still true, but it may not be a real reset.

RC: can we say initial value?

RS: if its the entire canvas the magnifier focuses on entire canvas

JMann: is there a default location?

RS: not that I am aware of

RC: can you make it invisible?

RS: yes

JMann: If we can tie it to the initial value (invisible) then that should cover the use case.

JMann: does this feel buggy or will developers get this concept
... this should also resolve 23987

<richardschwerdtfeger> https://developer.mozilla.org/en-US/docs/Web/Accessibility/AT-APIs/MSAA/States

RS: states are different on each platform. MSAA has the concept of unavailable


RS: offscreen means not visible, but not necessarily gone.
... that means disabled, so that wouldn't work
... if the author really wants to remove it from the AT they should just remove it from the AAPI

JMann: are we expected developers to delete the element and add it back later on?
... the recommend approach for removing your controls would be to change the ARIA state or delete the node. Maybe we should just a no op if the path is empty

RS: so do nothing?


JMann: should be as easy as updating the processing model

RESOLUTION: Update the process to include "if there are zero sub paths in the current path, abort these steps"

RS: we can close 23986 and 23987 when this is done.

<plh> https://www.w3.org/Bugs/Public/show_bug.cgi?id=23982

PLH: 23982 focus ring out of canvas is that one closed?

JMann: I don't think the bug reflects that, but we need to get the spec updated and then update the bug

<richardschwerdtfeger> It was agreed that the focus ring would be clipped to the bounds of the canvas area but that the location of the corresponding fallback element would not be clipped and would match the current default path passed.

Publication path/timeline for Canvas (from PaulC) http://j.mp/KfCUkH

<plh> https://www.w3.org/Bugs/Public/show_bug.cgi?id=23981

<paulc> Canvas CR bugs: http://tinyurl.com/kxmqysj

JMann: I'll have Jay close the bugs as soon as the spec is updated

<richardschwerdtfeger> "Informs the user of the canvas location for the fallback element, based on the current path. If the given element has focus, draws a focus ring around the current path following the platform or user agent conventions for focus rings as defined by the user agent."

PC: When will all the changes to the spec be made, and when will the bugs will be closed.
... next week we should have an agenda item to confirm these bugs are closed

RS: I can confirm that 23978 has been completed

JMann: scrollPathIntoView has been moved to L2 because it requires Path

<plh3> " the bounding box of the area to which the user agent is scrolling as the bounding box of the current selection."

<JatinderMann> https://www.w3.org/Bugs/Public/show_bug.cgi?id=23983

<Zakim> plh, you wanted to ask about a potential new issue

<plh3> http://www.w3.org/html/wg/drafts/2dcontext/html5_canvas_CR/#dom-context-2d-drawsystemfocusring

<richardschwerdtfeger> http://www.w3.org/html/wg/drafts/2dcontext/html5_canvas_CR/#dom-context-2d-drawsystemfocusring

PLH: If i have two checkbox in my canvas and two canvas elements to match, and I call drawFocusIfNeeded. Now, I move focus to the second element, second checkbox and I all drawFocusIfNeeded. I am assuming it does not remove the focus once focus is removed it is up to the author to remove the focus ring


RS: also when author calls .focus

JMann: is dominic in agreement with all of our changes?

RS: I think we should send him an updated draft

JMann: I will follow up with Jay to get all changes implemented so we have a new draft.

PC: I talked to W3C Team and as long as we are removing material to canvas and not adding substantive new material, we can waive the disclosure period, so that is no longer a factor.

PLH: we have two problems. We have added some text to this spec. Also, with changing the name of the method...

PC: PLH can you add this to the discussion with Ian Jacobs?


PC: doesn't look like we will have much of a delay going into LC. If we go back into CR what the min time will be. If another disclosure is required, the min length would match the 6 week period anyway, so we should be OK.
... main thing is that we need to get this into LC ASAP

RS: when we get this draft, there is an apple guy that needs to look at it as well.

RC: I think we have addressed his concerns as well. He may need to be convinced.

JMann: will be worth it to share our changes and reasoning with him.

RC: I'll email him.

JMann: how long to get changes in?

RC: Should be a small patch, the AAPI stuff might take a few weeks.

PLH: Need dominic to update implementation in Blink

JMann: we also need test suites as well

PLH: I have some tests, but nothing that looks at the AAPI

<plh3> https://github.com/w3c/web-platform-tests/pull/457

PC: During this week, if the canvas editors are closing out bugs, they should be prepping LC document as well

<plh3> ACTION: plh to update his tests in https://github.com/w3c/web-platform-tests/pull/457 [recorded in http://www.w3.org/2014/01/13-html-a11y-minutes.html#action01]

<trackbot> Created ACTION-225 - Update his tests in https://github.com/w3c/web-platform-tests/pull/457 [on Philippe Le H├ęgaret - due 2014-01-21].

PC: we have to do a CfC, earlier the better
... probably what we would do is have a heartbeat for L2 as well
... we will be doing heartbeats in early Feb

Next Meeting

MS: same time, same place next week

Summary of Action Items

[NEW] ACTION: plh to update his tests in https://github.com/w3c/web-platform-tests/pull/457 [recorded in http://www.w3.org/2014/01/13-html-a11y-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2014-01-14 15:28:46 $