W3C home > Mailing lists > Public > public-canvas-api@w3.org > January to March 2011

minutes: Canvas Subteam weekly telecon 2011-02-21 [draft]

From: Gregory J. Rosmaita <oedipus@hicom.net>
Date: Mon, 21 Feb 2011 21:05:10 +0000
To: public-canvas-api@w3.org
Cc: david.bolter@gmail.com, dmazzoni@google.com, chuck@jumis.com, franko@microsoft.com, bhawkeslewis@googlemail.com, public-html-a11y@w3.org
Message-Id: <20110221210419.M57360@hicom.net>

minutes from the Canvas Subteam of the HTML Accessibility Task Force's 
21 February 2011 weekly meeting can be accessed as hypertext from:


as an IRC log from:


and as plain text following this announcement -- as usual, please log
any errors, omissions, clarifications and the like by replyiing-to 
this announcement on-list...

note that the following RESOLUTION was agreed to at today's Canvas

   RESOLUTION: Submit 
   Change Proposal to HTML WG as it has been available for review 
   and individuals have been asked to review it specifically


                               - DRAFT -

             HTML Accessibility Task Force Teleconference

21 Feb 2011


   See also: IRC log - http://www.w3.org/2011/02/21-html-a11y-irc


          Chuck_Pritchard, Rich, Gregory_Rosmaita




     * Topics
         1. Getting Started/Updates
         2. Outstanding Bugs
         3. Strategy for positioning fallback content relative to
            corresponding canvas UI drawing counterparts.
     * Summary of Action Items

   <trackbot> Date: 21 February 2011

   <richardschwerdtfe> Meeting: HTML Canvas Accessibility Subteam

   <scribe> scribe: Gregory_Rosmaita

   <scribe> scribenick: oedipus

   Google Web Toolkit (GWT) 2.2 released this week now includes support
   for the HTML5 Canvas tag, which enables building "dynamic
   images/animations/transitions all in HTML with GWT."

   <richardschwerdtfe> Agenda is here:


Getting Started/Updates

   RS: if FrankO doesn't join, will submit proposal as last circulated
   to the list

   <Downchuck> http://mugtug.com/darkroom/

   RS: chuck, do you have an example?

   CP: still bug-fixing

   canvas 2d API change proposal cover letter:

   canvas 2d API change proposal:




Outstanding Bugs


   RS: pixel resolutoin -- what happened with that?

   CP: hixie and mozilla against that kind of work -- is a problem --
   MS already exposes it
   ... hixie handed over to CSS WG where it sits

   RS: what to do?

   CP: "slip in" via CSS
   ... for FF can do collectoin of media queries
   ... problem because vetoed by mozilla -- do have a workaround, but
   don't know what to do other than to bring to webkit'

   RS: close this bug?

   CP: assigned to CSS OM
   ... CSS resopnse "backlog is huge" -- sitting for at least a month

   resize event during zoom:

   CP: everyone in agreement -- needs to be added to specs


   <Downchuck> #canvas.CompatibilityMozScreen { width: 1px; } @media
   all and (min--moz-device-pixel-ratio: .3) {
   #canvas.CompatibilityMozScreen { width: .3px; } }

   <Downchuck> that's the trick for Firefox zoom

   CP: asignee in bug tracker, don't need to do ourselves


   RS: addressed in latest drafts and change proposal
   ... in change proposal under "Positive Effects"
   ... text metrics in Canvas 2D API

   CP: setCaretSelection API -- if selection not in focus, still works

   RS: changed that to accomodate that

Strategy for positioning fallback content relative to corresponding
canvas UI drawing counterparts.

   RS: will take a lot longer than we have -- create issue for LC next
   step -- either accept or not
   ... strategies for positioning content
   ... first, imagemap provides coordinate info -- can do that or use
   javascript according to FrankO

   CP: imagemap proposal been struck down

   RS: not doing imagemap, but interesting that elements with imagemap
   can take coordinates, so can fallback content for CANVAS provide
   coordinates through fallback

   CP: might be able to track positions via jscript object or can stick
   directly into DOM -- doesn't harm anything
   ... may have an effect --- no one uses fallback content to toggle on
   and off -- not available, but supposed to be a long-term option

   RS: reasonable strategy for doing this? get width and height, right?

   CP: yes, full CSS spectrum of adding info to elements
   ... keeping eye out for use case for exposing all of that data --

   \CP: not addressed yet in CSS Transformations

   CP: CSS Transformations don't provide AT with updates that screen
   has moved; if canvas animated, as soon as move over few pixels, have
   to update a lot of elements -- complexity we need to keep in mind

   RS: imagemap -- coordinates relative to context they are in?

   CP: yes

   RS: if UA moved, UA would update coordinates; app author only has to
   worry to extent that are visible on screen
   ... can't scroll and imagemap

   CP: could get same using SVG to define all shapes and z-index in
   relation to each other
   ... allowed to use transparent SVG elements to do these things (get
   a whole lot of data to AT) -- use DrawFocusRing and CaretSelection
   and telling it only what caret is on

   RS: could use x,y coordinate to move magnifier to whereever one
   wants -- not tied to caret in A11y API -- could position wherever
   one wants -- zoom to drawning object, put caret there
   ... second issue: what other ways can we set positioning other than
   x,y coordinates -- javascript?

   CP: not sure what FrankO means just by "javascript"

   RS: can do via CSS

   CP: may be referring to CSS absolute positioning
   ... in demo i have on darkroom site, am able to use our semantics
   (DrawFocusRing and CaretSelection) -- have TTS via google translate
   API running -- alternating via Mac and PC so can recreate each step
   on each platform -- works as well in my browser as in their apps --
   when scroll, some events don't get set up the way they should so
   accessibiolity focus ring gets stuck until type something in

   RS: if scroll window, position moves -- drawFocusRing around object
   is necessary step for retaining focus when scrolled

   CP: doesn't actually update where FocusRingSelection is on element
   -- does update in some cases, but only if change between 2 elements
   ... if send focus event on same element, will ignore repeated calls
   to update x,y position
   ... using our semantics, trying to get same result as apple

   note: CP referring to http://mugtug.com/darkroom/


   RS: add something to references in Chanqe Proposal so that
   http://mugtug.com/darkroom/ cited
   ... what to call this? canvas focus tracking simulation?

   CP: good -- focus management too

   RS: put under "development"

   CP: fine with me

   RS: now under "references" in change proposal

   CP: JAWS doesn't work as well in IE9

   GJR: RC beta of IE9 very sporadic with speech--problem updating info
   when switch tab to tab; forms controls often don't respond to
   keyboard input, but need routing of JAWS cursor to PC cursor, etc.
   -- very unstable

   CP: shift-based part necessary at moment -- every UA has different
   method to get focus
   ... working through a lot of quirks to get same result on the
   software combos we discussed

   GJR: am running FF so can use Chatzilla so can't check using IE9
   right now -- will do when CP alerts us that bug-fixing de-quirking
   done for testing

   <Downchuck> I'm using "shift+space" to turn a11y on the app at the

   <Downchuck> yes, will dequirk


   RS: propose we adjourn

   CP: ok

   GJR: ok

   ISSUE-131: Should we add a caret location API to canvas, or is the
   focus API sufficient?

   GJR: we have done our due dilligence in notifiying others

   RS: barring no objections at this time from Microsoft or Mozilla
   does Canvas Subgroup agree to submit change proposal for HTML WG

   CP: no objection

   GJR: no objection

   proposed RESOLUTION: Submit
   Change Proposal to HTML WG as has been available for review

   Change Proposal to HTML WG as it has been available for review and
   individuals have been asked to review it specifically

   RS: will send email to Sam Ruby, Paul Cotton and Maciej to alert
   that change proposal for HTML WG Issue 131 complete and vetted by


Summary of Action Items

   [End of minutes]
Received on Monday, 21 February 2011 21:05:49 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:31:52 UTC