See also: IRC log
<scribe> Scribe: Andi
davidb
DB: reviewed events.doc
CS: have not reviewed it with the
IE team
... HTML element types that support certain events, ARIA
semantics that map to similar HTML elements should support the
same events
DB: modifying the DOM
events?
... don't like the premise of modifying DOM events
... think ARIA's place is for desktop events
CS: don't necessarily have to
change DOM events but ...
... DOM events create OS events
... look at onchange or onsubmit events, some get mapped up to
OS level, not mapped to API events
DB: if <div> has
role=combobox, selection via ARIA attribute, need to fire
event....
... Firefox goes right to platform event
CS: when semantics of element
have been modified by ARIA such that they are the same as HTML
element...
... if the corresponding HTML element would throw platform
events, then the ARIA element should fire the same events
... third table in events.doc is the one that is like what we
need in the UAIG
DB: would be good to know if Safari is firing DOM events
CS: need to compare this to the tables we have and see if we have covered everything
DB: something that someone else could do - go through various attribute changes and see what the implementations do in terms of firing desktop events
CS: what do ARIA test suites test?
MC: still in planning stage, started wiki page for test planning
<cyns> I'm in
<MichaelC> ARIA test plan
MC: need some unit tests for what
UAs do for a given role and properties - several hundred unit
tests
... also need to do feature tests with combinations of roles,
states, properties working together
... don't know if this is required for implementation or not
but certainly required for interoperability
... might be room for adding suggestions to wiki page
DB: suggest adding events
MC: going to add a third category
for dynamic tests
... plan is to be able to import these into a test tool
DB: FF runs on three platforms so have internal event model
MC: adding to the wiki page: "reaction to events (fairly UA specific, need to expand on this)"
DB: testing would be UA specific but want to aim for interoperability
CS: many will have to be
operating system specific - something happening outside the
browser that AT has to react to.
... ARIA is a bridge from Web application to platform
implementation
MC: hoping to generate using XSLT, Jon G. has tests, want to automate what we can but there will be lots of human created tests
DB: created FF tests in Dojo
MC: hard to focus on test suite
until test harness is done
... for human created tests, need a way to ensure they are
valid
CS: writing a spec for each test, could be done without the test harness
MC: send note to Rich - UAI TF has thoughts about testing
<scribe> ACTION: Cynthia to send note to Rich about prioritizing testing on Monday's agenda [recorded in http://www.w3.org/2010/07/13-aapi-minutes.html#action01]
CS: section 4.7 rules
... element is focusable - doesn't necessarily mean there are
actions
... why are menus treated differently? do any of the roles in
the first table in 4.7 need this level of description?
DB: nice keeping the menu things in their own space
CS: we have a bunch of things that are separate one off things - the menu section is one of the last ones of those
DB: what does the actions column mean?
CS: aria states/properties that
cause elements to have additional actions associated with them
in the browser
... doDefaultAction on a button clicks
... <div> with aria-expanded, doDefaultAction comes in
from OS, what should the UA do?
DB: in FF, doDefaultAction simulates a mouse click
CS: maybe the format we use for menu events is a better way to describe the actions tables
DB: Javascript is between user
action and event
... browser has to fire event if state (aria-expanded)
changes
... Javascript is handling actions
CS: if AT simulates arrow behavior...
DB: AT should let user keystrokes go through
CS: but what about things that simulate a keyboard - voice reco or onscreen keyboard
DB: they just synthesize the keystrokes
CS: for UA, keystrokes and mouse clicks need to get passed through
DB: for 2nd table in 4.7, specifying the default action might be useful, might depend on the role
CS: most are going to be
click
... grabbed is interesting, exposes "grab" action
... back to bullet list at top of 4.7
... what actions get exposed when these are true
... just because something is focusable, doesn't mean there are
actions associated with it
... click handler is obvious
DB: if it's in the tab order, FF exposes click as the default action
remove the xlink bullet
CS: maybe we need a table of user
events
... but is that ARIA?
DB: more like HTML
CS: some discussion about moving
authoring advice out of HTML spec and into a separate
document
... been talking about doing a mapping document for HTML to
accessibility APIs
CS: actions and events topic could be given to someone else
9363 could be given to someone else
DB: could work with Joseph on events/actions bugs that are assigned to Cynthia
This is scribe.perl Revision: 1.135 of Date: 2009/03/02 03:52:20 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/davidb// Succeeded: s/1 min// Succeeded: s/primarily for the HTML/for desktop events/ Succeeded: s/exposed/exposes/ Found Scribe: Andi Inferring ScribeNick: Andi Default Present: Andi_Snow-Weaver, Cynthia_Shelly, David_Bolter, Michael_Cooper Present: Andi_Snow-Weaver Cynthia_Shelly David_Bolter Michael_Cooper Got date from IRC log name: 13 Jul 2010 Guessing minutes URL: http://www.w3.org/2010/07/13-aapi-minutes.html People with action items: cynthia WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]