See also: IRC log
<Andi> Scribe: Andi
http://www.w3.org/WAI/PF/Group/comments/details?comment_id=203
<cyns> can you put it in again
Most AT uses the DOM, some AT uses the accessibility API
Desktop applications remove the children. User agents need to behave the same way.
<Stevef> voiceover uses accessibility API exclusively (this came from maciej), no DOM access
CS: image with alt="", treat it as graphic with emptry string as name
http://www.w3.org/WAI/PF/aria-implementation/#mapping_general
<Stevef> Andi: OK, thanks
DB: in FF, if table with role presentation, don't expose unless focusable, has onclick handler, has another role, etc.
<Stevef> i think structural descendents should be removed example <table role="presentation"> should remove child table elements from the heirarchy, otherwise it will have to be applied to every element in the table
http://www.w3.org/WAI/PF/Group/comments/details?comment_id=203
<davidb> right Stevef
<davidb> and they are removed unless further overriden with another aria role
<davidb> or if they are focusable
<davidb> then we keep them
<Stevef> yes
CS: the text will remain but the roles and structure of the children will not be exposed.
<Stevef> what about other elements within table cells such as <p> or <hx>
<Stevef> it would be better not to remove these, but is there a relaible way to romve on table related elements for example?
CS: what happens if there is a
labelledby?
... example: progress bar that is a table
... one of children is text
... if parent uses labelledby that points to the child, does
this work?
... it should
<Stevef> CS: agree
DB: role=presentation - just
talking about the element that has that role
... if <div> with role=presentation, doesn't mean nested
<div>s are presenational
... tables are special
CS: if remove table, having cells sitting there without table doesn't make sense
TB: if leave the children, there wouldn't be any context
<Stevef> suggest tables and lists should be special cases
CS: Stevef: maybe
... maybe the only special cases in HTML but there may be other
special cases in other languages
TB: what about mathml?
CS: seems like lists and tables are places where the children without the parent are invalid
DB: structural elements that have related structural child elements, the child elements inherit role=presentation
CS: example, table is presentational, so are cells
DB: unless overridden by explicit roles in the children
<davidb> DB: or focusable
descendents that are text content are not removed from the accessible tree
<davidb> Stevef, when creating the accessible tree, FF skips over table structure if within a table role="presentation", but includes everything else otherwise. (with exceptions)
<AndiSnow> http://www.w3.org/WAI/PF/aria-implementation/#mapping_general
<AndiSnow> scribe: AndiSnow
http://www.w3.org/WAI/PF/aria-implementation/#mapping_general
<Stevef> davidbL cool!
<Stevef> CS: yes
DB: if click handler is on ancestor but the child is what gets clicked on, how do you determine that you need to expose the child?
CS: pattern I was talking about,
click handler on document
... if element is button, call function
DB: or certain ID, class, etc. -
can get pretty hairy
... can't cover that case
CS: probably need a WCAG
technique or authoring practice technique or both
... shouldn't put the click handler on the ancestor
DB: talk about in aria-activedescendent as well
<Andi> scribe: Andi
CS: bug - need WCAG and/or ARIA techniques
assign to Steve
CS: Gez would be another
option
... if have click handler that's only on the ancestor,
shouldn't be using inherited presentation
DB: if put ARIA role on object
that is a button, handler can be on the button or the parent,
don't care
... but if there is a click handler on a parent, and there is
no aria role on the child, that's a problem
CS: a click handler on an element that doesn't have either a native interactive role or an ARIA role that makes it interactive is an error?
DB: not sure about that
CS: the thing the user is
supposed to be clicking on either needs to be a button or have
a role that makes it interactive
... if have click handler on parent that contains things users
can click on but they don't have interactive roles, then it's
an error
DB: also wondering about
tabindex
... if element has tabindex="-1", wouldn't we expose that?
Section 3.7 Actions, add "the element is focusable"
<Stevef> tahnsk andi...
<Stevef> if somebody asks gez to join PF (like janina) he will and he will help
<davidb> Stevef, \o/
starting with "main"
<Stevef> Andi: is the above the correct link?
http://www.w3.org/WAI/PF/aria-implementation/#mapping_role
main is a landmark role
use AriaRole for UIA and object attributes for IA2 and ATK
marquee
DB: should we use ROLE_SYSTEM_ANIMATION for MSAA? waiting for decision to implement in FF
David to send e-mail to Cynthia about that
<davidb> DB: sent
<Stevef> note: AXGroup, AXSubrole: "AXLandmarkMain", AXRoleDescription: "main" on Mac
menu: okay
menubar: okay
<Stevef> latest Mac tests here : http://www.paciellogroup.com/blog/aria-tests/ARIA-Safari-RoleTests280909.html
option
CS: wouldn't be surprised if
Apple is mapping "select" to menu
... Apple: "select" with multiples is a listbox, "select"
without multiples is a menu
bug - CS - synchronize ARIA mapping of options with HTML 5 mapping of options, specifically how it relates to HTML 5 commands
DB: should "timer" map to
ROLE_SYSTEM_CLOCK? send e-mail to CS
... FF has history of breaking rules wrt MSAA, would like to
keep doing that for backward compatibility
CS: tools flag as error
DB: only way to give access when there was nothing but MSAA
DB to send CS an e-mail, explain what doing, why important, etc.
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) Found Scribe: Andi Inferring ScribeNick: Andi Found Scribe: AndiSnow Inferring ScribeNick: AndiSnow Found Scribe: Andi Inferring ScribeNick: Andi Scribes: Andi, AndiSnow ScribeNicks: Andi, AndiSnow Default Present: Andi_Snow_Weaver, Cynthia_Shelly, David_Bolter, Tim_Boland Present: Andi_Snow_Weaver Cynthia_Shelly David_Bolter Tim_Boland WARNING: No meeting title found! You should specify the meeting title like this: <dbooth> Meeting: Weekly Baking Club Meeting WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Got date from IRC log name: 20 Oct 2009 Guessing minutes URL: http://www.w3.org/2009/10/20-aapi-minutes.html People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]