<davidb> Andi: use logs instead of minutes

<davidb> and world-visible

<davidb> instead of public

as-built browser table

http://www.w3.org/WAI/PF/wiki/HTML_4.x_as-built_browser_comparison

CS: plan is to look at what browsers do now with HTML elements, find inconsistencies and gaps, move towards normative mapping reference
... for example, everyone gives <a> elements role of link but nobody gives <area> elements a role of link

<davidb> I think gecko exposes aria role="link" as link

CS: <a href> - most expose as "link". Opera exposes as "text" which seems incorrect.
... <area> not exposed to MSAA - seems like should be exposed as link
... <area> name should come from @alt
... <area> - default action should be "jump"
... <br> Gecko maps to white space

DB: looking at code, looks like Gecko exposes <area> as link

CS: should we have SHOULDS and MUSTs in this document?
... mapping <br> to white space seems like a SHOULD
... in IE, title maps to name on a lot of stuff

DB: may need to augment name computation algorithm in UAIG

CS: IE treats <div> as grouping - everybody else treats as text
... <div> is supposed to be a grouping
... having <div> be a grouping is useful for ATs to be able to navigate by <divs>
... but does making <div> be a grouping make grouping navigation too noisy?

DB: in Gecko, gets exposed through sub-API

CS: question about <div> - go by spec and treat as section or ignore it?
... is anyone relying on it being "grouping"?

DB: probably worked around what the implementations are doing

CS: AT doesn't seem to do much with "grouping"
... think new document section things should be mapped to grouping
... fieldset mapped to grouping - seems right
... hr - Gecko maps to separator which seems right. nobody else does anything with it
... might make sense to start deprecating use of title as name

DB: ask authors to use aria-label

CS: hr to separator seems like a SHOULD
... does HTML5 still have separator? seems like most people don't use it anymore
... h1 - h6: not mapped by anybody - should be mapped to heading role
... name from children, no actions
... iframes are changing a lot in HTML 5, not sure how to handle
... todo - figure out how iframes work
... img - IE uses title if no alt but DOM-based AT doesn't see it

DB: Gecko - if null @alt, don't map at all, treat like ARIA role="presentation"

CS: only Gecko supports @longdesc - HTML5 trying to deprecate

CS; img usemap - map structure with area elements should be part of accessible tree

CS: input image - Gecko has as graphic

DB: Gecko does use title on img if it's there, not sure if it maps to name or description

CS: input image - IE maps to button, Gecko maps to graphic, Opera doesn't map

*** stopped minuting as Cynthia is taking notes ***

<cshelly> <a href> role=link, name like IE and FF

<cshelly> <area> role=link, name from alt, default action jump

<cshelly> <br> = whitespace (SHOULD)

<cshelly> <fieldset> = grouping, name from<legend>

<cshelly> <hr> separator (SHOULD?)

<cshelly> <h1>-<h6> Heading role, name from children

<cshelly> Usemap has children in tree?

<cshelly> <input image> role=button, name works like image, label

<cshelly> Label, legend, li, ul, ol like Gecko

<cshelly> Table name s/b from summary.

<cshelly> Textare name from label

<cshelly> TH is column head or row head

<cshelly> Span = text (not groupign)

<cshelly> now I will paste the rest

<cshelly> Open issue:Title mapping to name. IE does it. Can it be replaced by aria-label? Are there places where it's more important to keep than others? There is a fair bit of deployed content that uses it.

<cshelly> DIV: do we go by spec (HTML 4 semantics) and treat it as a section/group? Ignore it because it's lost its semantic through overuse? How does HTML 5 treat DIV? anybody relying on it being a grouping.

<cshelly> IFRAME: these are really different in HTML 5. Work item: look at what to do here. Can the name mapping be more intitive?

<cshelly> IMG ALT: Gecko and Opera both make <img alt=""> not in the MSAA tree. IE makes it graphic with empty string for name. Gecko/Opera approch seems better. Would that break deployed content?

<cshelly> Does multiselectable on a <select> change the roll?

<cshelly> textarea: check that IE gets default value from @value, that others update value based on user entered text

<cshelly> How should headers and scope an such impact MSAA?

<cshelly> Work item: document events in IE and ask Opera to do the same

<cshelly> Work item: How to handle elements that are obsolete in HTML 5 but mapped in browsers? Does it matter how widely used the technique is?

<cshelly> Work Item: add new tags from HTML 5 to table.

<cshelly> Next call 2 weeks from today.

<cshelly> zakim make minutes

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)

No ScribeNick specified.  Guessing ScribeNick: Andi
Inferring Scribes: Andi

WARNING: No "Present: ... " found!
Possibly Present: Andi CS DB DIV IFRAME cshelly davidb textarea
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy


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: 14 Aug 2009
Guessing minutes URL: http://www.w3.org/2009/08/14-aapi-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.