W3C

- DRAFT -

User Agent Accessibility Guidelines Working Group Teleconference

01 Oct 2009

See also: IRC log

Attendees

Present
Greg, Jan, Jeanne, Kelly, Simon
Regrets
Jim, Mark, Kim, David, Henny
Chair
Jim_Allan
Scribe
Jan, jeanne

Contents


 

 

<trackbot> Date: 01 October 2009

<kford> - UAAG10 Techniques

<kford> http://www.w3.org/TR/2002/NOTE-UAAG10-TECHS-20021217/

<kford> - ATAG Techniques

<kford> http://www.w3.org/WAI/AU/2009/ED-ATAG20-TECHS-20090814/#gl-Web-based-accessi

<kford> ble

<kford> - WCAG20 Techniques http://www.w3.org/TR/WCAG20-TECHS/

Logistics (Regrets, agenda requests, comments)?

Meeting survey http://www.w3.org/2002/09/wbs/36791/20090930/

<Greg> We'll discuss the first item, "proposal for adding test conditions to the Definition of User Agent".

<Greg> Kim voted accept, Kelly was neutral, Jan had one suggested change, and Greg had three questions.

<Greg> If XML is stored as a text file, read by a text editor and rendered as text for the user, would that trigger part 1 because it is "using…a w3c specified technology" (XML)?

<Greg> Simon says it should be limited to when the w3c technology is used to present or render content.

<kford> Greg and Simon discussing issue.

<kford> Simon: Javascript isn't a W3C technology.

<kford> Simon: If the user interface is generated using HTML that's fine.

<Jan> JR: THinks saying W3C tech is not necessary - since what about Flash etc

<Greg> Greg suggests that the older wording, although longer, was easier for him to understand. Simon replied perhaps we should keep 1 and 2 from the previous version and replace 3 with the two new lines.

<Greg> Simon noted that in a previous conference call people seemed to find the older, original proposal more confusing, leading him to create the new proposed wording.

<sharper> 1) If the user interface is generated by the interpretation of either

<sharper> a procedural or declaritive language; and

<sharper> 2) If this interpretation is by a Primary User Agent, User Agent

<sharper> Extension or Plug-In; and

<sharper> 3) If the generated interface, intentionally or unintentionally, hides its interaction from the technology used to generate it.

<Greg> Greg says, as per my written comment, that I find the term "technology used to generate" much more ambiguous than the earlier terminology of "the primary user agent, user agent extension or plug-in".

<sharper> Now it seems that from my mail:

<sharper> http://lists.w3.org/Archives/Public/w3c-wai-ua/2009JulSep/0082.html

<sharper> It is possible to capture keystrokes and not pass them on AND

<sharper> make changes to the DOM without forcing a DOMMutationEvent.

<sharper> I now want to get this sorted so I can move on to my other 2

<sharper> (dependant) action items. So I'd like to make another proposal for

<sharper> part 3, thus:

<kford> Discussion continues.

<kford> Greg: Revisitng comments from a 9/3 meeting around this.

<Greg> On the 9/3 conference call we disucssed the possibility of using more general guidance for purchasing agencies as to what would need to comply.

<Greg> Simon says we keep revisiting this through the history of the project, people questioning what would be a UA. He finds it's difficult to word SC without knowing what components would need to comply with them.

<Jan> scribe: Jan

SH: Let's take eg web app created in Javascript
... Going to create interface using JS
... Or they have created interface in HTML but program logic and screen updates using JS
... I was trying to say if they capture inpout and don't pass it on to the rendering...just use internally
... Or they decide they are not going to put things into the DOM....
... Or they decide they will remove DOM event mutation listeners...
... Will cause updates of screen but not DOM...so
... Assitive techs won't know there are changes
... So...if they are doing this...they are pretending to be a user agent...so they need to conform to our guidelines.
... They are a user agent that needs to communicate independently with ATs
... But if they are passing things along properly to the accessibility APIs and DOM they are actually web content and only WCAG would apply

KF: New recent release - Chrome plugin that runs inside IE....
... Like Chrome browser inside IE frame

SH: L Not covered by part 3...but is covered by part 2...
... which is UA extension or UA plugin...
... And should conform
... So that example should conform
... But what I'm tryoing to get to with Part 3...web app not passing info must conform too

JS: What may help is an an example

SH: OK give me an example
... Chrome plugin example already done

JS: Want example to be vendor neutral

SH: e.g. Silverlight, Adobe Air
... Both fall into the extensions/plugin category

GL: Are there things we want to exclude?

SH: I've tried to think of all the possibilities

KF: So we are trying to think of something that wouldn't

GL: Right - so e.g., text editor used to edit HTML would not be a user agent

SH: first thing...is a stand alone app
... But doesn't interprt technology

GL: Highlighting syntax?

SH: That's a diff thing
... THen isn't a extension or plugin
... Then the last part is the hiding of input/output

GL: OK so take Adobe Acrobat reader plugin
... It's Part 2

<jeanne> scribe: jeanne

<Jan> SH: If new PDF window it's outside of UAAG if it's in viewport its coverd

JR: That example, practically, if IE was making a conformance claim, they would not include pdf, it would be up to Adobe to make a conformance claim for pdf.
... If we take the example of the text editor, take an authoring tool like Dreamweaver. The difference between user Agents and Authoring Tools is rather arbitrary for WAI to organize the documents. We should have a note that refers people to the authoring tool guidelines.
... If they were producing HTML, they would be covered by ATAG. If you are web based and your UI is displayed by a browser, you have to follow WCAG. If you aren't web based, then you need to find out the platform accessibility requirements and follow them.

<Jan> http://www.w3.org/WAI/AU/2009/ED-ATAG20-20090909/#def-Authoring-Tool

Any software, or collection of software components, that authors can use to create or modify web content for use by other people.

JR: For use by other people, doesn't refer to things like Firebug, because it is not changing the experience for other people.

SH: The authoring tool is very clear that it is for authoring. User agent is less precise.

JR: Where that matters is with the accessibility APIs. We should be encouraging standard APIs.
... it's got to communicate down through the user agent in a standard way.

SH: HTML5 Canvas worries me. Something could be drawn there that looks like a user agent but is really a graphic that doesn't update the DOM.
... Looking at the HTML5 specification, many things could be implemented that look like browsers and user agents in a really superficial way that doesn't pass information tothe DOM.

JR: But that would mean that there is a gaping hole in WCAG. Is there?

SH: I'm trying to future-proof it.

JR: The line between authoring tools and user agents is pretty fine.

SH: I think this is far simpler when it is applied. If someone takes a javascript form, is that a web application? How do we define web application? Or is it just a form?

JR: Features have been added to meet UAAG. If this application is running in the base browser, and the web application is hiding the information that the UAAG features need, then those features need to be implemented on the addin side.

<Jan> JR: Proposed test of whether a web application is also a user agent: Given a user agent conforming to uaag (ie with "uaag features"), if a web app is running/rendered by the user agent and the web app hides information required for the uaag features to operate, the web app is a user agent and needs to replicate the uaag features

<Greg> Looks pretty good but may be better to say "the web apps needs to" rather than "the web pass is a user agent and needs to".

<Jan> e.g., an app that takes SVG, converts them to gifs and sends down to browser as HTML+gifs

<sharper> If the following condition is met then it is a Web-based User Agent and Must Conform to UAAG:

<sharper> 1) if a web app is running/rendered by the user agent and the web app hides information required for the uaag features to operate

<Greg> If it's on the client, the SVG viewer would count as UA under Simon's second definition, since it renders within the browser window.

<sharper> GL: Needs to be evaluated in the context within which it runs.

<kford> ation: Simon to update User Agent test.

<kford> ACTION: Simon to update user agent test. [recorded in http://www.w3.org/2009/10/01-ua-minutes.html#action01]

<trackbot> Sorry, amibiguous username (more than one match) - Simon

<trackbot> Try using a different identifier, such as family name or username (eg. sharper, spieters)

<sharper> ACTION: sharper to update User Agent Wording [recorded in http://www.w3.org/2009/10/01-ua-minutes.html#action02]

<trackbot> Created ACTION-234 - Update User Agent Wording [on Simon Harper - due 2009-10-08].

Summary of Action Items

[NEW] ACTION: sharper to update User Agent Wording [recorded in http://www.w3.org/2009/10/01-ua-minutes.html#action02]
[NEW] ACTION: Simon to update user agent test. [recorded in http://www.w3.org/2009/10/01-ua-minutes.html#action01]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2009/10/01 18:35:20 $

Scribe.perl diagnostic output

[Delete this section before finalizing the 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)

Found Scribe: Jan
Inferring ScribeNick: Jan
Found Scribe: jeanne
Inferring ScribeNick: jeanne
Scribes: Jan, jeanne
ScribeNicks: Jan, jeanne
Default Present: kford, Jan, Jeanne, sharper, +1.425.895.aaaa, Greg
Present: Greg Jan Jeanne Kelly Simon
Regrets: Jim Mark Kim David Henny
Found Date: 01 Oct 2009
Guessing minutes URL: http://www.w3.org/2009/10/01-ua-minutes.html
People with action items: sharper simon

[End of scribe.perl diagnostic output]