User Agent Accessibility Guidelines Working Group Teleconference

06 Nov 2009

See also: IRC log


Kelly_Ford, Kim_Patch, Jim_Allan, Mark_Hakkenin, Jeanne_Spellman, observing-Michael_Cooper, Shadi, Andrew, Sally
Kelly, Jim
kford, jeanne




<trackbot> Date: 06 November 2009

<jeanne> http://www.w3.org/WAI/UA/2009/ED-UAAG20-20091105/

<kford> Scribe: kford

Kim Patch giving demonstration of browsing by voice.

Kim: I'm closing some windows.

Enters some voice commands.

Kim: I can combine things.

Says "w3c site".

Browser launches to W3C page.

Kim dictates a few more sites and pages open in new tabs.

Demonstrates how she adds a new site by saying add site, giving site name.

Kim: I'll show you a little bit of getting around a site.

says 2 tab back, 1 tab forward and releated commands.

Demonstrates link activation with a firefox extension that numbers links. Dictate link number and link is activated.

Trying test on a twitter page example.

Continue demo, look at example on http://twitter.com/goodies/widgets where links are not actual links.

Kimdemonstrates feature of her application where you can add coordinates of the screen to be activated by voice.

Introductions all around. Shadi and Andrew are members of other W3C groups.

Kim demonstrates more touch click actions, where multiple touch actions are combined.

Demonstration of form completion with speech.

MC: I work on PF and HTML. HTML 5 group has become large. Much communication on e-mail. Accessibility issues very personal but others lack some context.
... Some mishearing of each other.
... On the other hand a lot of accessibility in HTML5 because it is the language of the web.
... There is an opportunity to address issues that were never addressed in 4.
... Also many new features that need accessibility attention.
... We've created an HTML 5 task force to help here comprised of PF and HTML5 people. Opportunity to build credibility.
... As of yesterday has leaders from both groups.
... Hope to have a call for participation today or Monday.

Will have to be a formal member of the HTML 5 group.

<MichaelC> --> http://www.w3.org/WAI/PF/html-task-force HTML Accessibility Task Force

<MichaelC> --> http://www.w3.org/WAI/PF/wiki/HTML_Task_Force HTML Accessibility Task Force wiki

<MichaelC> --> http://www.w3.org/Bugs/Public/buglist.cgi?product=HTML+WG&keywords=a11y HTML accessibility issues

<jallan> KF: How can UAWG help?

<jallan> ...HTML5 seems to be close to a UA spec. should we work through PF?

<jallan> MC: html5 is monolithic. covering many areas. some advocate for splitting into specific areas (e.g. authoring, etc).

MC: HTML 5 is really a lot of content.

<jallan> ...no suggestion for breaking out UA specs

Authoring, UAAG and more.

Some discussion of splitting things out into separate specs.

MC: Don't think HTML5 would be objectionable but probably have to come with an editor.

<jallan> objectionable to UA split out without an editor

More discussion about how UAAG and HTML5 work together.

MC: UAAG and PF need to work more closely.

<jallan> should UA have a rep on the TaskForce

<jallan> MC: Yes. need UA perspective

<jallan> ...good synergy to find things missing in own spec when working on other specs

More talk about HTML5, ARIA and general coordination issues.

MC: End goal is that the host language drives accessibility, ARIA used when host language doesn't do enough.

Talk about ARIA, HTML5 and tool kits. Important to get these to do things right.

Talk about the black hole of Javascript.

What can we do at the script level to drive accessibility.

possible wai coor items.

uaag and html.

accessibility api for script.

PF coordination with UAAG.

sharing of technical knowledge.


<mth> fyi ... the video group is talking about codecs at present. not a11y

<jeanne> shadi: ERT works on the the language to exchange information about evaluation and tools. Accessibility needs many different tools and EARL can bring those together.

<jeanne> ...EARL is meta data so it could be an annotation resource that browsers that can do things with: e.g. you can search for web pages that are keyboard accessible. there are use cases, we are monitoring the work being done with Dublin Core, but we aren't working actively on it.

<jeanne> ...WCAG Tools Test Sample Development. A test suite for WCAG.

<jeanne> ... there is some resistence to calling it a Test Suite, because "suite" implies a pass/fail, but Samples have a wider variety of possible answers.

<jeanne> ... it would provide a resource for tests with specific Techniques.

<jeanne> ... if there is a mapping betweeen WCAG techniques and UAAG techniques, where WCAG could say this is what the content looks like, ATAG could say this is what the author writes, and UAAG could say what the user should experience.

<jeanne> jeanne: Since UAAG will need to write a test suite for Candidate Recommendation, what advice could you give us

<jeanne> Shadi: it is a series of XML files. THis is a way you could make it machine testable. Like H37 may be alt text for the submit button. So if UAAG was testing for a submit button, they could use the WCAG files, but we could need more specific UAAG files.

<jeanne> ... but it could cause problems if WCAG changed their file, but that could be worked with

<shadi> http://www.w3.org/WAI/ER/tests/

<shadi> http://www.w3.org/WAI/ER/tests/WebInterface/

<jeanne> shadi: [shows the WCAG Test Samples pages]

<jeanne> scribe: jeanne

scribe: we haven't been able to publically publish the tests. We need some people to help us do the last mile to get it finished.

Andrew: The review process is good, but it doesn't scale to 10,000 tests, if a major company gave us 10,000 tests.
... it needs peer review. It takes about 20 minutes per test sample presently to accept it.

Michael: ARIA will have several hundred tests.

shadi: we are still doing quality review, let me know. [laughter]
... a test can only relate to one technique, so if it applies to several success criteria, then the test has to be copied.
... but reducing it to one technique, it simplifies the review of the test.


<andrewA> -> WAI-AGE project page http://www.w3.org/WAI/WAI-AGE/

Shadi: The WAI-AGE project is a W3C/WAI project funded by the European Commission. It compares how accessibility recommendations/requirements compare to what is needed for ageing.
... conclusions are drawn between accessibility and usability, that are not helpful overall -- like a conclusion that a font size should be 12 pt, rather than a font size should be flexible.

Andrew: Like the default font size should be the user's default font size.

Shadi: there is overlap between user agents and ageing needs. Like font size widgets are recommended because many users (and trainers) do not know how to increase default sizes in the browsers.
... more intuitive browser design is desireable, so it is easier to find the features

Andrew: many older people do not admit that they have a disability, because their impairment may have developed over years, or in combination with other problems.

Kelly: We recommend an accessibility preference wizard

Andrew: Should we even call it an accessibility wizard? Or just a customization wizard that includes accessibiliyt features. Many people do not want to think that they need accessibility.

Shadi: There are the necessary techniques for font size, but there are advisory techniques that are just good to do.

Kim: The ease of use label makes it more likely that people will find the customization features. It would be very useful to be able to see an outline of a page in a very tiny size and expand it again in a singe click for speech users.

Michael: The issue is that the user may have a preferred font size and the author has a different one, but the author defines all the relationships in his preferred font size. You can't override individual font sizes.
... if there was a CSS attribute for the user's default font size, that would help things considerably.

Andrew: There are a lot of usability things that have come out that are good for everyone, but they are included in WCAG because they are particularly userful for older people.


todays new master document: http://www.w3.org/WAI/UA/2009/ED-IMPLEMENTING-UAAG20-20091106/MasterUAAG20091106.html

<kford> Group looking at section 5 on understandability.

<kford> Jeanne shows example technique from yesterday on sequential navigation.

<kford> kford: How would we summarize main points about "understandable" in wcag and au?

<kford> MC: WCAG talks about language, words AT might not know how to pronounce, jargon, reading level.

<kford> MC: We talk about pronunciation. screen reader tough with content.

<kford> Mistake avoidance, predictable.

<kford> Jeanne: mistake correction, good documentation.

<kford> Jeanne recaps currnet UA 5.

<kford> Andrew: the one that I see missing that is in ATAG is about appropriate language. Don't assume everyone has a doctorate.

<kford> Kim: We've talked about predictable with focus but maybe we need to address further.

<kford> If you are expecting one thing and something else happens.

<kford> Shadi: how does this relate to some of the security work?

Kelly: Understandable and security are interesting topics. Back in IE7, we moved to the notification to the address bar and called it the trust badge. We worked with vendors so that when the badge appeared, the user could be confident that the page could be trusted.
... if you are in IE or FIrefox, and you click on the trust badge, the sighted user gets more information. The screenreader user gets a short phrase saying that the page is sucure.
... people do not understand what it is for, because it isn't used often. Making a user experience understandable, the feature has to change often enough so that the user sees that it has a function.

<kford> Let's walk through our 5 guidelines.


Michael: the ARIA reference isn't really appropriate there, it's a technique, not a guideline.
... it also is about text messages, while ARIA relates to live regions, changing text.

Kim: It's about control.

kelly: Control Interruptions. The user agent provides a mechanism for the user to determine when notifications can interrupt their workflow iwth notifications

Michael: WCAG uses " a mechanism is available" because sometimes the author does it sometimes the user agent does it

Kelly: A mechanism is available to allow the user to control what notifications will interrupt the user's workflow
... Intent: Users with disabilities are using software in ways that involve a lot of input to the computer. Anything that interrupts that context and cause trouble for them to get back to where they were. The goal is to let the user decide when to be interrupted and to be taken off-task.

Example: A speech input user has started to dictate a series of commands for launching a web browser. The browser wants to let the user know that a software update is available. The browser has been configured to hold all those messages until the user asks the browser if the browser has any messages for them
... notifications that grab focus are not as big a problem for screen reader users -- we love dialog boxes. but devleopers are moving away from them, because they pull people out of context.
... another example: A switch user is clicking a single switch when focus goes to the tab key. The web browser blocks a popup so that the user can continue tabbing without losing focus

another example with an ARIA case: A user is browsing a web page to a web based chat client. The chat client is built using WAI-ARIA. The chat application allows the user to configure which chat messages are read automatically and interrupt other speech with the assistive technology.

scribe: correct A user is browsing a web page with a screenreader.

kim: when the user changes focus to an application is a really bad time to give notifications that interrupt for a speech user, because it yanks the focus away because the user is queuing up another operation of 5.1.
... the example of cut and paste between applications.


kelly: the problem I see with this guideline, I see that this is the only mistake (form submission) we have SC about.
... we have a bug with HTML that we need to undo DragnDrop.

Michael: how much is UAAG trying to be unique - since much of this is covered by WCAG.
... example of a form error

jeanne: this sounds like it is not a user agent issue.

Kelly: disagree. Supporting the ARIA invalid or the HTML "you screwed up" attribute. The implicit assumption is that the user agent supports that.

Michael: the need for marking the field in error, and the need for navigating to the element in error is different.

kelly: Support error notification from relevant technologies, and do an example of ARIA invalid.
... and WCAG
... did you really mean to fast forward and skip 30 minutes?

Michael: the html video element that fast forwards, so we need an api that looks for that error. Do we want to put that in UAAG, or leave it for HTML to instruct the browsers?

Kelly: do we do a overall message that the User Agents must meet other guidelines like ARIA, or do we want to draw out individual ARIA or WCAG guidelines where they are appropriate message.

Michael: The example of form submission not with the enter key isn't a violation of any guideline, it is a problem of the browser behavior in current practice.

kelly: the example of undo of drag and drop is a good example. It is very hard for the user agent to undo a DnD because all the user agent knows is that a script executed. It is easier for the authur to write the undo script.
... Most of the mail applications with cut and paste or DnD do it totally at a script level. This stuff is hard.
... what are the oh, shoot! moments as a speech user

Kim: losing focus, you want to be able to undo. If people know they can undo, they are more likely to try things.

Kelly: the back command is the undo here. Clicking on the wrong link is most common with older people.

kim: If you click on the wrong link when you ahve text somewhere, and lose the text, that is aproblem. But if the page is secure, you probably don't want to do that.

Andrew: couldn't you give an alert: are you sure you want to do that?"

Kim: that cuts both ways, because it adds extra steps.
... if you have a control of dialog boxes and set that level that you don't want to be interrupted, that would be a good thing.

issue: Are there other Undo or errors that UAAG needs to consider? Do we need new SC?

<trackbot> Created ISSUE-59 - Are there other Undo or errors that UAAG needs to consider? Do we need new SC? ; please complete additional details at http://www.w3.org/WAI/UA/tracker/issues/59/edit .


jeanne: reads current 5.3

Kim: keyboard shortcut section and discovering key commands should be a reference here.

[no discussion]

the examples are easy to create.

Michael - anywhere WCAG says " a mechanism is available" that would be a cue that UAAG needs to have a guideline.

<kford> ACTION: KFord to review wcag 2.0 for instances of a mechanism is available for places that UAAG needs to address. [recorded in http://www.w3.org/2009/11/06-ua-minutes.html#action01]

<trackbot> Sorry, couldn't find user - KFord

<scribe> ACTION: KF to review wcag 2.0 for instances of a mechanism is available for places that UAAG needs to address. [recorded in http://www.w3.org/2009/11/06-ua-minutes.html#action02]

<trackbot> Created ACTION-242 - Review wcag 2.0 for instances of a mechanism is available for places that UAAG needs to address. [on Kelly Ford - due 2009-11-13].

<andrewA> kelly: a mechanmism for setting global config as to whether UAs can set default focus is provided

<andrewA> ... users need to know that the nav on a page will satrt in a predictable location

<andrewA> ... recognising tha for a11y, stetting default to other etna the fist link, the user shuld be in control of tis behaviour

<andrewA> ... eg a user navigates to a page where the focus is set to the search box - affects scrolling, should be selectable globally

<andrewA> ... has to take additional actions to scrtoll to thecontent they really wanted to view

<andrewA> ... eg transit time for the ferry page, and focus is into searrch-box, you need to tab at least once to get to scroll to content you wanted

<andrewA> kim: problem when focus changes during an action

<andrewA> kelly: sounds like WCAG 2.0 3.2.5 Change on Request - redirection or change of context

<andrewA> michael: eg enterng postcodes and sometimes it moves on, sometimes not

<andrewA> ... with multi-part postcodes

<andrewA> kim: tricky as sometimes you wnat this, but not always

<andrewA> kelly: uninitiated focus change - ...

<andrewA> Jim: we alsready have a GL about that - 3.10.8

<andrewA> Jeanne: we should have it again here as theis about predictability, and reference the previous SC

<andrewA> kelly: the use cases we have here are about web content

<kford> Group picking up review of action items.

<kford> Action 174 updated to a December date.

<trackbot> Sorry, couldn't find user - 174

<kford> Action 182 completed.

<trackbot> Sorry, couldn't find user - 182

<kford> action 182 closed.

<trackbot> Sorry, couldn't find user - 182

<kford> closing action 182.

<kford> Giving action 183 to Mark for a December date.

<kford> Closing action 184.

<kford> Updatng action 188, leaving open need a good glossary.

<kford> Closing action 189 event handling langugage solid.

<kford> Closing 191, preferences seem pretty good.

<kford> Updating action 194 to a December date.

<kford> 4.5.4 Portable Preference Settings:The user can transfer preference settings across locations onto a compatible system. (Level AAA)

<kford> Closing 195 as done.

<kford> Updating action 198 to a November deadline, Jim to send Jeanne survey text since we have it ready.

<kford> Action 1999 update ot jan 2010.

<trackbot> Sorry, couldn't find user - 1999

<kford> update action 199 to jan 2010.

<kford> Keeping open, handling with other preference setting item.

<kford> That's 201.

<kford> Closing 202.

<kford> Closing 203.

<kford> Closing 207.

<kford> Close action 210

<kford> Update action 211 to a December date.

<kford> Closing 212.

<kford> UAAG links.

<kford> http://lists.w3.org/Archives/Public/w3c-wai-ua/2009JulSep/0091.html

<kford> Closing 214, 216, 221 UAAG subreviewed HTML 5. Good job everyone.

<kford> Closing 220.

<kford> Leave 222 open.

<kford> Updating several multimedia items to a new december data. Includes 226, 227,229, 231.

<kford> Close 230.

<kford> Assigning 234 to Jeanne.

<kford> Action 235 closed, bug 8187 filed against html 5 spec.

<trackbot> Sorry, couldn't find user - 235

<kford> closing action 235.

<SCain> Thank you for letting me observe and participate.

<kford> Jeanne: Right now our document format are in HTML and we are limited in the views we can create of them.

<kford> I'd love to try and move them into XML.

<scribe> ACTION: jeanne to follow up on the spec tool demonstrated at TPAC lightning talks by November 12 [recorded in http://www.w3.org/2009/11/06-ua-minutes.html#action03]

<trackbot> Created ACTION-243 - Follow up on the spec tool demonstrated at TPAC lightning talks by November 12 [on Jeanne Spellman - due 2009-11-13].

<kford> Video draft minutes.

<kford> http://ur1.ca/fa6o

<kford> +/me wow, anyone know much about this. Seems like some accessibility implications.

<kford> http://wiki.orbeon.com/forms/doc/developer-guide/xforms-refresh-events

<kford> MH giving examples of alert notifications in campus environments.

<kford> Recap of section 5 discussion.

<jallan> KF: what are things that make things understandable?

<jallan> MH: knowledge of operation of the application. some level of intuitiveness.

UNKNOWN_SPEAKER: highly symbolic for people with cognitive processing issues, procedural for people with memory issues.
... highly individual for people depending on their level of sophisticaion, technology being used. Symbology is a big issue across different cultures. It is very difficult thing to do.
... it is all about personalizing the UI.

<jallan> this is usability. a11y and usability are intertwined. what can to UA do to communicate, facilitate or discover the intuituveness of the functionality?

<andrew> -> Relationship Between Web Accessibility and Usability http://www.w3.org/WAI/EO/changelogs/cl-accessibility-n-usability an upcoming output from WAI-AGE and EOWG. Thoughts welcome

kelly: If I look at what I have today,we have document the accessibility, document the software, don't let me make a mistake. What if a web browser implemented a ribbon, what would they have to do to meet understanding.

mark: start with the simplest reduced set, then be able to add functionality.

Kim: keyboard shortcuts are not discoverable in the Office Word ribbon.

Kelly: maybe you need a tutorial of what the features are and wehre to find things

kim: the user agent needs to make it easy.

Jim: how can we measure "easy"? How can we use it.

Andrew: Make it easily discoverable, maybe a wizard, maybe a menu option.

<jallan> KF: if requirement to say explain your UI to me. can software do it.

<jallan> ... the software has all these features, heres houw you use them. pretty high standard

<jallan> Who is me? what does developer tell person x about their individual needs for using a product?

<jallan> KP: ...

<jallan> ... ability to configure the interface

<jallan> AA: sitemap 2.4.5 in wcag - multiple ways of finding information in a group of pages.

<jallan> KF: what if button push to give a list of all the functions available in the UA.

<jallan> KP: reference all configuration, options, etc in other guidelines in GL 5

<jallan> JS: audience of UAAG is developers that should know about user centered design

<jallan> ... need 2 slices

<jallan> .... use xml to sort by keyword for video, keyboard, etc.

<jallan> ... need multiple tags on each SC

<jallan> ... to ensure nuance and synergy between SC that may not be eveident to person who does not get a11y

<jallan> ... tag by who benefits blind, deaf, cognitive, noisy, etc

<jallan> JS: csun 09 presentation on cognitive interface

<jallan> feature bloat...start with a simple interface let the user add new features as needed.

<jallan> how does user know how to add a new feature, is there a button, etc

[discussion of interfaces and simplicity and how to transition to more sophisticated user.

<jallan> intelligence in the interface that watches usage and suggests features. how is user supposed to discover a feature that is not shown

Mark: need adaptive interfaces that see that you use a feature often, and prompts you to add a button.

<jallan> issue: need a requirement for out of the box installation present options for interface

<trackbot> Created ISSUE-60 - Need a requirement for out of the box installation present options for interface ; please complete additional details at http://www.w3.org/WAI/UA/tracker/issues/60/edit .

jeanne: When a software program is installed, the default configuration is a simplified user interface.

Mark: with an easy button for the advanced user to bypass it.

<kford> Let's do issues next time or something.

<kford> http://www.w3.org/WAI/UA/tracker/issues/open

<andrew> http://www.ehow.com/how_2303743_compete-bowling-blind-partially-sighted.html

blind bowling article that explains the rail. http://www.ehow.com/how_2303743_compete-bowling-blind-partially-sighted.html

Summary of Action Items

[NEW] ACTION: jeanne to follow up on the spec tool demonstrated at TPAC lightning talks by November 12 [recorded in http://www.w3.org/2009/11/06-ua-minutes.html#action03]
[NEW] ACTION: KF to review wcag 2.0 for instances of a mechanism is available for places that UAAG needs to address. [recorded in http://www.w3.org/2009/11/06-ua-minutes.html#action02]
[NEW] ACTION: KFord to review wcag 2.0 for instances of a mechanism is available for places that UAAG needs to address. [recorded in http://www.w3.org/2009/11/06-ua-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2009/11/07 00:59:19 $

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)

Succeeded: s/cack/backi/
Succeeded: s/backi/back/
Succeeded: s/partisipation/participation/
Succeeded: s/H37 may be the submit button/H37 may be alt text for the submit button/
Succeeded: s/Shadi, how doe/Shadi: how doe/
Succeeded: s/saerrch/searrch/
Succeeded: s/documents are somewhat primative./document format are in HTML and we are limited in the views we can create of them./
Found Scribe: kford
Inferring ScribeNick: kford
Found Scribe: jeanne
Inferring ScribeNick: jeanne
Scribes: kford, jeanne
ScribeNicks: kford, jeanne
Default Present: Kelly, Mark, Jim, Jeanne, Kim, Michael_Cooper(observing)
Present: Kelly_Ford Kim_Patch Jim_Allan Mark_Hakkenin Jeanne_Spellman observing-Michael_Cooper Shadi Andrew Sally
Regrets: Greg_Lowney
Found Date: 06 Nov 2009
Guessing minutes URL: http://www.w3.org/2009/11/06-ua-minutes.html
People with action items: jeanne kf kford

[End of scribe.perl diagnostic output]