WCAG WG Techniques Task Force weekly telecon

13 apr 2005


See also: IRC log


Becky Gibson, Chris Ridpath, Christophe Strobe, David MacDonald, Jenae Andershonis, John Slatin, Tim Boland, Yvette Hoitink, Wendy Chisholm, Ben Caldwell
David, Becky


Look at example use of link to mark navigation objects. Discuss proposed technique for feedback to UAWG and PFWG

<ben> http://lists.w3.org/Archives/Public/w3c-wai-ua/2005AprJun/0006.html

<wendy> http://cita.rehab.uiuc.edu/presentations/2005-03-css-tutorial/example/career-redesign-links.html

wc: look at "link" item want out feedback to PFWG
... source code link "rel" UA can identify that and skip blocks

wc: creating a firefox extension to implement and demo it, would replace skip links that we use now

BG: I'm confused, won't work til UA get on it
... will show pieces of doc but won't take you there
... need # sign

js: it's looking for a file not a bookmark (without the #)

wc: think there is an error in there code

js: implementing in 4.01 and XHTML 2.0 backward,

wc: UA wg driving it cause they want to know what to tell UA to do

js: might be "future" and "repair" will be easy in Xhtml2 but tricky in HTML1

bg: right now it only works in Moz or opera..

js: and moz and opera don't have screen reader support yet
... don't understand Mozilla, support if firefox puts text to speech

bg: Mozilla will die, everybody should switch to firefox.

wc: main question, what do we think of the technique, are there issue

tb: UA issues

wc: but apart from that, do we support UAWG in there efforts to push this on UA

<Becky_Gibson> example that works in Mozilla: http://lists.w3.org/Archives/Member/w3c-wai-pf/2005AprJun/0019.html

<Becky_Gibson> this is in member space

js: wondering about Screen readers, how will it work

bg: put this technique , these links, each UA will have its way to get to main content

js: is here any impact on the way a screen reader would behave in "say all" mode will it be convenient

bg: don't know if that has been thought through

js: it needs to be thought through

tb: nervous how it will test

yh: no side affects from using it. won't break code if UA don't support it.

bc: point out, interesting technique. create a standard outside of the spec..what are the implications of that. what motivations are there for browsers

js: this stuff is in spec for xhtml 2.0, the idea is to try to make it available to older versions of html
... can present it as future or optional but not for now

dm: might mean that usable current techniques won't be used even though this is not supported yet so it will mean that site may remain inaccessible

js: what is the expected for a result for a screen reader in say all mode, will it affect the way a SR reads a page
... take it to CG

<wendy> ACTION: wendy take summary of concerns to CG

Test files review, #75, 76, 77, 78¸ 79, 27, 128, 129, 183

<ben> david's summary: http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0117.html

js: important to use same terminology in tests and techs and guidelines
... term text equiv is still being used should be text alternative
... also statements must be testable: X is Y rather than X must be Y

ja: will be checking test files for consistency with wording

wc: looks like David's questions were about the test files

dm: Chris indicates that I should make suggestions form correcting them

tb: also found Chris's post helpful about procedure

dm: questions about test files being WCAG compliant in all respects except for error we are trying to identify
... isn't there a way we can make the test files compliant except for the specific error? looking for feedback from group
... tests seem they are written for machine test rather than human testing
... for example, some tests leave out the object tag in order to pass

js: didn't understand that from David's post

<wendy> http://www.w3.org/WAI/GL/WCAG20/tests/testfiles/75-2.html

<ben> example: http://www.w3.org/WAI/GL/WCAG20/tests/test75.html

bg: there are a few tests that require the page to be usable with object removed

<ben> (test file 75-2)

bg: pass test just removes the object tag - but there is no way for a machine to determine if the page is still usable

dm: agree

js: so how can we fix?

dm: another example is object must be keyboard operable. the pass test just leaves the object tag out
... that doesn't really test for keyboard operable

wc: so want to make sure test contains valid example and requires human testing

dm: can we do that - require human testing? this is why we are different from other WCAG groups

cr: agree with David - need to make human's part of the process; need to improve the test files to make more realistic
... example where the test file is not finished - encourage people to find good examples

dm: yes, we all need to help with the writing

wc: action item for Chris and David to develop the test files for object?

<wendy> ACTION: david and chris develop test files for tests in this set

dm: yes but have to agree that we have test files where just looking at the good can't guarantee a pass

cr: need a good test that a person can look at and understand

bc: test that removes element in question is a test file that illustrates when the test doesn't apply rather than a test file that passes the test

wc: does anyone disagree that test files should be usable in sense that they illustrate what we are testing for
... no one is disagreeing with David - issue is that the test file isn't finished
... Chris can you make the changes John suggested to the test files

cr: yes will do some editing

js: can you work with Jenae?

ja: is actually looking at the approved tests to double check

cr: willing to share :-)

<wendy> ACTION: chris edit these test cases (75, 76, 77, 78, 79, 27, 128, 129, 183) based on language feedback from john

dm: does group agree that it is possible to make WCAG compliant tests or will it bloat or confuse the tests?

bc: concerned that it will bloat the test and make too much work and harder to user to figure out what test is trying to accomplish. e.g., lang

cr: not sure what WCAG 2.0 compliant means yet :-)

bg: offer html with lang attribute as an example where it starts to make code larger

wc: what about video test - should it include video

js: yes and shouldn't be too hard to find
... issue with finding text alt for the object element
... example am hard of hearing and have various media players installed; if have stand alone audio file inside object element
... one tech says include transcript in body of object element - how would a user ever find that alternative
... when the technology is installed?

bg: trying to work that w/screen reader developers. want a way to access text in the object tag when the tech is enabled.

bg: currently, HPR offers the content. all will show you the title.

js: is there a way to provide text alt. explicitly associated with the object

bg: if have an object that is not accessible, could get to alternative.

js: doesn't solve hearing impaired issue since not using AT

bg: recommend putting title attribute

js: but can't put an entire transcript in title

wc: this has already been raised an an issue for GL 1.1
... could that prevent us from accepting these test if we don't have a technique to solve this issue

bc: have been recommending this since 1.0 but still isnt' supported in UA
... so do we change what we are recommending because UA haven't moved forward or do we put more pressure on UA?

wc: need more work on the technique

js: talk to UAAG and PF

<wendy> ACTION: john raise issue w/object support with UAWG and PFWG via CG

wc: helpful for Michael, Chris and David to develop technique proposal and test files all at once

js: since there are possible implications to GL 1.1 as well - can Wendy and Ben be added to that action as well

wc: Chris and David accept that change to previous action
... David, can you be lead?

dm: ok, but no conference space

wc: can create conf space if necessary

js: just want to make sure GL 1.1 still works and there are techs and test for implementing it

<wendy> action 2 = david, michael, wendy, and chris coordinate on changes to object techniques and test files (develop together to synchronized. wendy there to help with guideline 1.1 issues - explicitly associated defn)

bc: explicitly associated is the "sticking" point

dm: 2 other things about object element to discuss

wc: will they addressed by action or are they philosophical

dm: just make sure not to have obvious WCAG errors in the test files that may confuse the authors
... try to keep test files as compliant as possible; other issue is issue with embed

wc: assuming something about support for embed?

dm: yes using embed is a bit of a hack - should it be a repair technique

wc: in prev. discussions we felt okay dealing with embed; there are DTDs that you can validate against that have embed
... think this will be covered by action item

<ben> http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0114.html

<ben> http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0115.html>

Examples of Baseline information in HTML Techniques

<Tim> need to sign off now - sorry swamped at work today

bg: I was trying to CSS = 5.11 when to use this technique...how could I use this technique...found it hard to paraphrase when to use this...you assume CSS is present, might want to use <b> <i> if you expect your audience has not css then back to 3.2 etc. don't want to say use <font>

<wendy> http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0119.html

<wendy> dmd: looked through techniques to see how effected by baseline.

<wendy> http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0120.html

<wendy> dmd: in browser world, eveything is free, right?

js: chapter for each baseline instead of sorting it out at the technique level "if you have chosen this baseline here are your techniques"..."if ...

yh: how can you test for browsers. ??? well we use VM ware that makes us test all ie.5.5, 5.0, etc on different virtual machines

<wendy> http://webmonkey.wired.com/webmonkey/reference/browser_chart/ although doesn't have assistive technologies

<Zakim> wendy, you wanted to ask "i18n - http://www.w3.org/International/geo/html-tech/outline/#ri20030510.102829377"

wc: 1) what do authors need 2) what resources do we have
... richard said, "i don't want to get sold on a technique then find out if it won't work in a browser
... understand user support...lots of data out there for elements in browsers, what will we include....there is tons of data out there, we just need to focus on AT
... internationalization is providing tons of info and less people than us

js: I like saying something up front to let people make good selections...whether chapter idea I had or wendy's idea
... must have feature support going back several browsers

wc: depends on how they would let us use there info...they don't deal with baselines. which is similar to the direction we are headed

bc: what do they do when something works in ie 5 and not 6 what happens i there support chart...

wc: our chart could get very complex quickly

js: then you get into video card etc....

<ben> http://www.w3.org/International/geo/html-tech/tech-lang.html#ri20040429.113217290

wc: we need to focus on intended default behaviour is complex enough

bc: internationalization have a simple matrix and a complex one

js: language easier than our problems internationalization is providing very cool layouts etc

js: we should be able to grow steadily like them

wc: maybe if we just focus on the scripting techniques...rather than all techniques

bc: yeah we could document the important areas like screen magnification , screen readers etc.

wc: any action items in progress for scripting?

bg: yup, mc and I

wc: feels we need scripting baseline stuff, develop those techniques becoming important. try a table,, what browser support is out there....etc.

bg: I say 3 categories, 1) script turned off 2) script can help accessibility 3) if writing scripts here's how to write it accessibly

wc: scripting , j- script, javascript, etc.. dosn't feel that we have the same broad categories in scripting as HTML, CSS,

js: 1st decision an author will have to make is scripting, matter of the order we present things in a decision tree.

bc: different consequences to baseline support, bunch of elements in 3.2 , is it worth documenting that

dm: even poor countries have 3.7

wc: is there action item or resolution

yh: wondering if planning "out of the box features of windows"
... people slightly short sighted might use windows magnifier and not know about magic etc
... see if our techniques work with windows accessibility features...

bc: can't see where our techniques are in conflict,,

<wendy> heard: proposal for "chapters," different issues for baseline for scripting than html/css (although some overlap. although "property/element support" is one aspect and "scripting off/on" another), i18n approach of matrix, focus just on testing asst tech and harvest browser support info from external resources, get info from asst tech developers

js: narrator...MS is explicit that its not a serious screen reader. deliberately limited...anti trust issues

wc: john had a proposal for chapters, different issues for baseline, int'l approach their matrix, harvesting external ideas

<wendy> ACTION: john contact asst tech developers to discover what info they provide about support

<wendy> ACTION: john write one paragraph summary about "chapters" idea for techniques

<wendy> ACTION: jenae send john contact info

wc: left over questions, safari, jaws unicode supports, and support of other languages

User Agent matrix

js: let's get preliminary info on safari it will become a more realistic option that blind users on a MAC...web developers who are MAC based. can do accessibility testing the platform

wc: a lot of this info available already

js: zoomtext reads aloud and does karaoke style reading as they go along

bc: jaws support for reading style sheets...provide a way to disable style sheets using !import feature. interesting implication...if SR read stylesheets it will effect us....ie "show/hide"

wc: what action item do you recommend

js: create the matrix

<wendy> http://lists.w3.org/Archives/Public/w3c-wai-gl/2004JanMar/att-0667/draft_test_matrix_v2.html

ja: matrix exists do we just need to update it

js: yes it needs to be updated, how far back does it go

ja: I 'll update it

<wendy> ACTION: jenae finish update of matrix

wc: let's adjourn

Summary of Action Items

[NEW] ACTION: chris edit these test cases (75, 76, 77, 78¸ 79, 27,
... 128, 129, 183) based on language feedback from john
[NEW] ACTION: david and chris develop test files for tests in this
... set
[NEW] ACTION: jenae finish update of matrix
[NEW] ACTION: jenae send john contact info
[NEW] ACTION: john contact asst tech developers to discover what
... info they provide about support
[NEW] ACTION: john raise issue w/object support with UAWG and PFWG
... via CG
[NEW] ACTION: john write one paragraph summary about "chapters"
... idea for techniques
[NEW] ACTION: wendy take summary of concerns to CG

Minutes formatted by David Booth's scribe.perl 1.94 (CVS log)
$Date: 2005/04/14 18:57:49 $