W3C logo Web Accessibility Initiative (WAI) logo > WAI IG Home Page > WAI IG Meeting 22 March 2003

WAI IG Minutes 22 March 2003 Meeting

on this page: attendees - W3C WAI - UAWG - ATWG - WCAG WG - RDIG - PFWG - ERTWG - EOWG - testing - EARL wrapper



W3C WAIstatus and plans

Judy Brewer:

User Agent Accessibility Guidelines Working Group

User Agent Accessibility Guidelines Working Group

Matt May: presentation online

Judy Brewer:

Mark Urban: how well define participation of user agents that don't fit into standard UA; e.g., Flash, PDF?

Matt May: we're spreading the word that UA is more than just browser - something we are trying to work through

Authoring Tool Accessibility Guidelines Working Group

Authoring Tool Accessibility Guidelines Working Group

Jan Richards: presentation online

Cynthia Waddell: about saving a Word document, I was vocal about that problem. also appreciate seeing on the list a content management system.

Mark Urban: talk about ATAG not being dependent on WCAG2.0

Jan Richards: we refer throughout the document to WCAG, then the tool that is saying it conforms will specify to which versions of which (e.g., guidelines., WCAG1, WCAG2) ... handle the mapping of techniques throughout... then for new guidelines we can publish a new note, and not have to do a new version of guidelines

WCAG Web Content Accessibility Guidelines Working Group

Web Content Acessibility Guidelines Working Group


Marja-Riitta Koivunen: checkpoints are going to be more abstract, at same time need to define what exactly needs to be done; however, seems opposite - have thought about making techniques conformant [normative]

Jason White: more specific in success criteria... techniques explain how implemented

Marja-Riitta Koivunen: possible for people do to things other than is in the techniques?

Jason White: may be possible to meet success criteria by a different techniques.. conformance is at the level of the guidelines document, techniques make it easier to understand how that is to implemented

Michael Cooper: one reason techniques are non-normative is then can update more frequently. if someone comes up with a way to meet it that is not in the techniques, then we want to add it

Bob Regan: say word or two of the roles of non- W3C technologies under the WCAG?

Jason White: because the guidelines document written independent of specific technologies... should be possible to meet guidelines with technologies not by W3C... at the techniques level, it should be possible for the non-W3C orgs to provide techniques in similar format; however, W3C will only publish techniques for W3C

Cynthia Waddell: Have a policy observation. Multiple levels of conformance can be quite complex. Earlier with WCAG 1.0 there have been similar complaints, and complexity increased with 2.0, strongly recommend redoing complexity and reducing proliferation of checkpoints.

Judy Brewer: Suggestions? Would you be interested in looking at that? One of the things that group has done is decrease number of checkpoints.

Cynthia Waddell: Much more. Think the value of techniques document is great, success of framework will be helped by what's in non-normative techniques doc. and will define whether meet success criteria.

Judy Brewer: Flag for discussion in the afternoon.

Research and Development Interest Group

Research and Development Interest Group
Markku Hakkinen:

Protocol and Formats Working Group

Protocols and Formats Working Group

Matt May: presentation online

Evaluation and Repair Tools Working Group

Evaluation and Repair Tools Working Group

Wendy Chisholm: <a href="@@WENDY@@">presentation online</a>

Bob Regan: for things other than [missed]

Bill LaPlant: Any relationship between EARL and Common Industry Format (CIF)?

Wendy Chisholm: have talked to Sharon about that

Bill LaPlant: purpose of that is to allow for internal evalation - seems there would be some synergism

Education and Outreach Working Group

Education and Outreach Working Group

Shawn Henry:

Bill LaPlant: has been some work on the normalization & terminology between usability & accessibility, suggest that we...

Judy Brewer: clarify that we are doing UCD as a process for redesigning the WAI site, not doing work on usability. there is interest in a Usability interest group at W3C.

General discussion on testing

Wendy Chisholm: How do we make it easier to test WCAG, without making WCAG bigger/more complicated? Michael Cooper has 530 test files. What is the balance between providing more testability - but if there are more tests, then become...

[some participants shocked to see a stranger emerge from underneath a table skirt. as the familiar red, plaid hat swings onto the strangers head, sighs of recognition release the tension. it was only Jim Allan getting inside the tables to pass the microphone around ]

Bill LaPlant: the danger of having any test suite is that the developer develops to the test suite, rather than guidelines

Sailesh Panchang: said WCAG1.0 was based mostly on HTML, now using different technologies, put techniques for each of the technologies separately

Mark Urban: if so complex, then people can't readily implement tests

[Bob Regan: "congratulations to Jim for his inside the box thinking"]

Bob Regan: designing exactly for a test suite is in some ways we want them to start - to develop a heusrist way to understand the pirit of the guidelines - we ahve to accep that that may be part of the learning process when people just get started

Judy Brewer: if we are serious about wanting to get most of the Web to be accessible, we need to aim for a completely different goal - that is press a button and get the answer about whether it is accessible or not, like an HTML validator. .. maybe need to sponsor research on how to do authomated evaluation of the things that are difficult to do today .... so that down the road evlauation woun't be an artform

Wendy Chisholm: clarify that the 530 test are mostly for authoring tools and evaluation tools and as example, rather than content developers needing to do each. There is an art to accessibility, there's an art to creating web pages, and therefore, there is an art of evalatuion. in WCAG2 beacuse we are so concerned about testability, some of the artful advice that we can give is sorta falling away, and that concerns me.

Preety Kumar: use cases to go along with the test

Matt May: people don't need the details... the quick tips is pound for pound the best thing for accessibility. need someway to describe ina meaningful way the overall design approach to web accessibility. best thing is to get authoring tools to do it - relieve burden from authors

Marja-Riitta Koivunen: like WCAG2.0 abstractions, don't understand the levels. worried about coming up with a numbering scheme that works for people.

Mark Urban: philosophically - late Dr. Len Kasday said a computer is not a human, testing for human access, not computer access, therefore, requires a human

Janina Sajka: complexity itself not a problem. autos got more complicated. interface relatively simple to deal with, while the underlying might be more complex

Michael Cooper: unscare people about the test files - developed for evaluation tools only. relationships with other test suites. set of requirements for test files is that able to test both the negative and positive case.

Jason White: WCAG2.0 has fewer guidelines & checkpoints - always trying to find simple, elegant solutions; however, need to cover broad, and be lasting. working group would welcome proposals & suggestions...  future technologies... within constraints

Preety Kumar: hearing test cases should be built at techniques level, but also at guidelines. need to have scenarios, use cases, test cases

John Slatin: problem isn't complexity so much as it is clarity - need to work hard on the clariuty of language & presentations - not only langugea, but format of document. agree very strongly that there is an art to accessisbility, and the guidelines have to help people how to do the art, but they are not sufficient to do that by themselves.

Judy Brewer: the first time that I heard agreeing noises across the room was when Joh n said clarity.

EARL wrapper

Shadi Abou-Zahra: idea is to combine the results output from different tools into one single report.

  1. evaluating is a complex action/event, each tool has its pros, cons, strengths, weakness - by combining, get andvantages of all tools
  2. current eval tools not concentrate on HTML. this project layer between the evalaution itself & the authoiring tools, editing, repair
Shadi Abou-Zahra: Hope I get help from evaluation tool vendors/authors/developers. & that we can also work on EARL.

Preety Kumar: assertoins that are made by the different tools - is there any way to determine the quality of the assurtion

Shadi Abou-Zahra: difficult to rely on what the tool - hope to find alignment in the assertions themselves - e.g., tool says I require human input, another tool can determine without human - real work is in alighing those too test togeher

Judy Brewer: some evalatunb tools venders have said that they would appreaciate getting feedback on their tools

Shadi Abou-Zahra: I think test criteria has pass, fail, or require human review

Last updated 3 April 2003 by Shawn Henry <shawn @w3.org>