W3C logo Web Accessibility Initiative (WAI) logo | EOWG Home

EOWG Meeting, April 26, 2002


Code Name
HB: Harvey Bingham, Scribe
JB: Judy Brewer, Chair
AC: Alan Cantor
JMD: Jean-Marie D'Amour
AG: Audrey Gorman
MRK: Marja-Riitta Koivunen
DS: Doyle Saylor
Kathleen Anderson
Helle Bjarno
Libby Cohen
Chuck Letourneau
Henk Snetselaar
Eric Velleman

Outreach Updates

AG: City of Naperville, IL (over 130,000 population) has a $56K contract for site upgrade, requiring accessibility on the website, based on the draft State of Illinois Accessibility Guidelines. Will send pointer to those guidelines. More stringent than Section 508, closer to WAI guidelines.

JB: Attended Euro-China 2002 conference in BeiJing, hosted by European Commission and China government, on information technology and business. Presented on citizens and information in society track. Spoke with representatives of disability organizations. Met with accessibility team in the Sun Beijing office.

JB: WAI is working on contract for additional accessibility outreach in Europe starting in October when current contract completes. They're very interested in evaluation approaches. Intense discussions are going on in the E-Accessibility Working Group under the European Commission on evaluation of websites. We should expect comments from them on our newer version. Judy is contact for WAI.

Implementation Plan for Web Accessibility


JB: Continuing our work, see Change Log: Implementation Plan for Web Accessibility


JB: "Explain more about what to look for in code when reviewing data tables in lynx"

JMD: Now use Opera, rather than Lynx, can make all tests there.

JB: We mention Lynx in the preliminary review.

AG: Concern that its mention is too tecnical. We should address data tables. We are seeking equivalent information from content, not code

JMD: Lynx addresses data tables.

DS: Concur. Lynx has limitations, uncertain.

JB: Should we include limits on how we describe tools such as lynx.

AC: What will be uncovered by doing this study in Lynx.

JB: Several different tools can reveal order in a table.

AG: We use lynx as only one example, of a text browser such as lynx. There is already an extensive note. Add another: don't use on data tables.

JB: Why not?

JMD: Lynx linearizes the table. Cannot see if data cell is related to header cell. Two things to do with voice/text brwoser. Lynx is a good way to see if we have alt text for images. Is information in same logical order. OK for a layout table. For data table less effective.

JB: Add new note in this section: do not use lynx to test linearization of data tables, as this may give ambiguous results.

JMD: Why do we need to include this note. The two tasks for lynx are the correct ones. We don't talk about tables.

DS: We shouldn't assert this. How can we know all the ways it doesn't work well?

JMD: Do not add note. For lynx assign two tasks for which it is good. We don't talk about tables.

JB: On our second defined task, is the information presented in a similar logical order as with a GUI. The second task is important for testing logical presentation.

JMD: Where we can find a negative, and search the second question with layout tables, some pages designed with layers, code order is not necessarily the logical order of content. Layers in layout tables can be tested in lynx. Few pages are designed with layers.

JB: We should replace the discussion about data tables.

JMD: Preliminary review test just layout tables.

JB: Question one: Logical order. Information in layout tables presented in similar logical order.

AG: Looking at whole thing, similar logical order, inadequately tested in lynx. Sepcific to Section #2, the problem we see is only with data tables, don't exclude all other information and its logical order.

HB: Layout tables deprecated in favor of style sheets.

AG: Leave question. Add note: make evident specific to #2, This will work for many things, but not for some such as data tables.

AC: As this is a document for data vs layout tables. There are nested data tables.

JB: Don't use lynx on datatables, provide link to where data tables are defined. DS: Simplify don't take out.

Conformance Evaluation Intro

DS: Evaluations are demanding for volunteers.

JB: Concerned about how realistic to do comprehensive monitoring if it includes usability testing.

JB: Not necessarily volunteer work, may be paid job. Deal with the ongoing monitoring.

AG: Needn't do everything doggedly every month. What are we really asking? Add language on ongoing monitoring.

JB: Clarify that full comprehensive eval not required.

MJK: Evaluate at times of change.

JB: Run automated tests, customized, based on problems arising in automated evaluations.

AC: A way to provide people the way to do quick and dirty testing that you can do periodically.

JB: That's what the preliminary review does. Ongoing monitoring is to identify what may change.

AG: In some conscientious groups, would pick up just the ones they have found problems with in the past.

JB: Need some limited automated/semi-automated tests. After some period of time, rerun comprehensive tests.

HB: Plan to do comprehensive test any time there is a significant design change. Include accessibility evaluation of any such new design.

JMD: Have never done a usability evaluation of a website, though have done evaluations of accessibility.

JB: Reminder that we decided to say "Conformance Evaluation", not "Comprehensive Evaluation."

JMD: Have trained blind people for 5 years on use of JAWS. That person won't be an effective tester at novice level.

JB: We can't conveniently combine "lack of resources" and usability.

AG: Usability testing helps more with other disabilities. So need it too. If you haven't gotten into accessibility, alibi is no resources. Include accessibility in bigger picture make life better for all. Usability piece adds to that appeal.

JB: Concern: "Conformance Evaluation" dancing on edge of comprehensive evaluation. Most W3C specifications have simpler test tools. In conformance testing, if we also open up to usability, we are diluting our emphasis. JMD comments as experienced reviewer are important. Need to limit scope of usability testing to how the accessibility features are implemented. Do some of those useability techniques interfere?

AG: Will tackle doing a grid of accessibility and usability. Those doing controlled useability testing. Some places they touch closely.

JMD: Problem is in heading "disability evaluation", a rigorous process to do. Propose: affordable and realistic in first paragraph. "Different levels of user technical experience and familiarity are important."

JMD: Ask testers to find answers to the most common questions. Refers to a usability testing process.

JB: "Useability Testing" to "User testing of implementation of accessibility features" "useful" rather than "important"

JB: If we want to reduce scope of user testing to focus on implementation of accessibility features, that's what we should say. Ask users to test to the guidelines.

JMD: Will focus on accessibility feature (second bullet). First bullet is general.

DS: Wish to see written version.

JB: "Useability testing change title to "User testing of accessibility features."

JB: Ask users to test accessibility features on site.

AG: You can ask people to answer questions to common questions, and observe or ask for feedback on how to use the features. They may not be familiary with accessibility features. Presume one-on-one. Could be done remotely.

AG: If assume too much, based on what they don't know, make wrong

Evaluating Web Sites for Accessibility

JB: Give guidelines.

JMD: Ask testers to find answers to most common questions. Where it is difficult or impossible to use the website, you can ask in some situations the person responsible for the evaluation can make a different choice.

JB: Confusion on page type. What is scope of page selection: One of each kind and all top pages. Vague what are entry pages.

JB: 3.1.1 vs 3.1.2 above. We haven't addressed page types like entry pages. Use something more vague.

MRK: any page can be an entry page.

JB: Remove references to entry page.

HB: Entry pages are normal top-level pages, simplest URL.

JB: Replace by definition of top-level pages where people are more likely to enter the site.

AC: Covers the possibility of redirection to somewhere else.

AG: ALA site, a loosely-coupled group for other sites.

Verify differences betweek #2 and #4 on legacy site.

JB: Legacy sites may be of often ambiguous authorship. They may possibly have some affiliation with another organization, a third-party source no longer under contract. Who is responsible? You may be in pseudo-ownership condition, with no clear owner.

JB: Clarify internal vs external.

AG: Break #1 into two bullets: identification vs obligation to do.

JB: Additional, deferred: Provide more content, more detail on how to do, commercial viability.

Meeting June 6 and 7

JB: Haven't met in Montreal, so that would be first choice. Haven't yet confirmed with Mary Frances Laughton.

JB: Try to finalize arrangements for those dates, will announce on list. Any opposition from those in good standing on the list.

AC: Great, good to get to Montreal.

JB: Next meeting 2002-05-03

JB: No meeting 2002-05-10 as that's when WWW conference in Hawaii.