Minutes from WCAG 2.0 F2F - Day 1, 20 June 2001

Table of Contents

Discussion Summaries

Summary of Usability Testing Proposal discussion

Detailed minutes of the discussion of the Usability Testing Proposal.

Refer to Proposal for usability testing of WCAG - David Sloan

The following suggestions were made and issues raised:

Usability testing of examples to be included in techniques documents is a different issue and process.

We resolved to discuss it on list and in upcoming telecons. We told David we would give him feedback in a few weeks, that it might take some time to gather our thoughts.

David aniticipated that it would take 6 months to complete the project.

We didn't feel we would have a draft ready for testing until September.

David felt this was appropriate, since they needed to get funding. He asked if WAI could offer any support.

Wendy took an action item to discuss possibilities with Judy.

ACTION AR1: discuss extending studies to Finnish academic and professional community

Testability of checkpoints

Detailed minutes of the testability of checkpoints discussion.

We discussed minimum requirements as being used by the recent ATAG draft and as in use in UAAG drafts for a while. Rather than discussing how minimum requirements might be worded or what we would want to convey in a minimum requirement, we instead debated the usefulness of minimum requirements. We did not come to any conclusions and have left this as an open issue.

Graphics/Multimedia Techniques break-out session

The notes from the Graphics/Multimedia session are available from the mailing list archives. Refer to the message from Charles McCathieNevile - notes from first f2f breakout on graphics. Also refer to Gregory Rosmaita's notes.

HTML Techniques break-out session

Wendy Chisholm's notes from the HTML session follow.

Participants: Chris Ridpath, Matt May, Paul Bohman, Katie Haritos-Shea, Lisa Seeman, Wendy Chisholm

We propose the following series of events:

In regards to the AERT:

LS is interested in implementing and testing techniques

LS is interested in working on CSS techniques

CR will work with us, particularly during the implementation phase, so that we can have an evaluation tool that can help authors determine if they have conformed to WCAG and checkpoint solutions.

CR says they are looking for a student to work on HTML Tidy. They hope to have someone by the end of summer.

LS is interested in the work on HTML Tidy. CR and LS will discuss.

We talked about test files. ERT used to have a set of test files for each checkpoint, CR is still developing, WC also developing. Test files are very useful for tool developers to determine if they are each getting the same answer or not. Also helps people evaluating the tools to determine which one they like better or find more useful.

We discussed an admin section of the DTD where we would record notes for ourselves and then strip out before generating a document.

We wondered about making the xml of the techniques docs available so that they could be used in authoring/evaluation tools. WC remembered that there had been a discussion about this in the WAI domain, but couldn't remember pros and cons of the discussion.

Detailed minutes

9:30 - 10:00 Introductions

WC: work for W3C, from the USA; I'm here to chair the meeting in the physical absence of our 2 chairs; hoping we can get a lot done at this meeting and that people will leave the meeting enthused and eager to move WCAG2 forward

AR1: European Union for the Deaf; interested in the work of the GL WG, web very important to the deaf, as well as to all people, disabled or not; working on distance learning software/solutions for deaf students

CR: from the Adaptive Technology Resource Centre (ATRC) at the University of Toronto; editor of Accessibility Evaluation & Repair Techniques (AERT); lead developer of A- Prompt tool; from Canada

JR: from the ATRC; co-editor of Authoring Tool Accessibility Guidelines ATAG; from Canada;

AR2: teach information systems at California State University-Los Angles; most of my graduate students work fulltime as IT people in corporations; have a great deal of first and second hand consulting experience meeting various needs;

DS: University of Dundee (Scotland); digital media access group--usability and accessibility consulting group; department of applied technology--provide solutions for persons with disabilities; Alan Newell is chair of my department--internationally recognized expert in the field; here to talk about usability testing of WCAG

MM: from the USA; background in commercial web site development and testing; working on HTML/XHTML techniques for WACAG

PB: at Utah State University; working on project called WebAIM--mainly an educational initiative to help universities make

LS: make web tools for people with learning disabilities that died a sad death due to bureaucracy; carrying on doing web accessibility stuff--more generalize, use global formats

CMN: with W3C; from Oz; staff contact for Authoring Tools; editor of ATAG; work extensively as member of the PF [Protocols & Formats] working group as well

GJR: from New Jersey; WebMaster and Minister of Propaganda for the Visually Impaired Computer Users' Group of New York City (VICUG NYC); been active in WAI since 1997, when was invited to work with the WG formerly known as HC, and now known as PF; am the Interest Group representative to the WAI Co-Ordination [read: Chairs'] Group

KHS: from the USA; work for US government agency as 508 coordinator for accessibility

DD: from France; technical manager of the WAI;

GV: from the USA; director of the TRACE Center at the University of Wisconsin at Madison; co-chair of GL working group; it's 2 in the morning where I am right now

// DD tries to call Jason White (JW) //

// the minutes go "live" via the projector //

// JW joins //

// the minutes are live!!! //

AR2: this being up to date, the content modes are not in 4.1 and I thought that they had been put in

WC: this is from March; no resolution on content modes

// administrivial discussion amongst the W3C staff people //

GJR [typed to screen]: is everyone who can see 'em happy with the projection of the minutes?

// YES //

10:00 - 10:30, 11 - 12:30 Usability Proposal discussion detailed minutes

DS: have a floppy with a PowerPoint presentation on it--could project it while I speak

// WL joins //

// GJR suspends active minuting so that his laptop can be used to project DS' slide show; minuting continued by CMN by logging the active IRC session //

GV: slide 11 you said that get feedback from participants -- assessment by experts of the content produced

// DD tweaks the phone //

GV: people who are deaf using interpreters have difficulty raising their hands because of lag in interpretation

GV: slide 11 -- 2 usability items--feedback from participants; assessment by experts of accessibility of materials produced -- how are you going to parlay that back into usability of the GL rather than the expertise of the users

DS: that stage is there to ensure that WCAG serves the purpose; by asking experts to validate info that participants crate allows us to judge conformance to specific GLs or priorities; which are in fact met; that's why that was in there--as a quality control over content produced by participants to see if WCAG doing what it is intended to do

GV:P temptation to draw causative conclusions -- they didn't do this, therefore the GL might not be useful in the are -- participant may have made a decision not to use a checkpoint because they didn't understand or agree; if experts are on hand, can look at it and see what is missing, which could be turned into a question t ask participants -- why didn't you do X -- did you not see it, etc. -- user might say "I saw it but didn't know how to do it".. etc. -- learn why didn't follow guideline, rather than assume wasn't followed due to usability

DS: agree completely--that's an extra stage--structured interview following expert assessment or observing a GL not being implemented followed

GV: do as close in time to the event as possible

LS: few questions: 1) said you give the participants the guidelines and I think that was the first thing that happened to the participants in this exercise -- no control no mechanism to know what they would have put in without WCAG -- give them the task first, see what they do (exercise common sense, for example) then introduce them to WCAG; 2) in selecting users -- do they represent a cross-section from an internationalization point of view; cross section of people with technological know-how and those who came to computers later in life -- not one community from which to draw participants; 3) another issue -- give them existing web pages and ask them to spot accessibility violations; one often comes across sites where people have made an effort to be accessible but have failed--sites tend to be created using a tool, should be involved in assessing / identify situations where there has been a genuine attempt to make a site accessible, but when analyzed by an expert, it isn't; think that when FrontPage makes a TABLE layout --draws picture] 2 columns, 2 rows -- top row has logo, etc. other has side menu -- that's how one would assume they put in links, in the embedded side menu; in a site I came across who had made a noticeable effort to make the site accessible, but the tool used to construct the page, instead of having an embedded table, have a rowspan of 2 with word "Welcome": -- new row that includes both columns (right-hand and left-hand) new products link in left hand column followed sequentially by "Who We Are" or "Contact Us" embedded in separate paragraphs as you read cell by cell; might be really good to compile a list of such violations as we come across them and give that info to the usability group

AR2: rather than have a list of violations, since new violations are being invented every day isn't realistic; need a testing procedure -- first order testing procedure for GL1 is "see if everything shows up in Lynx" -- if it does, then GL1 is about 99$ met; what we are going to need for GL2 is something were instead of just getting a list of possible violations is to have these people work in the kind of tool environment where they have an interactive view of accessibility and potential problems while working on the document; what people would expect to work with -- base computing environment -- isn't a set of requirements on paper, -- expect an interactive tool that incorporates the GLs and lets them put them in a web page

MM: similar issues to LS; diff perspective; give them something to do from scratch -- present them with something to design from the beginning -- different use case for the guidelines -- creation, evaluation, and repair; we need to succeed on both of those; wanted to make sure that there is a relatively good definition of who is a designer and who is a developer -- more than just defining an amateur as someone using a tool -- many paths to writing a web page, whether skill is server-side development, Flash, javascript -- issues are the same for presenting the content -- designers see the language as obstacle to design, so they hack it until design is what they want; giving them a document that has certain things in color, but structure solid, ensure that they don't lose structure of document as change presentation

JR: follow up on LS on role of tools -- would be interesting to see which tools are being used and whether there are patterns of violation and which allow WCAG2 to be complied with; what are the liabilities of the tool, what are the strengths;

WC: few comments -- accessibility issues we want to catch; usability issues should be caught as people use WCAG; interesting to look at those using tools -- a lot of people use more than one tool; one thing doing with WCAG2 is moving away from being HTML specific -- tests with SVG and SMIL? One big concern is server-side generation -- make sure that someone using a database to drive a site does it accessible -- are the tools that allow that

MM: want to see if we can get subjects to self-identify skill levels with certain technologies -- do they know what should be done, but were prevented by a tool; new technologies -- that's what I meant by designing from scratch -- no need to go back to go forward (to SVG); skill levels are built up

LS: think important if they are being asked to design a page with accessibility in mind, if I were them I'd do a plain unaesthetic way not using tools coming out with something accessible, but that would not reflect a real life situation -- have to insist on an amount of tool use and somehow mimic the pressure and pressure for look and feel and bells and whistles and gizmos; not sure how we can do that -- perhaps can tell them that they are being tested on ability to build corporate type web site -- ability to show off design skills

JW: very good proposal; important to recognize purpose of testing is to determine how effectively people can read and apply a set of GLs and associated documentation; think most of the GLs should be built into software so not necessary to refer to document directly, but work through authoring tool; this type of test is an invaluable resource for improving guidelines; WCAG has a number of constraints on it -- what can be learned by testing discrete parts of WCAG2 -- modularized WCAG; interesting to see if could cover a range of WCAG2 modules among those being used as subjects of the study; will help evolution of the guidelines

GV: one other way to test is by looking at the tools that build evaluation into them; a number of places where people take WCAG and put them in the tool without phrasing from ER and either massacre or slightly bend the rules -- which rules don't people understand when they build that type of device into an authoring tool

LS: another spin on the usability questions -- if have scenario where given a task to do without having read WCAG and then a site with the guidelines, what about asking disabled users to test the usability/accessibility of the test pages -- what did they prefer? What worked? What didn't? what info did they miss? -- fulfilling the guidelines to the letter doesn't always make the end user happy -- test not only how easy it is to apply WECAG, but how effective is the end result from an accessibility viewpoint

WC: could look at usability testing of WCAG and of techniques; database of techniques with some sort of usability rating along with each technique -- test them before recommending that people use them -- how formal the process? Need to separate from what DS and Helen Petrie are doing; concerned about how we separate -- WCAG understandable and usable by users of tools; testing of tools themselves; authors and their needs -- main focus should be that WCAG itself is usable -- would hope then that an author or tool developer could use them -- is that a good assumption? Concerned that when add in all the variables of tools, how we would deal with that

DS: think it is a very important piece of information -- in proposal said wanted to test people in their own environment (office, home) using the tools that they regularly use and are comfortable using;

WC: makes me think that we aren't just defining GLs, but a process -- some have commented that we need to do that anyway -- are the GLs just a part of the process? Can we give people scope over use?

PB: more than one version of the guidelines to test -- if test only one version, have good feedback on that version, but with the current model of W3C documents they are long, one page documents; my first encounter with WAI documents was intimidating -- stinks of technicality; useful to have a document that is more geared towards someone who doesn't read technical documents, but just wants to know how to make a site accessible; not suggesting that we disregard the W3C format, but the idea of investigating another format is a useful one

AR2: I'm concerned about the ecological validity of the study -- beginners don't normally use paper documents; those who are not professionals would never use a GL document; unless rig an interactive interface to WCAG, won't have ecological validity for users not familiar with underlying technology; way to achieve is to limit study to professionals used to reading specs and applying them -- those are the ones who will be using the actual guidelines, in creating tools for others;

JW: comment subsumed with by AR2's last comment; assume a certain level of knowledge of underlying technologies necessary to provide detailed technical feedback; appropriate to get comments from those without that type of background -- are those using tools without detailed knowledge of underlying tech can use the higher level documents that form WCAG2 to create accessible content and evaluate content for accessibility; need to be clear about the audiences of each level of WCAG2; if tools are developed during CR or PR, one could assess how usable the tools are on populations that aren't likely to read or apply WCAG directly; sharpen parameters

GJR1: minuted by WC

AR1: what PB said ; had a project were our developers tried to follow WCAG -- he said we are doing this for people, not for the W3C's sake" -- spec usability is important -- REALLY important -- dealing with artists and designers -- get nervous when you warn them about not doing this and doing that in a specific manner; shouldn't be led to hat accessibility/usability -- ways of communicating with them; when take WCAG, you can put more into your creative process --that's how we have to sell it to actual designers

WC: ideas?

AR1: lot of testing at our university whose contact info I can exchange with DS; working with deaf students -- language is sign language -- that is their mother tongue

// ACTION AR1: discuss extending studies to Finnish academic and professional community //

MM: what PB and AR1 are saying underscores need for us to go after those who are not experts -- need to get this message out to as wide an audience as possible; can't just ID professionals in technologies -- will lead to WCAG being oriented to a small group of people; address the broadest audience possible; plain English version of WCAG to simplify the lower levels; for time being, we are cornering ourselves by focusing on people we believe are creating these sites

WC: we need to be working very closely with AU WG so that as move forward, can get these things in the tools, not just on paper/screen/ears; get to point in CR where to exit, have to show implementation experience -- part may be that our techniques are being included in tools, both authoring and evaluation and repair tools; in usability testing, would be useful to have a tool like A-Prompt that helped them use WCAG

CR: would be really helpful for the tool makers if the GLs can be easily interpreted -- they need to say what you intend them to say -- mark sure good set of technical GLS

GJR2: minuted by WC

CMN: one approach is to look at AU tool developers directly as one of the target populations for usability testing of the guidelines -- those are the people who interpret them, produce the tools that are the way that most content will be generated

WC: don't think that we're changing the levels -- have to specify in the levels the information appropriate to that level; main thing is ensuring that tools help people have to ensure that tools are implementing WCAG in the best manner possible

GJR3: minuted by WC

DS: written down a lot and will analyze the feedback; very keen that we involve the sort of people who aren't technical experts, but have the responsibility to look after the web content (managers) -- are told must be accessible, but may not know anything about either accessibility or underlying technology

// WL leaves //

// 15 minute break commencing at 11:03 local time //

WC: summarize discussions from break

AR2: need end user requirement document that says that people with such and such a disability need to be able to do this, then this is what content they need, this is what UAs need to expose this is what AU tools need to support; the documents of the 3 WGs would be subordinate to end user requirement document and would b organized on that basis

CR: ER tools need a document to guide them, so need coordination between AU and ER to develop them in concert with WCAG

AR1: discussion with PB about coordinating usability studies as mentioned earlier

LS: had work related conversations, but not necessarily about this topic -- page map, HTML techniques, AR2 had a great idea about moving forward to address problems with Semitic languages; still wondering how deaf dyslexics could benefit from our work

MM: talked with CR about syncing up and the DTD that we are developing for Techniques in WCAG2 -- could help coordination among diff groups

DS: a lot of small talk and discussion of usability testing and the extension of them ad sharing resources, knowledge and expertise

WC: 43 minutes until lunch -- like to come to some conclusions -- next steps; how we can give HP and DS feedback on how to proceed; have a few issues that need to be taken back to WG on list; 2 issues: 1) overlap with tools in defining a process for how someone might use a document in concert with other tools or on their own; 2) audiences: experienced users, naive users, etc. -- based on DS' proposal, -- who thinks we should proceed with proposal as is? Just proceed or are there changes we need to make

AR2: specifically the proposal we are voting on is?

WC: what DS presented today

WC: 3 words or less how you would modify DS' proposal

LS: suggestion: proposal as a stage map -- line by line of the proposal which stages are involved, then people can write in what is necessary to add

WC: whiteboard it? Brainstorming session rather than discussion

DS: trying to think how I would make such a map

[PB in charge of whiteboard]

WC: information gathering on how people -- discover current state of people's knowledge of WCAG;

DS: people responsible for putting up web content or telling others to do so and web development professionals -- testing their knowledge of accessibility issues -- how PWDs access the web, importance of accessibility to an organization or company; legal requirements; and their perceived effort in making things accessible -- what resources do they think are required to create accessible resources -- don't know if this type of info has ever been gathered before; attitudes and knowledge about accessibility

GV: like to see some expansion of coverage of issue around how experts evaluation fits in -- easy to get off track in that area; other part is that it wasn't clear from discussion exactly how going to separate -- good thing to get straight up front -- measure twice and cut once; most people using tools today -- unless you know exactly why they didn't do something is because they didn't think to look under the advanced or annotations button to add annotation to a graphic using the tool familiar to them -- missing: 1) have to have a better method for handling and finding out why didn't do something -- that feedback should be obtained instantaneously to capture why what was done was done

WC: GV's comment is on the second part of the methodology --something to highlight; not making firm decisions today -- just a sample population of the WG -- need more discussion from the WG after a chance to read proposal--next steps: put on agenda for next week's telecon and have more discussion on the list; today I'd like to quickly gather thoughts people having right now, capture them and then discuss them on list and at next week's telecon; comment quickly on both parts of process

PB: helpful to see all stages and points in proposal so as to organize my own thoughts -- at least the main points

WC: need to spend more time with proposal

DS: sent to list this morning -- will resend corrected version to list shortly

WC: just get initial reactions today

JW: like to say that I've reviewed the summary of the proposal--agree with GV's comments; can we clarify issues around tool usage and making sure that populations of web content developers will be ok to proceed along lines already set forth; should also work out what the resource constraints are and where we can fit that discussion into our process -- need to wait until full report available to asses usability; run study when WCAG2 ready to go to CR


  1. tools -- how are we going to do usability testing on the guidelines with or without tools -- how to separate?
  2. timelines -- where in the process does usability   testing occur and how does that fit in with going to CR?

WC: any other issues we need to take to WG?

AR1: is it possible to test students in concert with developers

DS: students of web design or students who use web content to communicate their ideas; students who didn't come from computing background but who happened to write web pages currently

WC: general issue -- identify audience -- groups being tested an open issue?

AR1: yes, many types of students -- do they really get a grasp of what accessibility is -- what are they doing a year from the first exposure to WCAG

// GJR's machine goes wonky/CMN picks up the slack //

// GJR resumes minuting //

WC: process -- timelines;

DS: six month project -- that's flexible; haven't approached any potential participants, plan is to submit research proposal for funding -- may take a bit of time; no rush as of yes -- happy to wait for a through discussion with GL WG; six month time frame could be adjusted

WC: six months from when funding comes through?

DS: yes

WC: info gathered going to be usable across WAI domain; talked about iterative process where we would give you an initial working draft, get feedback from you, then make changes

DS: proposal could be done with a small pilot group with some of the limitations pointed out this morning (all from UK, all with similar experiences, similar institutions) -- smaller group running through each iteration to see if viable and contains what they are looking for, and then feed that back to the WG; could be a pilot study with initial feedback --that's possible

WC: if could have separate usability studies going on in Asian counties, as well as Israel, US, elsewhere -- in general do you have feedback about other groups with whom we could coordinate testing

LS: could hook you up with Hebrew testers

// AR2 has already suggested contacts //

WC: will be translating documents, but not during the development of them

LS: can limit pilot to English speaking developers developing sites in non-English natural languages

WC: LS might be able to do informal studies; perhaps we can outline a streamlined process to give to people around the world

DS: a spin-off of our deliverables -- developing a methodology

AR1: have some really useful contact info [will give to DS]

PB: idea of breaking it down into 3 audiences of web developers, AU tool developers, and ER tool developers would be very valuable

// OPEN ISSUE: audiences and subjects //

LS: can take whatever feedback we can get -- what we do with feedback is up to us; if have feedback, can only help us; are we showing to people who aren't intended to read that level anyway -- feedback can only help; need to document the assumptions or framework that each "testee" comes with -- this is not someone who normally reads technical material, not someone who directly does development (of content or tools)

GV: usually what happens is that you discover things and share them at any time; have to have all diff users use it to discover if it is a good idea; usability feedback is valuable even if testing performed on small groups

KHS: ongoing process -- not something that's ever "done"

WC: need with each public WD -- every 3 months to see if changes made since last WD work; if resources of usability study takes 6 months, and we are talking about putting out a WD every 3 months, so having more tests would be good

KHS: Steven Pemberton's people -- good to have testing done in house as well

WC: series of open issues to take back to WG; proposal to read and review; what happens at this point is DS and HP move forward on funding front, while WG works on providing feedback; earliest discussion at next week's telecon -- probably won't hear from us for a few weeks

DS: in terms of funding, could there be a contribution from W3C or have to fund entirely by ourselves

WC: have to discuss with Judy Brewer

DS: not asking for a huge amount

KHS: new usability group under S Pemberton may have funding resources

WC: Steve Pemberton -- one of our hosts and chair of HTML WG -- has proposed that W3C start usability group to look at usability issues of the web -- as we are looking at accessibility issues on the web

KHS: his email mentioned funding

WC: talk with him at dinner tonight

AR2: accessibility is usually the prow -- leading edge -- of usability and productivity issues

WC: that's a whole `nother kettle'o'fish which I don't want to talk about now

MM: usability IG will be usability of specifications; web usability for users -- what's more important to that WG

WC: no--not usability of W3C content --that's Quality Assurance (QA) -- DD is co-lead of activity with Karl Dubost

MM: how to make authors use specifications and the language, versus how to help users use sites constructed using specs and markup languages

WC: activity similar to what we are doing with accessibility

MM: still a line -- who the end product is for

WC: QA looking at usability, accessibility, DI, and internationalization of W3C documents

DD: W3C Publication Rules being extended to include more quality work, such as readability, usability of spec, inclusion of tutorial, test suites, etc. -- usability of a spec, closest thing we have is UA work in WAI -- don't have a usability WG, which is what SP proposed

WC: is an IG, but yes

// the WG thanks DS //


// BREAK FOR LUNCH at 12:24 local time //

// CR takes a picture of the WG members in attendance -- a highly naturalistic pose, not at all staged //

// GJR shuts up and starts minuting //

// that last comment was for your benefit, Jan! //

// GV and JW rejoin by phone //

1:30 - 3:00 Testability of checkpoints

WC: next on the agenda: hour and a half to discuss testability of checkpoints -- in requirements document

WCAG2 Requirements: http://www.w3.org/WAI/GL/wcag20-requirementsURI of draft being discussed: 28 March 2001 http://www.w3.org/WAI/GL/WCAG20/WD-WCAG20-20010328.html

WC: Checkpoint 2.4 --

JW: useful to have a format -- separate question than separating minimal requirement from something more than a min requirement; format to be followed in document and distinguishing between minimums and more advanced possibilities; part of overall document? Useful to do first -- make distinction between minimal and more detailed or advanced implementations; will also want to maintain some sort of format so that information is structured

KHS: good idea because separating it into chunks is good not only for back end but makes it easier for people to use

WC: chunk but sub-chunk of a bigger chunk

JW: if did in XML, could have presented with or without labels

WC: easier extraction -- if just wasn't rationale, could just get rationale

JW: think it is time we do it

WC: hard finding minimum requirements

JW: for 1.1 we could say that the existence of a text equivalent which is related in markup or data model to that which it is an equivalent -- giving the same functionality or communicating the same information as the visual or auditory content -- could specify minimum conformance criteria for each checkpoint

LS: way to avoid this becoming obtuse is to put minimum requirements in technology specific techniques doc, so that has a practical application -- that's what minimum requirements are

GV: minimum requirements sound like P1 to me; saying that you have to have something in ALT is the essential part meaningful and useful is the second part; worried that we don't get to choosing what is easily testable and what is important

JW: not a solution but something that might be done

WC: where would minimum requirements go -- checkpoint or checkpoint solution level; minimum requirements are minimum requirements to satisfy a checkpoint regardless of priority -- minimum thing you need to do to satisfy a checkpoint; how to we state that something is easily testable

GV: could creep in if we aren't careful in formulation

WC: how can we state checkpoints in a way that makes them easy to test -- been restating them to make them easier to understand, but does that also make it easier or harder to test

LS: first, the process is technique specific/technology specific test cases -- have a QA person who sets up test

cases -- can volunteer that person to establish test scenarios -- should be done with black box testing -- regular paraphernalia of usability testing; should be designed for a QA department -- follow same paradigms that QA department does -- they do site testing, so let's give them accessibility criteria; most big sites have site testing

WC: big sites

LS; but if had paradigm on plate, they could use it

WC: integrate with WCAG?

LS: could have basic requirements which we could hand over to the A people and link to it

WC: ideally, as part of WCAG need to ensure that people can use tit to see if they have passed it -- having something separate doesn't make WCAG better

// ACTION LS: propose a QA activity based on discussion //

CMN: to make easy to see if piece of content meets a checkpoint, provide code examples that say "this is how you should code" or provide a functional practice/illustration; in technology specific work, that's what you want to do -- in HTML use the ALT attribute -- it is possible to describe the graphic in the text, but that is not a good practice because..." -- wooly phrasing, but need a "if I use this document, did I create accessible content"; will be examples at technology level where we say "use this type of code construct" at tech independent level, quite often you won't 0--provide functional description of what needs to be achieved

WC: both levels?

CMN: yes

CR: boils down to a level of detail -- more detail than you want to get into in a document like this, but to test, you have to be very precise

AR2: need to have at the highest level of conformance, something that is ecologically valid -- actual testing with a human subject who has the disability for which we are attempting to find solutions; if tie to technology, you're stuck -- that's what happened with WCAG1 with client-side image maps -- better technology became available, but still have requirement use client-side image maps even though something better superseded it; with technology at time of publication of these guidelines this is the recommendation for this technology, but real standard is will a PWD be able to use page

CMN: in making this thing usable generally, that always has to be our guide -- are we making this stuff accessible to PWDs -- have to be careful that what we are recommending is in fact beneficial; not all testers will have access to range of people with disabilities who will have access to those people's content; what would happen if you were disabled -- using a tool like the WAVE -- so that authors can perceive how PWDs perceive and interact with their products; take knowledge gained from testing so that the next person down the line who can't do as much testing, is still assured of the quality of WCAG -- distilling for them the user testing

LS: probably be allowed to go to the National Business and Disability Council might allow us to go in and perform testing there -- have almost every device available; if work out a QA paradigm that isn't technology specific but another that is non-technical, we might be able to do proper testing there for the paradigms that have been defined -- if WAI/W3C can join in, that would be good

AR2: human subject testing should be available as an alternative way to qualify a site or a document -- if someone finds a way to make a site accessible with technology that wasn't available at the time, should be able to claim conformance on basis of human testing

MM: only concern is that people who are seeking compliance with respect to legal requirements especially, would be creating things they believe satisfy WCAG, but which are inaccessible for other reasons -- legacy browser support, older computers, etc. -- having technologies that are proven to be accessible is good, but WAI needs to be on top of that, not delegating it to the general public

AR2: good idea -- if h ave alternative way to qualify a site, application, may need to have a filter at the W3C level so that it cannot be abused; could have some kind of volunteer organization in open source community style that would be available as a board of referees or jury

WC: not certifying anyone -- giving guidance to people who are developing sites, responsible for overseeing a site, -- how can we give the best guidance to people

AR2: what is necessary in the doc is to say head-first this is how you will test it with available technologies but ultimate goal is to make accessible to PWDs -- if you can show to us

GJR3: minuted by WC

CMN: difficulty I have with end user testing as justification is that it is extremely fallible -- example of a blind guy who demonstrated a site, but didn't show 2 or 3 major features, because he had no idea they were there; small group with limited testing service/skills/capacity is going to make mistakes; board of testers is a helpful service, but not rigorous enough to provide the information; we need a lot of end user testing to ensure that we've encounter all of the situations

MM: ultimately, the idea of an ombudsman wouldn't be for an org to buy indulgences, but to contribute to the form of the techniques document which are living documents -- need to work on ways of getting g people to publicly claim when they claim they have come up with a better mousetrap

AR1: problems with testing with deaf people is that there's no compact minority of deaf people having the same aims -- we are all individuals; should have end user testing, but has to be properly planned;

WC: not saying usability testing with end user is bad idea at all and should encourage more in document, but have to give guidance to people so that they can determine if they have met WCAG; developers of sites, developers of site construction tools, and evaluators

JW: question you originally raised was do we want to distinguish between minimal and advanced requirements for each checkpoint and that question hasn't yet been addressed; that's the issue you raised, and that's what we should suggest; LS at technology specific level, CMN at both levels -- have to decide if this is an appropriate distinction to make and clarify surrounding issues

LS: belongs in techniques level, but could also live in top level as well; end user testing can be a practical requirement -- though I'm not a QA person myself, from what I understand or have absorbed via osmosis is to create a paradigm that ensures that each route can be traversed by people with all types of disabilities; black box paradigm needs to be followed so that situation CMN described about blind demonstrator isn't repeated; wouldn't certify them,

JW: not talking about testing now, but whether or not we want minimum and advanced requirements for each checkpoint

LS: take consistent navigation -- if have testing paradigm that says ensure that someone from the following groups can traverse all of your site, that to me is a minimal requirement

JW: that presupposes that it is going to be tested with limited groups of users;

AR2: do see this as a matter of level -- at highest level should say that any technology specific tests are just proxy tests -- if can be shown that site is truly accessible, than using technology specific techniques could be bypassed

WC: last checkpoint solution is "here's what I did, here's -- would put at the front -- first of every technique; one example of human testing doesn't meet any of the psychological design criteria that are used in testing; technology based testing leads to inferior products, where ecological testing would have been the better more efficacious route

WC: we understand your point

JR: wondering if the best place to test minimal requirements is at whatever level you decide to put priority levels, whenever you get around to that

MM: have a structure of priorities in existence -- most prudent way to go about this, when apply priority levels, apply them to checkpoint solutions, then derive priorities for checkpoint level based on an extrapolation of that

WC: priorities are what you focus on

MM: priorities are "less than impossible" "this makes it better" this isn't necessary", those mean that P1 is a minimum requirement

WC: can have minimum requirement for each checkpoint despite priority level -- P1 is not the same as the minimum requirement we are discussing right now

CMN: requirement as used in ATAG and UAAG is "the minimum you have to do to satisfy this checkpoint" -- there precisely for testing purposes; need a baseline that people can use to say "I've reached the goal" -- prefer to have functional requirement for each checkpoint -- for technology specific things, code is better; where I'd like to see the strong emphasis on user testing is the "How to Use This Document" section -- test, test, and test again; do tests with users as minimum requirement is kind of silly -- if it works with a specific technology it might not work for others using similar, but different technology; testing is fairly complicated -- most people who will be using this document won't be experts or even competent testers

// WL rejoins //

GJR4: minuted by WC

AR1: question of priorities is still open -- testers in Finland need a minimal requirement; need technology specific resources

JW: issues: 1. value in specifying very clearly what is needed to satisfy a checkpoint -- not convinced that the idea of a minimal conformance requirement and a less minimal requirement is going to work without -- take 1.1.1 -- conformance requirement would have 3 parts: textual equiv for non-text content, 2. related to the non-text content for which it is the equiv; text equiv contains same information or performs the same function as the non-textual original -- tripartite criteria -- if you meet those 3, you've met the requirement; take 2.2.4 -- suggestion that there be a minimum of a 10 second delay before any reaction from user as an absolute threshold; varies from checkpoint to checkpoint; go back to 1.1.1 in the case of complex audio/graphical content give it a label, less minimal is provide a description or caption that conveys the same info (LONGDESC); can't actually provide caption s that convey same meaning as a Beethoven symphony -- not totally sure can apply minimas across all checkpoints -- is this a feasible approach? If so, need to assign some action items

WC: there are a couple of layers: guidelines, checkpoints and CP solutions -- difficult to define minimums for checkpoints, because they are expressions of minimal requirements, but might want o ad a functional requirement[WC explains functional requirements by walking through some checkpoints]; at checkpoint solution level, those might be the minimal requirements for the checkpoint level

JW: had an issue to write rationale for everything -- this is one way of doing it -- what and why in a couple of sentences; not satisfied that all of the conformance requirements or criteria that determine if CP has been satisfied belong at the checkpoint solution or technique level; could specify some of them at the general level as I did previously with my tripartite requirement statement for 1.1.1

PB: true that some checkpoints don't have minimums; introduces another level of complexity -- we have GLs, rationale/functional requirement, priority levels, cascading level of granularity -- all have benefits for good reason, but add to complexity -- as we are looking through ATAG I can understand why they are there , but it had to be explained to me, and if that is the case, that is problematic -- potential confusion between minimum requirement, priority levels easily confusing set of conflicts

LS: have 2 potential proposals to get around these problems -- my problem is that when people see minimal requirements, they see this is all I have to do and I don't need to bother with the rest" -- no matter how many times we say this is a worst case scenario; example JW gave where text equivalent is merely a label, undermines capture functionality in text equivalent -- may have lost where it wasn't necessary 2 ways around: "worst case implementation" instead of "minimum" and a "typical" which is what we consider idea -- there is no best case -- that's the one we haven't thought of yet; other suggestion is to provide this as a separate document so that we've removed the complexity -- that would remove PB's problem, keeping the mandate to state it clearly, but it is a trouble shooting document -- no matter what you do you can't get single-A, so here are some minimal solution s for you

CMN: sensitive to the comment about increasing complexity; ATAG is a challenge -- to use it, you need to know it and WCAG; in ATAG, we have a summed up version of all the pieces -- look at checkpoint, get idea of what asks for says why in a sentence, the least acceptable work in a sentence (agree that there is a risk that people will say "oh, I only have to do the minimum"); having minima helps because sometimes people push for a low minimum, but sometimes they push for a high one, so it is good to put them on the table to get consensus; 2 types of techniques -- implementation techniques and a new document containing "Evaluation Techniques" -- how can I evaluate and authoring tool to see if it meets the requirements; minimum requirements stuff is the BARE minimum piece of work you can do -- answers the question "how do I know I've satisfied the checkpoint"; we will never ever write GLs that people understand first-off -- have to give them a few bites of the cherry;

GJR5: minuted by WC

GB: when we look at WCAG as a whole and try to establish minima, in some items we may say in this situation you need to do this, and the minimum you need to do is this, which will make it more useable; that which is the minimum you must do is that which you must do to satisfy the guidelines as a whole shouldn't be any part of a P1 guideline that isn't essential; if that is what we are supposed to be doing, why are we having this discussion -- people must think that there is something that were are missing -- think that what you are looking at is sufficiency which is different from a minimum -- do the checkpoint -- there are many ways to meet it, and the following are sufficient, and here are some other things you can do -- if you did the following things, you would satisfy this checkpoint -- can be tricky; need to word original guidelines so that they are clear -- should be automatically testable

JW: GV already said part of what I wanted to say -- sufficiency is the concept, not minima -- the tripartite criteria are satisfied for 1.1.1 that is sufficient, but more could be done; we should go through the guidelines to see if a sufficiency requirement would work -- one solution would be to divide up sufficiency criteria -- maybe we need to assign an action item to go through

WL: you just volunteered for the action item

JW; well, not really, but if you want to volunteer, William

WL: priority level is the equivalent of what you just said, so I think the work has already been done


WC: only have an hour -- want to discuss some topics while JR and CR are still here; outstanding issues from previous topics will be dealt with by me

// ACTION WC: organize open issues and fruits of discussion into cogent form and post to list //

// WC discusses whether WL and GV will stay on the phone //

// WC comes clean -- we want to be outside //

// GV gives the plan his blessing //

Multimedia/Graphics Techniques break -out session detailed minutes

Present [compiled from Gregory's memory, apologies to anyone not listed]


LS: what would be very useful would be to specify colors that can be used to provide maximum contrast and minimize perceptual black holes -- 2 quite different color disabilities; leave the color to the rendering; what might be nice to do would be to use namespacing

JR: that's when color is important

LS: that's when color is important

JR: wouldn't want to change the color of a traffic light

AR2: redundancy of position

CMN: talking about both cases -- it is worth noting that there are 3 different types of color disabilities: diminishing perception of blue, red-green color blindness; grayscale vision; in the traffic light example, changing the color is tricky, but the redundancy is by position and brightness -- only one bright at a time

AR2: there are those that can't perceive brightness -- need degradation scheme to go from a photo to a single line drawing, first removing color, then scale, then anything other than the outline -- once you get something distinguishable at each level of degradation you've met the degradation criteria; another criterion is the cognitive components of vision which most people doesn't know about agnosia, but it is quite frequent; focal cognitive disability -- inability to identify any visually presented information despite normal visual acuity and ability; noticed in reference to icons, representations of realistic things -- editing a film by showing an eraser on a piece of paper doesn't work -- cutting is a means of editing film, so it would work; some icons you can make sense of very easily, but some that required complex cognitive processing/associating

CMN: this is something that SVG as a technology handles very well -- SVG Access Note contains some details about how to do that; what do you do with these technologies? What are the color disabilities? What kind of information should be presented? How do you draw a picture that communicates? Use of color is very useful for a lot of people -- what do we want to keep in our representations of visual objects; icons discussion seemed to lead to a preference that led to color coding -- each a different color

JR: process color rather than picture which can be much much easier

CMN: pointers?

JR: perceptual research -- they have numbers, which make for good minima

CMN: color disability information -- references scattered through the list need to be collated

JR: motion agnosia -- can't identify motion; facial agnosia -- can't recognize faces; reverse agnosia -- mirror vision

CMN: end up with a list of things you can't rely on, but we also need a list of things you can use

AR2: an outline drawing of a face with a specific expression can be perceived faster than a picture of a face

CMN: for things that are more abstract than Charles, where a photo is useful, outlines are faster to process?

AR2: yes

JR: no, color

AR2: yes, color is faster -- color, light/dark, outline

// ACTION JR: find reference and send to list //

CMN: associating motion -- is that generally a good thing, given the proviso you must be able to turn it off

AR2: people are too good with motion -- it blocks perception of everything else

CMN: association of sound

JR: will reinforce a visual thing -- having the 2 out-of- sync can mess people up

AR2: if have things that are expected to be vertical or horizontal, if are even a few degrees off, that screws

things up

JR: prettiness has a lot to do with symmetry -- if cut a picture of a face in half whichever one of the 2 sides is flipped, that will appear prettier than the original

JR: higher processing can be affected by degrees of brain damage

AR1: size of documents can have big impact -- long streaming video, the alternative can be as big or bigger than the original; subtitles are easy, but ASL or other sign language used as subtitle takes up a lot of bandwidth; synchronization is essential; ability to switch between subtitles and signing

GJR: aesthetic concerns of content creators can potentially be allayed by use of SVG which allows the user to restyle the graphic so that it can be perceived while allowing the author the requisite specificity as regards the designation of colors, shapes, etc., with the added benefit that (a) the graphical content is scaleable, so it will not distort if enlarged or shrunk and (b) SVG has a robust title, description, and text-flow models

CMN: right, but the underlying principle is that one must have access to content in order to change it in order to discover what the content is

AR2: semantics conveyed by layout can be misunderstood; when reading from right to left, placement of iconic information can be confusing

CMN: expressing relationships in SVG -- RDF and other technologies allow us to express those relationships; browser takes that relationship info and renders it graphically, which is a useful thing for someone who can't see original, but also because it allows you to transform it appropriately

AR1: metadata important -- can help people discover what they are missing or what the author intended; but as GJR pointed out several times today, writing such info is an art

$Date: 2001/08/08 23:06:01 $ Wendy Chisholm