29 March 2001 WCAG WG minutes

Summary of action items and resolutions


jw: agenda is 1) consider general issues with latest draft of WCAG 2.0 released yesterday and 2) discuss checkpoint 3.5 - write clearly and simply

wc: issues concerned about. action item to write introductions to each guideline. seems displaced to write intro after guideline. instead, added overview in introduction. became very metaphorical. is this the right direction? would this be more appropriate in executive summary?

kh: likes categories added to each guideline.

wc: suggestion from f-2-f

wc: still need to update checkpoint mapping

wc: edits are all marked and listed in change log

wc: checkpoint 2.1 reworded. used to say "provide a variety of navigation mechanisms". too vague. new wording is "provide more than one path..." At CSUN, disussion about providing more than one path may not be appropriate for all sites, especially if they are very simple.

as: what is the rationale for this checkpoint?

wc: different people find things in different ways

pb: the way 2.1 is now, read it that any one of the suggested methods would meet the checkpoint

wc: WCAG 1.0 13.4 - use navigation mechanisms in a consistent manner. WCAG 2.0 history is that this got split into 1)provide a variety and 2) provide them in a consistent manner

jw: primary driver is for cognitive disabilities. propose we limit it to relatively large and complex web sites. must be defined; i.e. more than xx documents and linked with great many cross references, structure is reasonably complicated. must do research to decide what complexity threshold must be.

wc: WCAG 1.0 defined "important". people had a very hard time dealing with it. maybe something like this would apply to some of the other checkpoints as well such as the one on writing simply and clearly. at f-2-f meeting in October, at device independent workshop, Tom ? said they found it useful to categorize web sites and apply guidelines based on category. seems complicated but keeps coming up as a possible solution.

pb: one method that stands out in 2.1 as not having to do with cognitive disabilities is the skip navigation links method. concerned that this one will get lost and not correctly categorized.

jw: trying to get rid of legacy issues

gv: where is the skip navigation links requirement?

as: in 2.1

gv: not sure this really fits under this guideline.

kh: how about "provide more than one path .... for ease of navigation"?

jw: invite proposals on how to handle 2.1.

mm: design of everyday things by Norman. decision trees and how to organize them. seems like that's what this checkpoint is trying to get to. e.g. if you have something that is 100s of items and 2 levels deep, do A. if you have something that is 1000s of items and 3 levels deep, do B.

gv: suggest removing 2.1 and 2.2 because have nothing to do with accessibility. they are usability

kh: why shouldn't they be in there?

as: because it confuses the issue.

gv: these are here for cognitive reasons. in example above, "complex" is different for someone who is profoundly retarded. in cognitive, no matter what you recommend, you will leave someone out. if pursue putting these things in, will come up with something that is not only untestable. you can't even agree with yourself.

gr: have a list of guidelines for accessibility, not usability. what we are thinking here is that the ultimate usability of a site is the baseline for our work here. Steven Pemberton (?) from f-2-f looking at creating IG on usability. for now, we are stuck putting in things that are really "usability" but have a quantifiable benefit for accessibility.

gv: but no one will legislate that a site be usable. accessibility guidelines or derivatives of them will be legislated. propose we remove usability items out of the checklist. if we do this, we will eliminate anything that we have now for cognitive disabilities. dilemma. could take usability things and put them in a separate section - general usability concepts, interact strongly with accessibility

kh: would this be in the normative document

gv: yes

gr: need to define accessibility and usability if we take that approach. muddies the waters a lot more than addressing some of the usability because accessibility is a subset of usability.

gv: usability makes it easy to use. accessibility removes barriers

bb: accessibility is also about compatibility which is not usability

gr: if we don't address these issues, we're not addressing accessibility

jw: should continue but need to address Gregg's concern about accessibility issues that have to do with cognition. no matter what you do, there will always be some people with cognitive problems that will not be able to access it. cognitive-related checkpoints can improve acess but are not decisive. is there a way of developing these in a way to say "the more of these you implement the more accessible the site will be. assign priorities based on type of web site being developed." unlike other checkpoints where meeting the checkpoint means it is accessible and not meeting the checkpoint means it is not accessible.

wc: need to provide rationale for each checkpoint

gv: all of cognitive ones will benefit all disabilities. can say you want a site usable by blind or deaf users but can't say a site will be usable by people with cognitive disabilities.

wc: because cognitive disabilities is such a broad category. you could pick specific cognitive disabilities and make a site usable by that segment. Ann has a lot of sites that are just pictures. the reason she is using them is educational. there are a lot of people who communicate through images. could provide guidance on specific cognitive disabilites.

gv: information about how to design targeted sites for people with cognitive disabilities should be taken out of WCAG document and published in a separate document. not recommendations for all sites. W3C can decide whether or not this document should be on a W3C site or somewhere else.

wc: we've learned so much about other disabilities.should do the same for cognitive and learning.

gr: same conversation we have been having since the winter of 1999. don't believe there is a strict demarcation between usability and accessibility. talking in last half hour about an impact matrix more than we have about accessibility guidelines.

mm: issue here is that when it comes down to cognitive disabilities, changes are made to content itself and are not quantifiable. with vision and hearing disabilities, the requirements are well defined because you are really targeting a "tool" that the user uses. with cognitive disabilities though, you are targeting the person.

gr: i was not born blind so my abitlity to conceptualize is different from someone who was born blind. we should honor the work that has been done over the past two years to try to solve these issues.

jw: is there anyone who wants to remove the usability issues from the document

gv: i argued strongly to include cognitive requirements in the document. keeping something in the guidelines should not be decided by how much work was put into it. been wrestling with the cognitive part a lot. not able to find a good solution. nobody appointed this committee to develop usability guidelines. some of the things should come out of the guidelines. "clear and simple" language checkpoint should be removes. need to separate those things that are absolute and measurable and people should do from those things that people should work toward. can leave in but in a separate section. we are watering down the other things by keeping these things in there.

mm: design practices document

gv: if pages are terribly laid out, it may be impossible for someone to figure out with a screen reader

gr: have to be adept at reading document source

jw: idea - move cognitive issues into a checkpoint "these are factors that need to be taken into account in designing content." In some instances there are measures - more than xx documents and yy links, complexity threshold exceeded. problem is that no matter what you define, you will leave out some people. people in cognitive disabilities communities are distrustful of numerical measures of reading grade level.

wc: checkpoint we hear the most complaints about in WCAG 1.0 is the "write clearly and simply" one. uncomfortable removing or moving cognitive things just because it is hard to do

jw: not that they are hard to do but hard to define when you have "done" them

gv: they are impossible to do. can't put something in guidelines that can't be done in all places on all web sites. if guidelines are "advice" you can write them a particular way. if guidelines are for compliance, people must be able to prove that they have met them. could put them all as priority 3 but that doesn't seem to be the right thing to do.

gv: can require that somebody require alt text. can't require that the alt text is clear and simple

jw: never had trouble with "write clearly and simply".

wc: in 1.0 don't have tests required to ensure that you have met a requirement

gv: who decides if it is clear and simple? author. fought for this to be priority 1. want people to try and do it.

gr: can prompt and assist them in making that judgement.

as: agree with Gregg that cognitive/usability issues cloud the issue. in IBM, I have to be able to tell someone how to test/validate that they have met a requirement. these are not testable.

pb: agree. could separate into objective and subjective elements.

gv: might be a good exercise to look at what we have and categorize as objective and subjective. can we come up with an objective measure for them. we might find that some we think are subjective can be made to be objective.

jw: checkpoint 3.5 - say you had a numerical measure of readability from 1 to 100. If you test a web site and come out with a measure of 65, have you met the requirement or not? if lower from 65 to 60, does that have an impact on readability for people with cognitive disabilities or not? could set a target but not achievable for all web sites.

pb: leave in subjective area and not try to come up with measure for it. allows you to include fuzzy areas in the document with the recognition that it is not an all or none criteria.

jw: still concerns that division is what we can implement and what we can ignore.

wc: but that is already occurring. Bobby has so many things that have to be manually checked. Most people just ignore those and say they pass.

gr: Bobby approved with subjective hat

gv: could have q&a session. answers to questions determine whether you pass or not.

wc: must be able to automate or describe process of how you test each thing. biggest thing that is not included in WCAG is what is the process and tools you use to assess a site

lg: whatever we write, the people who are developing the web sites will have to understand what we mean. we have things in mind when we write the checkpoints but it may not be clear to the web site developers

gv: how do you apply "write clearly and simply" to quantum theory? very few people in the world can understand it.

wc: WCAG 2.0 is not written clearly and simply. not trained to do that.

lg: most people in the world are not trained either

wc: impact matrix

jw: go through the draft and indicate where tests are required

wc: have already for WCAG 1.0 with AERT

gv: need applied to WCAG 2.0

jw: indicate where there are well-known criteria and where there are judgements

gv: create parallel table to use as we study it

jw: with 2.1 for example, there is an issue of when to do it. is it applicable to your particular web site and have you achieved it? a lot has been done in AERT document. not always technology specific.

wc: matt is going into AERT to create HTML techniques. Jason is suggesting that we go through WCAG 2.0 and split them up into things that are subjective and objective. think that this can only be done at technology level.

jw: automated is totally different thing

wc: in ER it is the same.

gv: things that are discernible by all of us that are not automatable

pb: quality of alt tag is subjective

wc: i have written alt text that people have told me is "wrong"

gv: just decide if it is passable. it can always be better

wc: still subjective

gv: think we could have a set of rules that would work

jw: reasonable semantic relationship where most people would agree on whether or not you have met the rule

wc: who will take the action item?

action: gregg, jason, and andi will take a pass through the document and state whether something is objectively discernible vs. not objectively discernible. also determine whether something is accessibility or usability. due 4/12/01.

wc: still have all issues that were on the mailing list and Kynn's rewording

jw: most of the issues are related to this

$Date: 2001/03/30 15:24:32 $ Andi Snow-Weaver (scribe), Wendy Chisholm