15 February 2001 WCAG WG minutes

Summary of action items and resolutions


Gregg, Cynthia, Wendy, Jason, Marti, Matt, Len, Loretta, William, Katie, Dick, Bruce

User agent capabilities

JW Issue still exists but latest draft may have changed context. Concerns text immediately after guideline 1. Capabilities of device refer to device currently have or able to get. I think the way in which the issue was raised is a product of earlier drafts. The mention of device capabilities used to be in the text of guideline 1. It's now in explanatory text under guideline. Therefore not normative. Still relevant to checkpoint 1.7.

WC Disagree, applies to all of guideline 1. Let's just answer the question.

JW Even tho the question may occur...

WC Think it's part of stating our assumptions.

CS Commonly have argument, this won't work in netscape 3 - make them upgrade to 4!

JW Same issue as far as how far backwards compatibility extends.

CS Some new tools won't run on 486. In the U.S. not unreasonable, but not sure. We have to balance feasibility needs of author: can't say has to work in oldest browsers vs. only ie5.

JW Want to avoid making a semi-arbitrary decision.

CS It is arbitrary. Look at usage numbers. That's how people make decisions. There is not a "right" answer.

JW One approach might be, these guidelines are accessibility. Then say that in relation to disability issues then if there are implementations that support a feature that will overcome an access problem then that solves it so far as the disability related character of the problem is concerned. Then becomes socio-economic problem of people upgrading. Once implementations as far as access for PWD, the problem is solved and content can make up own minds how far backwards they want to support.

CS That's reasonable.

WC Multiple platforms?

WL Absurdist position: hire someone who has the latest equipment..availability is complex.

WC Platforms? Language?

JW No longer accessibility question. Perhaps say backwards compatibility is important but not part of conformance requirement. Not intrinsically disability related matter. There is an assumption that someone needs to be able to use the technology independently.

WL Persuaded me.

CS Good idea.

KHS Makes task feasible.

LK Related: should blinking be minimized or avoided. Minimizing optimal accessibility with other factors. If we apply to blinking image...

JW If a particular feature is implemented in latest versions of 1 or more user agents, then become more precise for technologies and compatible with assistive technologies then we can say the backwards compatibility feature has moved outside the realm of access and then problem with people upgrading.

WC Therefore, as a concrete example, I use tables for layout. Home page reader exists and can navigate the table. Therefore, I can use tables without worry.

JW Yes.

LK Someone has a linux program that can read alt-text in images. Would we drop the requirement for alt-text for images?

JW Depends on how you want to say, "this is the requirement you have to meet before we regard something as being implemented." Therefore, how would we describe the situation.

WC various axes:

LK Different thresholds for different sites. e.g. government has to accommodate for much further back, vs. commercial entertainment site. Also, different countries. We draw a line and let users draw their own.

JW We draw line at latest, broadly with considerations of platforms and assistive technologies. We provide people info to let people provide backwards compatibility route.

LK That should be provided in conformance declaration rather than something to do on the side. i.e. W3C+these enhancements or W3C-these exceptions.

JW Good idea. Then you can say, "we're conforming to this and we're only taking into account the latest versions or what's been around for 2 years." although want to make it more precise.

MM Backwards compatibility endorsement on top of compliance. We have compliance levels based on: it's impossible to do things w/out this level of compliance...we're dealing with audience. If we take the technology and separate it and cast a wider net. If you have an endorsement system for backwards compatibility people can expand definition to cover that.

DB A level of compliance you can climb up if you say "we are also backwards compatible."

JW As far as conformance is concerned you can say this is accessible to people with disabilities under a certain level of conformance, that they have the latest versions of user agents and ATs. Then define platforms, etc that should be supported. In addition in conformance statement additional claim to make that ensure backwards compatible to support technologies that have been around for more than x years. Basic level of conformance does not require backwards compatibility. Then people use latest technologies if want to. Additional access to claim conformance.

GV Doug Wakefield on other line, asking about questions we're discussing. Could repeat what you said. i will pass along questions he had.

DB What are the practical implications? They would no longer mention script? If they do have a way to be backwards compatible with scripts, say that in conformance statement?

JW Depending on definition we put forward.

GV Can we take the same topic, so we're not talking about specific technology. If you have XML, where it does not make sense w/out a style sheet, saying it needs to work without one it doesn't work. Unless change it on the server side. If say, have script but always render so that screen reader is one thing but that it must work with lynx is another thing. they can't handle scripts. there is an issue of, will we abandon the position where lynx is a test for accessibility. Will pages be getting so interactive they resemble programs rather than documents. Most pages still look like documents.

CS I would venture to guess that most of the hits on the web go to those that are more programmatic rather than document.

GV Increase in bandwidth and other factors cause this to further grow. You can claim conformance if accessible to latest technologies and ATs?

JW Yes, plus additional claim about backwards compatibility.

GV If I have a $5,000 talking browser that a page could claim accessibility if it was accessible using my browser.

JW No, that was the other issue we discussed. We would qualify what we meant by "latest version" based on compatible with ATs, browsers, platforms, etc. Then we would separate backwards compatibility issues from that.

GV That begins to address the issue, but it is normative and then how do we make that list? Means the list doesn't change. The list of criteria, but no one knows what it meant.

JW I don't think that will be a problem. You can either put the criteria as normative and provide informative list of how it applies today. That could be updated. Like techniques. Or user agent support page. These are the criteria in the latest version, we provide documentation that tells you if that is the case or not. If you disagree with our assessment you apply the criteria independently.

GV That sounds complicated, but I don't see another way to do it better. Let's think of how to simplify it.

DB The practical implications are tricky.

GV Political reality: dangerous if we say there are 2 expensive screen readers that accessibility depends on. Not sure that it's practical that pages need to be made accessible to screen readers that do not have web savvy built in.

JW We could advise people to be backwards compatible, but we don't preclude conformance to older technologies.

WC What about 1.7?

JW Advisory.

GV Or priority 3. Then may cause us to redefine priorities.

JW Low priority and can specifically state.

WC In ERT, writing a language that can let someone claim above and beyond Level A or some level of conformance. Therefore, this would not be the only checkpoint that one could claim about. I am Level A plus these 5 checkpoints.

GV If there are too many at each level, it gets too cumbersome. I am level A plus these 18.

JW Or, I am AA minus these two checkpoints. That would be equivalent to A + 7 checkpoints. Could restrict, can't claim anything below A.

GV Concerned that that is false advertising.

WC Need to go through to determine our assumptions.

DB That's my action item from a couple weeks ago. I will begin work on that.

JW Are people comfortable with handling the issue by stating conformance claims in these ways? If so, then what are the minimum requirements? If claiming conformance to only latest versions, what will be the requirements?

WC Not comfortable with first part: claiming conformance based on latest version. However, like the idea of conformance in steps. Should move forward, but cautiously.

MM Need to state it simply.

JW Move forward with idea and test.

DB Will it make a difference? Don't see many conformance claims. More important to determine if have met a particular checkpoint.

LK I hope that people who say, "I can't do this AA thing, it's too hard." Then they ignore the rest of the checkpoints. Presumably it will help them if they make a claim, but even if not, might help the mind set. See that there is good in going past A even if don't reach AA. Just a hypothesis.

JW Yes, Kynn has emphasize this. It could operate as a disincentive.

CS At MSN one thing I struggled with was become compliant and filtering out browsers that we had decided not to support.

JW If we go this path, it's reasonable to support a subgroup of user agents. Then go through checkpoints to determine how this criteria would apply.

GV A language to use would be "recent" rather than "latest." On AT say recent? May be web-aware. The thing to do is to build the list to see if we can move forward.

JW Also, development of techniques.

Report on Access Board questions

GV Was asking about screen flicker. Is it talking about blinking or flicker? What do authors have to control? My response was that we don't have control over hardware flicker. Animated gifs, authors can control rate. Flash and other programs, no way to turn it off, authors have control. Times when author create a page and if moving stuff around may flicker. If know happen, should control that. Should not purposefully flash something, particularly at bad flicker rate. WCAG also has stuff that 508 has chosen not to address. e.g., animations can be distracting. It's not a seizure issue but drawing attention away. 508 does not address any cognitive issues.

The bigger issue is forms created by XML where style sheets are required. I asked if also running scripts and if the form is interactive? If so, have a form of this that could be used by people using older browsers. He raises the same discussion we were having. At some point, don't we assume that people have more reasonable browsers?

For employees of government, if part of their job is to work with the Web, you can assume they have up to date browser and screen reader. However, public sites, like the IRS, you can not assume that people will have the latest and greatest. Therefore, what is it safe to assume that they do have? We have not defined that in the past.

In our discussions, discussed making 1.7 a priority 3. It seems to me, that not supporting new technologies and viewable even if it doesn't is either an accessibility issue or is not. If it is, should be at whatever priority, if not should not be in here. We are now changing the definition of priority. Somehow we need to future out how to separate pure accessibility issues from doability.

JW Yes, that was the early part of the conversation. Once a feature is supported to a certain level it no longer becomes an accessibility issue.

GV What if we did it with plus: I am AA+. Not only was it accessible but made it accessible to people with older technologies. Can't get plus unless works on older technologies.

LK Yes, that's what I had in mind.

GV Then can go right on the conformance logo.

DB Then figure out 2 levels for each checkpoint?

WL Only work on older technology.

WC Beyond A but not AA.

CS Need a way to say I'm taking into consideration older technologies.

WC But that's a checkpoint.

GV AAA as well as latest and greatest.

Synchronization of audio only events

WL What are we trying to achieve.

JW GV Raised this issue before the draft became public. When have a media specific event that requires time dependent response, do equivalent shave to be synchronized.

GV An example, if the thing says "do such and such hit it now, do this then hit it now." The transcript does not help since don't know when now is or was so do n't know when to do something. if translate to a link then one way to satisfy. otherwise synchronize audio with interaction.if there is time dependent interaction, got dropped out.

JW My non-chair reaction is that wouldn't it be preferable that instead of adjusting synchronization that you must provide alternatives to time-dependent interactions. We already have a requirement that requiring someone to respond to something is an accessibility issue. There should be a way of having a version that avoids that which avoids the synchronization issue. Therefore make that clear rather than adjust synchronization requirement.

WC Agree with that.

JW Never guaranteed that person with disability can interact in time.

GV That is kind of like saying we don't need to worry about large print b/c can use as if they are blind. there may be people who are deaf, with good reflexes, who would want to use regular interactive game. but can't b/c you didn't synchronize the captions.

JW The problem we ran into last time was getting the language right.

WC GV's proposal was: 1.2 Synchronize text equivalents with multimedia and time-based interactive presentations

WL Is it possible that timing itself is content?

CS Interesting way to look at it.

JW Add wording into checkpoint and add example.

GV Something just occurred to me: if basic multimedia, already covered. Only time need this is if there a visual-only with a response that had to response or an auditory only where needs a response. I have not seen auditory only and that people who see are drive to present things visually. will we do anything as a pure auditory "hit now."

JW I have encountered one, but not on the web. A piece of software to measure people's reaction time to auditory events.

GV Right, and we'll caption our spelling tests as well.


GV Perhaps it is so unlikely to happen, we don't have to worry about it. e.g. "are you still there? if not i'll reset the page." Usually people have to put up something visually. Therefore, would be surprised not be synchronized.

JW Press the space bar when the sounds are equal volume.

GV A hearing test. Not necessary. Is there ever a time when something is presented visually and you are supposed to react when something visual happens. Click when proper one is displayed.

JW If you see them you can interact with them, if you can't then no reason to have time dependent interaction.

GV Right. Usually need mouse. Let's raise this again at the end of another session to see if anyone can think of an occurrence of when this will happen. If we can't, then we can drop it. I'm feeling it may be esoteric or hypothetical.

Resolved: Issue 41 is still open, although if no one can come up with a real world example we will close it out.

$Date: 2001/02/15 22:47:01 $ Wendy Chisholm