06 February 2001 ERT/AU telecon meeting

Summary of resolutions and action items

Participants

XML Authoring

JT We are working on 2 deliverables. 1 - issues for next version. 2 - organize techniques into various views. As part of 2 we've been looking at various authoring tools. One suggestion is XML tools. How well does our document handle XML tools.

/* lauren leaves */

Resolved: AU will discuss XML authoring tools at the 20 February 2001 meeting. Lauren will attend.

AERT integration into ATAG-TECHS

LK We have a variety of techniques, some are heuristics. How will these be handled in AU to determine a level of compliance?

GJR So far, the need to repair has driven the priority rather than how complex or difficult it is.

CMN The priority is primarily based on WCAG priority in terms of what have to do.

WC Not level of priority, concern over normative vs. informative. Techniques are informative. AERT is smorgasbord of informative stuff.

CMN To be normative, have to demonstrate that if you don't do it access is not possible.

GJR Example is the alt-text registry.

CMN For most of the ATAG life we said, you must be able to configure prompting. We dropped that since it was not a requirement for accessibility.

WC Does this answer our concern, LK, that we can suggest anything for inclusion in ATAG-TECHS?

GJR Oh, your question is if AERT become normative?

LK The simplest test is to check alt attribute to see if alt-text is present. Do I have to do that to conform to ATAG?

CMN No, can give the author a note that says, "Please check this."

LK All an author has to do is to have a single sentence that says, "please make sure that everything in WCAG is satisfied."

GJR There is also guiding the author, providing documentation, alerting. It's both these are things the tool does, and other things the author needs to be aware of. What if you have a tool that includes a copy of WCAG, does that conform? no.

CMN No, that does not meet the requirement. For example, what Bobby does: here is what we've checked, here is what we have not checked. That would pass. There are automatic checks that tools may not implement. If you say, "make it accessible" that is not checking the problems but telling the author to deal with it.

JT The criteria is that the tool will create an accessible page and should therefore be an integrated part of the tool. Dependent on user interface of the tool.

GJR Feedback from developers has been that rolling AERT into ATAG-TECHS gives guidance that they felt was missing.

LK If I create an authoring tool and it does not flag images without alt-text, can i do that?

JR OK to have manual checks, that's not the same as "go to WCAG" as long as you have an itemized check.

LK You could have a wizard and each dialog box quotes a WCAG checkpoint, and it says, "check that all images have alt-text." once click OK button, get next quote.

JR Yes.

LK That's not much more that pointing to WCAG.

JR It's less than a providing a list.

GJR No, that fails. It has not looked at your document.

CMN No, it's walking you through WCAG step-by-step.

LK Some people want to see all in a checklist.

JR What would people rather see?

LK In terms of passing ATAG, this will pass.

JT It would not be integrated into the user interface.

JR But that's a different checkpoint.

GJR Integration is essential.

LK This is an extreme example, and someone wouldn't want to do this. This sounds to me that we aren't requiring any automatic checks.

WC This is just a test case to determine that any crazy-automatic check that the ERT WG may suggest to AU for inclusion in ATAG-TECHS will not be required by anyone to implement.

JR There is a special checkpoint that says you have to prompt for alt-text.

JT Right, Techniques are not normative.

GJR Are there any places in ATAG 1.0 where stronger verbiage is needed, like the special checkpoint for text equivalents. For example, the minimums in UAAG.

LK My personal opinion is that if you have a checkpoint for evaluation techniques, then it ought to have some substance. e.g. checking that alt attributes exist.

GJR We ought to attach this from both points of view. UAAG says "these are the things that must be exposed" and ATAG then says, "these are the things that must be checked or included."

LK There is a checkpoint that says, "detect automatically where possible."

GJR More comfortable with minimums or maximums?

JT The developers feel strongly that that is difficult. Dependent on the style of the interface. Not enough developer freedom.

WC That's what they always say. <grin/>

GJR They are balancing deadlines and all kinds of pressures.

JT Authoring tools are too varied. Can't generalize.

WL MS Word can save a document as HTML. To accomodate 508, MS could take that feature out and say "We don't have to worry about ATAG-like issues now."

LM The principal is that any product that saves as HTML will be market driven. NotePad saves as HTML, will it incorporate a process? Probably not, due to the robustness of the product and the purpose of it. We'll look more at VisualStudio - authoring environments.

LK 4.1 says, "check for accessibility problems"

HB Where possible.

LK Do developers like that kind of requirement or does it leave them never-knowing when they have satisfied it?

GJR Techniques will eventually lead you to AERT where it clearly says what can and can not be automated. Give them a variety of choices.

LK In effect, this says, "implement everything that is an automatic check in AERT?"

JR "Where possible" it doesn't say "where possible to automatically do it."

LK Where you have the development resources.

GJR Or feasible for the product.

CMN Essentially we allow some weasle space. On the one hand you can develop a tool where you can not do any checking if it fits into 3K of memory. We started doing product evaluations to give a sense of what the working group thinks does or does not meet the requirements. What are the acceptable ways of satisfying the checkpoints. Doesn't make AERT normative. Some are straightforward and give a reliable result. For a lot of them, they are neat ideas but they do not give you a firm answer either way. However, help the author. ATAG applies to a variety of tools generating a variety of content.

LM In theory, an authoring tool can provide an interactive checklist. However, if FrontPage does that we would not conform?

GJR Where appropriate with the look and feel of the tool. If you provide feedback on other things should do it for accessibility as well. e.g. in notepad wouldn't expect a pop-up window, but in word you might want or expect that.

LM The more the robust the product, the higher the expectations.

GJR Notepad with a copy of WCAG might meet Level A, if you stretch it. However, single A not be popular with disabled users. Integration are priority 2 checkpoints.

JR Everyone must remember 3.1, which takes care of a lot of the things we want you to check for.

LM I can prompt the author a the beginning when I load an image. You're talking about an image by image basis.

JR I agree that there should not be a time to fully disable. We need to talk about this.

WL There has to be an "it" there. We have spoken about, "you can disable it." However we require there must be an "it" there. Being able to disable it and controlling when applied, but the function has to be there.

LM True, but if I'm going to say, "the it has to be there" I might as well as how it should be implemented.

GJR If you're not on a network, then the tool may not be accessible.

LK Conclusion?

LM Discuss it further.

LK What do people see as the points of debate?

CMN The specific question asked by Lisa is one that came up specifically during the last call. There is a question from Adobe.

GJR It's like the seatbelt law, you should wear it but don't have to.

CMN Right, we can't force anyone to use anything. My first impression of LK's question was, "is ATAG-TECHS normative b/c if so we will be careful what we put in." However, they are not and we want as many techniques as possible to let author's decide what is reasonable and feasible.

LK In 4.1, when it says, "check for an inform author...where possible." This does not, except where mentioned otherwise is a judgement call as to what is reasonable and feasible in context of the tool. The conclusion then is not sub-checkpoints under that.

GJR Are you asking us to add minimum checkpoints or to make normative more techniques?

LK I can not speak for ERT WG, but personally it makes me uncomfortable. It seems that "check for" does not have a well-defined meaning if you say "do what is reasonable." it would feel more comfortable to check for prescence or absence of an attribute. Then, at a minimum warn the author if any of those are absent. so that a developer knows when he has accomplished it.

HB To help the author deal with it.

CMN Accessibility problems defined, basically WCAG. Tool has to check or ask the author. Comes from checkpoint text. If you think that it doesn't then any person in the universe can send comments. My personal check is getting into formalized subcheckpoints will be a major hassle.

LK Checkpoint stands and interpretation is minuted.

Resolved: no change to checkpoints in ATAG concerning checking for accessibility issues. Interpretation of how this applies to a variety of tools is as minuted here and discussed in ATAG 1.0.

AERT open issues

LK How are these going to be resolved. We've started to go through them in ERT WG. There are about 2 dozen left.

JT Have we separated out the WCAG issues?

LK Yes. Where is the final list?

CMN ERT has an issues list for AERT. Expect that ERT come up with techniques - you're the experts.

LK Will you be content getting resolutions on issue even if they come from a small subcommittee?

CMN Up to ERT to decide. We just like having the content provided. Be aware that these are not normative and welcome as many suggestions as possible and will not say "you must do this." Look at the incorporation draft, it says, "if you do these things, we reckon you've met the checkpoint."

LK Then we can close as, "it might be figured out, we haven't, there you are."

JT Obviously, the more you figure out the better.

LK The main thing is not to have anything that is wrong or misleading. However, having figured out every last check we'll go as far as we can.

CMN If we don't pick up that it's wrong and include it, it's our problem.

LK We will not deliberately include anything wrong.

Resolved: ATAG Techniques are not normative and should contain as many ideas as possible, therefore open to all suggestions by ERT WG. It is up to ERT WG to decide how to pass info on to AU WG and what info to pass for inclusion in ATAG-TECHS.

Next Meeting

CMN It's the joint meeting at the face to face, the first Thursday at the month. Do we want a follow-up meeting after that?

LM Any more feedback on if we can call into the meeting?

CMN It's appears to be technically possible. We don't know the final answer to. We have to check if each group can get a space that is callable. Microsoft is saying they don't want to turn up but dial in.

LM I wouldn't have said it that way, but MS is saying that no working group members attending any WAI meetings.

CMN We'll have to check across the working groups. Given that early hotel registration closes soon.

Action CMN: get a genuine answer in the next day about telephone participation at the face to face meetings.

LM If we can't dial in does not mean that we will show up. Fighting this decision.

Resolved: next joint meeting is the face to face, 1 March. After that, the next joint meeting is Tuesday, 3 April 2001.


$Date: 2001/02/06 23:31:44 $ Wendy Chisholm