WC I had an action item from the joint AU/ER meeting: WC take idea of other types of evaluation tools for ER. What exactly do we mean by evaluation? Does it include evaluation of tools or limited to evaluation of content?
CR our work is based on WCAG and therefore on content and not on tools.
WC CMN made an interesting point that if AU and UA become the interest groups for ER WG, that the division between the groups seems to be: "IGs" create normative guidelines, ER creates tools and documents techniques. therefore if an "IG" needs a tool they come to us.
WL Putting a tool on our list is an evaluation?
CR we put it there but not evaluating it.
WL we're not putting an evaluation in any other way.
DB Not sure we should evaluate tools. Not sure what we would base our measures on.
WL conformance testing. We are publishing reports on authoring tools.
DB that's how it conforms to AU guidelines.
WL The only tools that we could evaluate on this list are tools that are evaluators of themselves.
Consensus: we will focus on tools that evaluate content rather than tools that evaluate tools.
HB valuable to annotate each of the tools as we have done.
WL We're evaluating but not rating.
CR if we're just looking at content, we do look at tools. Where does Bobby fit in? It's not an authoring tool but it's evaluation.
WL It's on our list.
WC combining tool list with AU. therefore, bobby included in with those.
WL The question on the table is do we evaluate tools.
MC I like the idea of tool developers evaluating their own tools. If we have a set of test pages for people to run the pages against, that will add consistency across the evaluations.
HB We add to the suite of tests.
WL That in itself is a tool.
MC That is another representation of the AERT guidelines. We describe, this is what a tool should do and if done correctly this is what the result of the test suite should be.
HB Then anyone who has built a tool could contribute their test suites.
WC Are we suggesting that we will build a test suite? If so, does it include AU stuff?
MB For AERT we are dealing with an HTML page, ATAG goes beyond that.
HB If in evaluating a document, if we could identify what tool was used to produce it, we could report that in the evaluation of the page.
MC We've discussed that recently. The reason for doing it is that if you discover that tool X generates alt-text incorrectly in one way, you could easily search for tool generated problems.
HB also an AU issue.
WL Perhaps a Priority 3 guideline in WCAG to identify what tool you use to produce.
HB However evaluation tools may strip that and insert that they created.
WC Variety of ways that they can include who created it. META, comment, or DOCTYPE.
Action: WC take to AU: we're looking at content. we're going to develop a test suite to evaluate content. This will help AU tools evaluate ATAG checkpoints that apply to WCAG content. If you want to include ATAG tests in the test suites, then AU needs to do that work.
MC Reason we use test pages rather than URIs is that URIs might get fixed and then we'll lose our test.
WL Does Bobby use something like that?
MC in theory, yes still working on.
WC 21 action items! Many people are not here.
Action WC send notes to people with open action items.
WC accessibility wizard in FrontPage. Dick?
DB Recently seen e-mail, will forward some info on the ER. Been bouncing around for a while.
Action DB: Forward info about the FrontPage Accessibility Wizard to ER. Get more info.
Action: everyone to comment before next meeting.
Action: WC add test suite to deliverables.
WC reads from charter.
HB naive to think that we'll describe all of the algorithms that will be used. What we describe are those that demonstrate what can be done not all the possibilities.
/* silence. people want to look at off-line */
WC Want to point to a concrete example of how we might do this.
CR Example language is helpful.
WC Don't see how we're going to do this.
WL We're the examples for their techniques. Example language could just be example tools.
WC What if we point to tools and say, here's what they say, or if we point to WCAG and say here's the rationale.
Resolution: Replace the "example language" sections in AERT with link to a tool (as an example) or to text with in WCAG (rationale).
CR we're setting up a relationship with AU, should we also set up one with WCAG? It seems very important for us to be involved in the work on WCAG 2.0
WC Has everyone looked at WCAG 2.0 Requirements doc? "2. Ensure that the minimal conformance requirements are clear " is most applicable. Joint meetings?
CR reluctant to increase meetings.
WC agreed. If we meet too much we won't get any work done.
Action WC: propose to CG that ER meets with AU, UA, and WCAG on a monthly rotation such that ER has one joint meeting per month and 3 solo meetings.
CR when we ask a user to confirm something, like is this a table for layout or data? Where is that info stored in the document or external?
WC Does there have to be one way or a few that are interoperable.
CR A few ways. But do we want to get into that?
WL what stage is RDF at? could we use it?
CR HTML could do that. perhaps "longdesc=""" means not required.
WL Same type of effort to make it a recommendation. That's a fundamental addendum.
CR It will bloat everyone's document. We would encounter resistance.
WL Legacy issue also a problem.
WC Suggest prototyping to see what feedback we get.
Action WC, CR, MC, BM think more about how to save state information during an evaluation session. Put on the agenda for the 17 July meeting.
WC I am not available, Len also has sent regrets.
WL We can celebrate emancipation day.
Resolved: Next week's meeting cancelled. Next meeting is 26 June.
$Date: 2000/11/08 08:17:43 $ Wendy Chisholmp