Minutes from 14 August 2000 telecon

available at: http://www.w3.org/WAI/ER/IG/2000/08/14-minutes.html

14 August 2000 ER WG Telecon

Summary of action items and resolutions
·       Action LK sendto the ER list a proposal for how to define and 
navigate instances of classes.

Participants
·       Len Kasday
·       Harvey Bingham
·       Chris Ridpath
·       Wendy Chisholm
·       William Loughborough
·       Brian Matheny
·       Dick Brown
·       Michael Cooper

Regrets
·       Gregory Rosmaita

Agenda
Agenda is available from the archives.

Action items from 7 August
·       Action WL Start the basis of "What are the accessibility features 
of RDF"
WL believes that RDF will help the device-independence movement move away 
from writing content for each device.
LK Two instances:
Knowledge in RDF, generate content from that
RDF to say things about pages
WL RDF will be a mechanism that will increase Web commerce even more.

·       Action WC: Check out schematron, how can we make it more precise?
WC In process. Not sure reading rules correctly. Play with more, send 
questions to Dan and Rick.

·       Action LK post links to ASSETS and CUU
·       Action LK post link to XML 2000 to the list.
done

·       Action WC ask JB who she knows in the area that might be able to host.
WC Lots of contacts, WAI domain discussion. PF also meeting.
LK With PF, good opportunity for us to discuss nitty gritty of standards.
WL PF is conduit to technical folks.

WAVE
WL Can WAVE detect non-w3c technologies.
LK General question, would be good to come up with an algorithm for this in 
general.
WL Could there be a WAVE logo?
LK There's no such thing as WAVE approval.
WL WAVE tested.
LK The approval would be W3C. WAVE tested is a marketing thing.

Javascript
LK This is a non-w3c technology. If something sent to a cgi, then don't 
know what's coming back.e.g., the IRS site, don't know if text or PDF is 
coming back. From a human point of view, once you've identified 
"document.write()" or ... rollovers can be decorative or cause a pop-up 
with information. Once you've alerted people to the javascript you can ask 
them to test the page with javascript turned off.
MC We're trying to provide automatic support for as many things as 
possible. That type of technique would not be appropriate.
LK Right, more of a WAVE thing. Trying to do everything automatic seems 
impossible. A-Prompt is more in the spirit of manual than automatic.
CR As automatic as we can make it.
WL Bobby, A-prompt, and WAVE have different approaches?
MC All similar right now, "ooh, you've got a script."
CR Sometimes hard to make a decision. e.g., hard to parse script to 
determine if it will flicker.
WL Does anyone parse javascript?
CR a-prompt no
MC bobby no, but want to
LK no. So, none of us do but we all want to.
MC We can pattern match, but we aren't parsing. therefore could find "open" 
but wouldn't know if it is window.open or a user defined open.
LK Can you just look at the code or do you have to simulate it?
MC Need an object model for the script. Don't need to simulate running it 
if we have the correct model.
LK I could imagine a program where "foo" is either a window or something 
else. But, you don't have strong typing. Let's say somebody likes to use 
foo as a dummy variable that they use and reuse. e.g., could be a number, 
characters, a window...
MC You have to evaluate the if statements, which leads you towards 
simulating the program. If we know we are catching 99% of things, have we 
caught it? If can't be 100% worth doing at all?
WL The amount of trouble decides that. Anything helps.
LK Answer is different depending on the type of tool. If have some user 
intervention and get a false positive, ok. However, if automatic then 
perhaps not.
MC Put in AERT let tool designers decide. However, need to say how many 
potential false positives exist.

Chunking
HB I extracted about 1/3 of their pages, w/out graphics. They are using 
horizontal tabs at top. They started adding new tabs for each new section 
of their site and recently simplified.
LK In terms of existing structures, list is nice. Headings are pieces of 
title. Nothing that says what it is a heading of.
WL You can have classes of headings and accessing what that class means. 
Discussion on WCAG about styles and class. should "sidebar" be an h3?
LK Additional conversation. Not obvious place in conversation. That is not 
a concept in HTML.
WL It becomes a visual class. Does jumping from a style to a style 
constitute structure?
LK We could propose that class should be made visible. I should be able to 
get names of classes and human readable names associated with classes. In 
WCAG a rule that you would use this. PF develop namespaces. In ER tools 
would hunt for places where classes ought to be used. We can start it here.
WL Is this a resolution?
LK Proposal: I have a document that defines a class "sb." I have some way 
to define that that stands for sidebar. A user agent comes along and reads 
what those classes are and allows me to navigate by instances of class.
MC How does it extend CSS?
LK it's like alt-text for a class.
Action LK sendto the ER list a proposal for how to define and navigate 
instances of classes.
MC Don't want to supersede XML.
LK If you look at the text of the guidelines (all) they define logical 
styles like example, checkpoint, etc.
MC Class is not part of the style sheet.
LK Where else can you use it?
HB It's up to the application.
LK Examples?
HB Use it in digital talking book.
LK So those user agents must do something with the subclasses.
HB It's XML. Allows us to say, "this H1 is a chapter." Reads class name.
LK Is there some way to have a string that is read aloud rather than class 
name itself?
HB They did not do that with talking book, no. It simply reads the class name.
LK There are some classes that I do not want public.
WC How specify?
LK Another attribute?

$Date: 2000/08/15 13:47:43 $ Wendy Chisholm

--
wendy a chisholm
world wide web consortium
web accessibility initiative
madison, wi usa
tel: +1 608 663 6346
/--