WL we're not evaluating the WCAG, we're evaluating Web Content.
HB but we're adding content.
WL it repairs the content.
LK "Techniques for evaluation and repair of Web content"
CR do we want to refer to WCAG?
WC AU does not. What about "Evaluation and Repair tool accessibility guidelines"
@@WC summarize discussion of title f the document on the list.
LK new news?
WC before WWW9 is not out, but more likely to be in Sophia than Amsterdam or we could consider some other place in April.
WL April is out for me.
LK cost from Amsterdam to Sophia?
WC don't know.
@@WC propose Sophia before WWW9 or Philadelphia in April.
LK easy to get a room at temple.
WC resolved in WCAG WG that defn. in Glossary is the proper definition,
CR row by row?
HB except if reordered for language (read right to left)
CR but row by row is still appropriate
LK if you have HTML source, one row table data name then table data address. the order in the source is name address. that's how it is in the source. if that is put into a browser, it will read address and then name.
WC still go by row, but column order needs to be switched.
@@WC do today.
CR according to HTML spec, STYLE only in HEAD. However, browser support STYLE in BODY, but differently. in IE5 it takes the last style and applies it to the whole document. Netscape applies the style until it comes to a new one then it applies that one. I only looked at STYLE element not the "style" attribute.
WL "style" attribute used to override at any element level.it does not cascade to other elements only specific to the one it is attached to.
CR for guidelines six, if have STYLE element or LINK then need to ensure that document usable w/out the style sheet.
LK wherever the STYLE element is we need to put up a notice.
WC but STYLE in BODY will not validate.
CR regardless, no matter where it appears we need to let them know it needs to transform.
WC but we need to validate to a published grammar to conform to WCAG.
/* discussion about techniques for STYLE appearing multiple places */
WC we may want to structure by HTML element like in the WCAG techniques document HTML element index.
MC yes, we have had to think like this in developing Bobby.
WC could also modularize as in WCAG techniques.
MC there will also be several elements that may have the same technique.
LK do not agree that we should go by element. focus should be on checking against the guidelines. if i were implementing this, i would have an alt-text check then a list of what elements this applies to. I like it the way it is.
CR what about both?
WC yes, we could have an index like the WCAG techniques.
LK concerned about effort to create want to focus on issues.
WC we use scripts in WCAG to generate documents. worth the effort.
@@Resolved: we want to have an index of HTML elements and attributes similar to the WCAG techniques HTML index.
@@WC take discussion to the list re: using scripts to generate document.
done.
in progress.
WC waiting for tool. /* describes how ETA works */ it doesn't automatically do things, it just tracks what we do.
LK can we sort by technique?
WC we could create a category for each technique or category.
LK what about a search?
@@WC can we do a search on ETA issues?
@@ WC and CR get in touch via e-mail
LK could MC highlight major issues from list?
/* MC's issues */
MC 1.1.B. currently ask for every image for longdesc and d-link. instead just ask for longdesc and then d-link.
WL value of bobby is a little mitigated by the repetitive asking.
LK from the user interface point of you could ask once "did you use a longdesc" rather than for each image?
MC yes, that's how it gets done, but it could be a long list of images. Instead we could have question w/out listing the images.
LK we go through trouble, then longdesc not shown.
MC IBM home page read shows it.
LK how about saying, the tool should have the ability to turn off any of these checks.
MC perhaps in preamble. provide a grouping of those that are semi-automatic that can turn off, vs. those that are automatic.
WL yes. one of the fundamental thing is can you turn off the alarm.
WC but that's user interface. we need to point people to UA and AU.
@@WC add links to UA and AU in regards to dealing with user interface (such as turning off checks).
MC don't do longdesc and d-link as separate points.
Resolved: treat longdesc and d-link as the same point, don't separate.
WL 11.4
MC if page hasn't passed automatic P1 and P2, then raise 11.4.
WL yes and when it says "text-only."
MC but then they have satisfied the checkpoint.
WL how do you tell it is updated? it is suspect.
MC with Bobby, if they pay attention to the issue once we assume they have done it all the time.
LK you could have some automated checks to compare text-only pg w/the standard page. e.g, check that all the links in standard appear in text-only.
WL my assumption is the opposite. it's a sign of cop-out.
MC the links may be pointing to other inaccessible pages.
LK I would find a tool like that handy because when i evaluate a site, I could say "this one is not equivalent anyway, so since you have work to do, go ahead and make the original accessible."
WL it's a red flag.
LK yes, but we're permitting it. it's easy to think of a highly interactive page where the best thing is to have a separate page.
MC yes, think of the macromedia site. their site is in flash. they can't not use flash since they developed it. the inaccessible version has a good reason to be there.
LK there is an educational site about bats, with experiments and illustrations that use Flash in a worthwhile manner. it would be easiest to create a text-only page.
WL isn't that like londesc.
MC down the road you would do that with nested objects.
LK that makes me change my tune. instead of completely separate page have nested objects?
WC yes, like MC said. however support of OBJECT is iffy.
LK forbid alternative pages all together, and only replace a portion of the page.
HB wanting a text-only rendition and getting jumps off to the side.
MC you lose context.
LK withdraws suggestion. I would like to see the tool check if the text-only page is truly equivalent as much as possible.
MC repair tool could take complex table and linearize as text-only pg.
WC there is TOM, but i believe it is outdated.
LK with 11.4 we add to the techniques that the tool shall help the author create the accessible page, shall check as much as possible if the page is equivalent.
CR aren't they going to determine if the page is accessible or not?
LK go through automatic and manual checks (interactive) if fail those, then pops up a suggestion, "here's how we can help make this page accessible. and suggests how to create a text-only page."
CR if you use Flash to make that accessible, can't you include a description of what happening in the graphics right on the page? do you have to create a separate page? it's up to the author.
MC CAST would recommend that. not sure if you can check it.
LK with Java, you may have an applet that brings up another page. therefore the text below does not change. In the future, with OBJECT would be similar.
WC both things are possible, therefore need to confer w/author.
CR if put in best effort won't need text-only page.
MC it's been there for a long time.
WC no, it's not outdated. look at the flash example.
CR can't you include links on each page? or the text?
WC try selling that to a Flash designer.
CR getting back to MC"s comment - only trigger on certain pages, should we do this for all?
LK doesn't really say.
CR then trigger on all.
MC but none of the priority levels is fully automatic.
WC
LK there be an alert only if haven't passed P1.
WL or don't pass the level they claim.
WC yes, i like that. start out with your aim, AA. If you don't pass all of those (automatically or manually) then you get the "here's another option - a text-only page."
HB let the user decide what level they want to check to.
CR part of the user interface.
WC agree.
LK defining when warnings pop up, is not user interface that's functionality.
CR in a-prompt you can select which ones you will check for.
@@WC summarize and take to the list.
$Date: 2000/11/08 08:17:43 $ Wendy Chisholm