6 December 1999 ER-IG telecon



Report Tool

WC I sent a proposal, Gregory mocked it up (sent to me privately), Daniel commente don my proposal, Gregory will be modifying mock-up based on Daniel's comemnts. I still need to rework how to walk a user through all of the WCAG checkpoints.

DD any one working on guess e-mail function? to look for author's e-mail on the page. in meta, or elsewhere on the site. I can try to find someone to do it.

LK Emphasizes comparison of visual presentation from alternative content. e.g. it puts alt-text next to the image. I've been writing this in Perl. Bobby is still stressing the auto-checks and the others are verbal at the end.

MC yes. we've discovered that people who use Bobby do the auto checks but not the manual ones. therefore we want to make the auto ones as best we can. we don't have the resources to do the interactivity required.

LK what language? does it do its own parsing?

MC Java. Does its own parsing. NOt sure what 4.0 is going to do. Looking for a good parser that support DOM and correction for well-formedness HTML.

LK might want to look at Tidy. An option is not only to correct syntax but change to XHTML.

MC on the back burner. we might incorporate Tidy as a module or steal some ideas.

LK using Perl for sake of doing something fast.

Table linearizer

LK you select and then it gives you the whole table. is there a way that when you get to a certain cell you can switch the options w/out losing your place in the table?

DD we discussed that a while ago. Thought about a link next to the cell that would let you change direction. It was discussed on the list. It is a feature I would like to implement. However, before add more features want to get it online. We are dealing with security problems.

ERT - tables

LK what's the next issue?

CR before next issue, could we review table discussion? guideline 5 techniques.

WC do we need the "suggested language" sections?

CR we have it so that tools are in synch with each other and we don't confuse the user.

BM it also shows more of the feature

WC someone may come up with better language and it doesn't show all aspects of what your tools are doing. perhaps say "example language."

CR remove entirely or change to example?

MC say "example."

WC it's like the separation between guidelines and techniques in WCAG. it's the implementation specific information.

CR what about an "example implementation section." where we link into Bobby and A-prompt?

LK in terms of time to get it done. quickest thing is to remove suggested language and put in next version of document.

CR not take too much time. MC how soon will you have something i can link to?

MC In January we'll have new version.

CR for now, change to "example language" in new year set up links.

WC need clearer algorithm.

CR reluctant to use pseudo language.

WC what about combining them in order, since some checks in 5.4 should be implemented before 5.1

CR & MC not in this document. need to be checkpoint specific to determine conformance.

MC we could suggest combined algorithm in Guidelines 5, then break down under each checkpoint.

WC developers will combine when go through

MC people may not. they may keep them separate.

CR didn't want to go for pseudo code so we could discuss.

BM plenty of cases where it is appropriate. e.g. 3 and 4.

WL nothing wrong with.

LK different types of programming, like AI where looking at things in parallel. sometimes when order is important, other times when it is not. therefore don't want to put order on something where not needed.

CR i'll try using it more, so that it's more precise. if not clear or not in order, let me know.

LK it seems to be saying that if you have 2 or more level of headers, need to use "headers." but first, that doesn't seem to be necessary. some tables it's straightforward algorithmically for figuring out what headers go with which cells. as far as guidelines go, it says, "scope and headers to explain more complex relationships." therefore, you would want to look at HTML 4 algorithm, only give the warning when have cases that don't fit the algorithm.

@@LK send HTML 4 algorithm to ER list.

WC should check that IDs are valid

LK would HTML validator pick that up?

@@WC find out if HTML validator pick that up.

LK it should be under "validate HTML."

MC is a validator check, but important for accessibility.

/* BM leaves */


LK next is on the 20th then the 3rd. go ahead with?

group agrees to meet as scheduled.

XML 99

LK presenting on ER with Harvey and Jutta.

WC slides? what talking about?

LK it's a day. Judy talk about web access in general. a whole over view of web access, representation from every group.

WC focus on XML for repair tools?

LK using as opportunity to present to audience of web people. hasn't been that much talk about it (XML).

WC have audience do some brainstorming on both solutions and problems.

WL therefore, XML means that someone gets to create their own markup language.

LK individual proprietary or industry.

WL so they can create elements, they may not be related to HTML markup elements. one of the things is an education process to create an awareness that when create an element does it impact accessibility.

LK it might be worse than that. are there facilities in XML to get around these problems. in XML define a "header" then define style that renders a certain way. a mechanism (aside from RDF).

DD XSLT map XML elements to whatever you want. could indicate way to translate XML to HTML. data typing, etc. there are a few mechanisms in place. you can also use HTML module - you don't have to reinvent everything. in PF we started an XML authoring document. I presented it and it is online. http://www.w3.org/WAI/PF/xmlgl.htm

WL when you say it could allow the possibility of must or require doesn't exist. you said it wold allow you to furnish semantics but there isn't anyway to require.

DD you can design your own language and be done.

WL if you wanted to deliberately make it inaccessible you could do that?

DD right. thta's why wee need guidelines. you could define photo with not alt or description attribute.