Chair: Jon Gunderson
Date: Thursday, March 30th
Time: 2:00 pm to 3:30 pm Eastern Standard Time, USA
Call-in: Longfellow Bridge (+1) (617) 252-1038
Chair: Jon Gunderson
Scribe: Ian Jacobs
Gregory J. Rosmaita
- Integrated some comments from Mark Novak.
- Wrote long descriptions.
IJ: I will publish a new one after the face-to-face.
1.IJ: Ensure that UA FAQ makes the EO WG agenda.
2.IJ: Reuse questions from ATAG FAQ.
3.IJ: Coordinate other longdescs for techniques documents with JG.
Status: Longdescs finished.
4.JG: Get MN and RS to review Appendix 5.
Status: MN Done.
5.DA: Review techniques for Guidelines 7 and 8
6.DB: Get Tim Lacy to review G+
Status: Not done.
7.DB: Review techniques for Guidelines 3, 4, and 11
Status: Not done.
8.DP: Review techniques for Guidelines 1 and 2
Status: I'm reviewing the whole document, will send comments.
9.GR: Talk to Paul Schrader at AFB (re: testimonials).
Status: Done. He will contribute one. And, as editor of "Access World" magazine, can accommodate an article on the Guidelines.
IJ: If you speak or write on behalf of W3C, please contact me (as part of the W3C Comm Team).
10.GR: Look into which checkpoints would benefit from audio examples in the
Status: Not done, but obviously HPR and JFW ones.
DP: Note that there is a multimedia presentation at the JFW Web site.
11.GR: Review techniques for Sections 3.7 and 3.8
Status: I sent comments on documentation, but others not done.
12.MQ: Review techniques for Guidelines 9 and 10
Status: Not done.
1. FTF meeting 10-11 April 2000 at RFBD in Princeton, NJ http://www.w3.org/WAI/UA/2000/04/ua-meeting-rfbd.html
IJ: Six registered today. Deadline for registration 5 April.
JG: Al Gilman may come. Rich may come.
DP: I still don't know.
MN: I still don't know.
DB: Don't know if Tim Lacy will attend yet.
IJ: Still no "no" reviews. One comment on post-REC tracking of user agents. I request that this be discussed at the face-to-face.
JG: This is now in the hands of the EO, who discussed this at the face-to-face.
CMN: I was at the EO meeting. They wanted more general questions, and came up with a bunch. Wendy and Chuck Letourneau took minutes.
JG: At CG meeting, Judy made it sound like EO would take care of the questions.
CMN: For ATAG, we sent some stuff to EO and they dealt with it. In the end, I think just Jutta and I verified that nothing was drastically wrong.
JG: I don't know whether we need two sets of FAQs (for media, for developers).
CMN: I think it would be a useful document to have for developers. Useful for new readers of the guidelines.
IJ: Yes, an executive summary, for example.
Action IJ: Draft a preliminary executive summary/mini-FAQ for developers. (No deadline.)
IJ: I still intend to talk to Tantek Çelik about IE on the MAC.
DB: I will continue to pursue my MAC contact.
JG: No information from Denis on this. I have an iMac now and could do a run through the Guidelines.
JG: Question raised by Phill Jenkins about whether a source view would satisfy checkpoint 2.1.
JG: My understanding is that the user be able to choose which content it wants rendered.
/* IJ reviews definitions of content and rendered content */
IJ: We've discussed 2.1 recently and decided not to reduce to through the
UI only. Instead, added a cross-link to Guideline.
CMN: Haven't we discussed before whether source was sufficient access?
DP: I think it isn't, since you're required to read the markup.
CMN: In ATAG, the tool is supposed to provide access to content through the UI.
IJ: I believe from the previous discussion on this that we did not want to make this a stronger requirement to make available all content for humans through the UI (at some point). In the 2 March teleconference (as I recall), we did not want to make such a change.
JG: Do our other checkpoints make the source view irrelevant if its not accessible? Do you satisfy 2.1 with one feature, or with a set of features?
IJ: Sounds like the "alternative equivalent page" case.
CMN: For HTML, you have to provide access to text content of elements, images, titles, etc. Those things collectively are what you must provide access to. You could have 39 powertools that work together, or offer a general solution.
GR: I don't think that 2.1 is not satisfied by a source view.
CMN: I agree.
JG: Then what does satisfy this?
IJ: I proposed that 2.1 be clarified to be "through the UI" and this was rejected.
JG: Is 2.1 a collection of a bunch of things?
IJ: I think that the source view and and question of 2.1 through the user interface are orthogonal.
- I think 2.1 is the general case. (There are other specific checkpoints for access to particular types of content).
- A source view and the rest of the UI are different kinds of beasts, and we need to make that clear in the document.
Process suggestion: We flag this as an important issue on the mailing list and we attempt to find there further input to have a clear resolution. Say explicitly that resolution will come from list input.
DP: What would we leave out by reducing scope of 2.1 to user interface only?
MN: Difficult for me to draw the line.
KB: I don't understand the trade-off. Does all content meant for humans mean through all modalities?
MQ: The phrase "meant for humans" is not explicit enough.
CMN: I think 2.1 should be limited to human-intended content through the UI. But the implication is: whether you require a source view. I can use a source view for information, but not all users can read markup. Having a source view is valuable, but using the source view as the way that you are expected to get information intended for people is not sufficient.
DB: I think that there's a possibility (notably in the future) for markup meant for devices so that when rendered, provides content.
IJ: That's what scripts do today (e.g., generate a list of numbers).
CMN: Non-renderable information is the sort of thing that a source view is reasonable to use to provide access (for skilled users). Stuff that's intended to be rendered should be rendered for humans.
DP: I like having a source view available as a coder.
MQ: I think that we're saying is that the source view is not sufficient for average users.
CMN: It's not appropriate what is "renderable content". It's appropriate (and may be a requirement) for making available machine-readable content.
Proposed: A note that says "A source view is not sufficient for providing access to content that is meant to be rendered for humans."
DB: I have problems with this since there may content intended for people but not for all user agents. Any kind of content can be considered renderable.
JG: We're not asking all browsers to support speech.
CMN: Longdesc is a good example of this. In the source, longdesc is a URI (it's meant for machines). There's not requirement to expose the URI, but there is a requirement to make available the content at the other end.
JG: I'm concerned that if 2.1 stays as it is, a UA would not have to make available an equivalent alternative.
CMN: I propose to add to 2.1 a note:
- For "renderable content", access must be provided through the user interface.
- All renderable content (subject to the applicability clause) must be available to the user.
- For content intended for machines, access to the user through a source view is sufficient in some circumstances.
(CMN: The implication is that the source view is different from a "canvas".)
- Content that is available its "final form" (i.e., ready to view).
IJ: Need to define renderable content and make sure that it's clear that information doesn't have to be available all at once.
CMN: That's covered by 2.5
DB: What about multimedia content that the UA can't render?
CMN: Covered by applicability.
IJ: How does a UA dealing with generic XML distinguish content for humans and content for machines?
DP: I think this is still covered by applicability.
CMN: I think the answer to IJ's question is work going on in schemas.
KB: Are there any user agents that meet this checkpoint?
IJ: Only if a source view counts: Amaya.
CMN: PWWebSpeak handles longdescs.
IJ: What about scripts? The script is meant for machines, but the content is meant for humans.
CMN: The UA doesn't know that. It does know that the script is meant for a machine.
IJ: While that's true, wouldn't you (in spirit) want access to that content?
CMN: Yes, but this is an authoring issue: use NOSCRIPT.
IJ: CMN, are you suggesting that a source view is a requirement for all user agents?
DB: I don't think that it's a requirement. You can't get a source view for Word.
/* Ian gets the door, Jon summarizes the situation */
CMN and IJ disagree on whether 2.1 say "all content must be available through the user interface".
Action CMN: Send a proposal to the list.
Adjourned 15:34 ET