Chair: Jon Gunderson
Date: Wednesday, November 3rd
Time: 12:00 noon to 1:30 pm Eastern Standard Time
Call-in: W3C Tobin Bridge (+1) 617-252-7000
Chair: Jon Gunderson
Scribe: Ian Jacobs
IJ: Deliverance of conformance statement
Status : None done.
IJ: Review of last call by RealNetworks:
MQ: Found possible contact at NullSoft (WinAmp)
JG: Last call annotation in issues list
JG: Contacted various people: Hans (no news), Wilson Craig, Tim Lacy.
MN: Active element text:
MN: Proposed 1.5 wording.
Madeleine Rothberg: Still waiting for techniques.
Dick Brown: Proposal to split 10.1?
Status: For this call.
GR: Issues about spawned windows?
Status: Not done.
GR: 2.5 cascading order is subsumed by proposed restructuring and I have
techniques for it.
GR: Providing author info:
HB: Left message for Steve Anderson. No news.
JG: Sent email.
GR: I will look further into email for Foteos Macrides (Lynx).
1) 21 people have confirmed that they will review the document in last call. More people still pending.
IJ: Please note that these people will also be good sources for press support at Recommendation time. Keep the list handy!
JB: Linked from our home page.
JG: Anyone know anyone at Apple?
Action MN: Find a contact.
MQ: DialPad is a Web app that allows you to call on the Internet. It uses plug-ins and is Java/Web application.
JG: Include in techniques document in some form?
JG: Should this be for all user agents or in the section on dependent user agents? E.g., graphical UA conforms by presenting visual information, summary info, header info.
/* Several people felt it was a general requirement for user agents */
AG: Table content already covered by another checkpoint (2.1). What's missing: 2.1 doesn't cover the *relationship* between the cell and the header. The proximity must show through when rendered to the user. You could include "association" in support for HTML. I think the requirement in the area of this checkpoint could stand to be clarified.
JG: Ensure that the user has access to relationships between elements.
AG: Yes, then line by line explanation in the techniques document.
MQ: Tables are so important that it's worth saying it directly.
KB: Is there a way to point out that the proposed table summary checkpoint is a special case of access to content?
AG: Content includes metadata.
DB: Are we asking the UA to give info about a table on demand?
AG: Yes. (e.g., context menu).
IJ: Should this checkpoint stand on its own as a specialization?
GR: Yes, specialization of 2.1.
a) Capture goal of making table understandable.
b) Make applicable to user agents.
DB: Not sure if proposal is specific enough.
a) Should we be explicit in 2.1 or elsewhere that relationship information in general is important?
AG: To be clearer about what you need to do (orientation), you need to capitalize on relationships in HTML. And this is critical in the table area.
IJ: Should we be vague (allowing binary verification, however) and just rely on suggestions in techniques?
JG: Any objection to adding checkpoint as stated? It seems to hold some info not in 2.1.
DB: Are we requiring the user agent to provide information that has not been supplied by the author?
AG: The repair techniques are more costly to implement than to implement HTML.
IJ: I agree with Dick that the requirement should be for what is supplied by the author.
Resolved: Incorporate proposed table checkpoint (without reference to "selected" table).
Action Ian: Send revision to the list AND include in document (without necessarily rediscussing on the phone).
Action Al: Send HTML discussion to list.
Refer to proposal from CMN on relative priorities as done in AUGL.
JG: I think there are very few checkpoints in the current doc that would require it. Also, 6.2 (implement W3C specs) is priority 2.
DB: I think consistency between guidelines is a good idea.
IJ: I support this proposal for the checkpoint in question.
DB: I realize that there's baggage in trying to explain relative priorities.
GR: I see both sides of the issue, but I've had problems with the sloughing off of problems to WCAG. Priorities in WCAG may be disputed there as well.
JG: We can revisit this in last call. May reconsider at the end to see if the issue actually arises.
DP: We can also provide direction in 6.2 to the effect of relative priorities.
IJ: I can live with Priority 1 through last call, but I think Charles' logic holds in our case.
JG: WCAG depends on some features of UA. So UA complying at level one should allow those options to be available to them.
GR: WCAG is clearly transitional. There are many user agent clauses. The user agent guidelines is more forward looking. UA needs to encourage authors to use new features.
AG: It's proper for UAGL to assign something higher priority in user interface than what WCAG assigns it.
IJ: I propose accepting Pri 1 for last call and raising as an issue in last call.
Resolved: Leave as Priority 1 and indicate that it's an issue in last call.
GR: When you send out last calls, please include a reference to the issues list.
Action Ian: Include this in call for review.
JG: Issue of whether user agent should be required to provide information about author-supplied configuration to the user.
DB: For UA-supplied, P1. For author-supplied, P3.
JG: Proposed: P3 checkpoint to provide info just about author-specified bindings.
AG: It's important to know what will happen when you hit a key. It's not uninteresting to know where it came from. It's also interesting to know what's product default and what isn't.
DB: So does this mean that 10.1 (input config) will remain P1 and will include author-specified?
JG: Anything the UA can reasonably know about.
DB: I don't think that the author-supplied info is a P1.
IJ: For me, there's a minimum of making known that some behavior is associated with a binding.
JG: So there is no consensus on 10.1 Priority. Possible candidate for relative priorities.
JG: Does input configuration apply to other issues than bindings? Can we solve the issue by adding a specific checkpoint to address that particular concern?
GR: I still think that it's a two-part question w.r.t. accesskey: - The specified key binding is not the only piece: the UA has to decide how to support if. UA needs to tell user how to invoke it. I don't think we should have a separate checkpoint.
IJ: Where's the burden on the UA?
DB: Not a question of burden but of priority.
AG: Dick is saying it's not P1. Gregory is saying not P3. I would split it as follows: If you do it, you must document it. There's also a question about the priority of implementing.
KB: I can see some benefit to knowing what's author-specified as opposed to global.
GR: Please review my proposed techniques that I sent to the list (27 October).
1) Make 10.1 for UA-supplied configuration P1.
2) Add a checkpoint that's P2 for author-specified configuration.
3) Considering merging them. Should the priority of making author-specified configuration be as high as user-agent supplied?
a) Resolved: Go to last call 5 November - 1 December with open issues noted.
Two abstentions on decision to go to last call:
(The following correction based on http://lists.w3.org/Archives/Public/w3c-wai-ua/1999OctDec/0245.html)
GR: I'm concerned that we've had enough time to work out issues.
Add GR: I'm concerned that we haven't had enough time to work out all of
the issues after effecting such a drastic change to the structure and purview
of the document
JG: I'm only aware of two issues that are still open.
DB: I'm not satisfied that I've gotten enough input on this document from the IE group. That's a problem of the MS Access group.
IJ: Note that Rich said ok to last call with current Techniques Document.
IJ: Note that new level of document maturity (Candidate Recommendation) is likely to be adopted by the Advisory Committee during our last call.
Action Ian: Incorporate changes from this call.
Action Ian: Continue to add techniques.
b) Resolved: Structure of the 29 October Techniques Document ok.
Action Ian: Make meeting page available.