January 28, 1999
4:00 - 5:30 PM EST
Longfellow Bridge 1-617-252-1038
DD: would like to add Frames / NOFRAMES to open issues. We have to decide if we want a NOFRAME. Not enough agents supporting NOFRAMES.
GV: in lynx you get a list of frames and can jump around. Emacs too. FRAMES may be the same or more extreme than TABLES.
CMN: Thinks it is reverse of other situations. User agents don't handle the situation well.
DD: thinks that CSS2 will be the saviour of FRAMES.
IJ: maybe frames aren't all that bad, to factor out some constant information this is good for frames. CSS won't be able to do this and will require more coding per page than frames require. FRAMES aren't handled well.
CMN: most people design frame sets poorly from an accessibility viewpoint.
GV: since some tools allow you to use frames, maybe it should be priority 2.
CMN: you can build NOFRAMES
GV: Frames, if designed properly, can even aid accessibility by hiving off global navigational links into their own frame and the user knows where they are.
GV: Frames are somewhere between inconvenient and very inconvenient, and the current user agents are not handled well.
Consensus: A.9.1 make it a Priority 2 checkpoint.
NOFRAMES is necessary for good design, but not necessarily needed to get at the information. Consensus: ask UA group to mark proper handling of frames as a UA priority 1
IJ: problem in UA frames are something the user needs serial access to (like tables sometimes).
JW: are we going to make the same recommendation for I-FRAME (inline frames).
IJ: IFRAME is similar to OBJECT in its application.
CMN: if anyone gets around to implementing it, they will do it right
CMN, JW, DD, some discussion about whether user agents or page authors have ultimate responsibility.
GV: NOFRAMES is what we think people ought to do, but the information is not entirely inaccessible without.
CMN: may or may not take it back to the list. Make it closed at P2, and he may go back at some time.
WC: larger issue: other priority 1 items that are backward compatible. AT the time we started Netscape 2 were prevalent, but aren't now. What to do with the issues that may fall due to time.
JW: priority one if well designed UA's cant handle it, and priority 2 otherwise.
GV: have tried to round up interim fixes and group them together in one area. These should be monitored and reviewed frequently.
JW: and these should be P1's for User Agent guidelines.
WC: Would like to revisit content negotiation.
JW: the accessibility issue is that to make it easier for the user to get what they want when there are choices some content negotiation should be done. In allowing automatic retrieval of the correct file format, as in the helpful appendix to A14. (Wording as suggested by Wendy and Jason)
More general guidelines as to the use of character sets, media type request in the header would seem to fall out of the area of accessibility and into the area of server configuration. He would set the demarcation point for inclusion in the guideline.
DD: thinks we should not spend any more time on this since it is a webmaster issue. Also, by the time we work on the second edition of this, there will be more tools available.
Consensus: The wording (re: content negotiation) as suggested by Jason and Wendy ought to be a P3 checkpoint in A.14.
DD: He is collecting information from various experts and sources on the effective use of CSS. He is not finished bringing together the various information he is collecting on CSS.
WC: thinks that most of what DD is describing will be in the Techniques. We now have A6 supporting the use of style sheets, checkpoint 4 gets more specific on the use.
JW: there were some checkpoints that relate directly to CSS usage. Use of style in general should remain in guidelines, while specifics to go into techniques.
DD: will not argue with this concept.
Consensus: no modification needed to guideline doc. Will add new information to the technique document. Issue closed for Guidelines.
GV: hold off until further discussion with Jon Gunderson and John Gardner.
JW: doesn't think there is any difference in strategy between solution for Math and science, e.g. In long-term use a DTD that is appropriate to that mark up. And there will be an agent that will reproduce that in an accessible way.
GV: new checkpoint in A.6 ("Indicate structure with structural elements, and control presentation with presentation elements and style sheets.") "Where it's possible to present content by using mark up (for example music, mathematical and chemical equations) use markup language instead of using images. [Priority 2]
per Jason's message, we'll list a sample of W3C technologies in A.14.1 so that it reads: If W3C technologies are used (e.g., MathML, SMIL, HTML, XML, CSS), use the latest W3C specification whenever possible. [Priority 2]
insertion of MathML - Intro to Transform Gracefully - "Since HTML, XML, and other W3C technologies (such as MathML and SMIL) are designed with the flexibility to create documents that may be formatted in various ways on a variety of platforms, by virtue they support accessible design. Non-accessible pages are a result of giving up this flexibility. Creating pages that transform gracefully is not more costly, but requires a different design approach that also makes pages compatible with emerging mobile technologies. The following section A guidelines address the issue of creating pages that transform gracefully."
JW: dislikes repeating concepts throughout guidelines just to make sure that people are reading the document carefully. Music notation also fits in these definitions.
Consensus: take these four items as amended to the list as measures being taken to address this issue and ask for comments for wording from the list.
Who is going to establish what minimum accessibility is?
CL: the Canadian Federal Government has unilaterally decided that conformance means all P1 and P2. The WG can only suggest what it believes is conformance.
GV: US DOJ may decide for the US. Has some problem for us (i.e. the WG) deciding what conformance means.
JW: standards and anti-discrimination laws vary widely across national boundaries. Would like to propose what is the level of compliance with respect a particular law, and what constitutes compliance with the guidelines.
DD: what is the difference between having a statement of conformance and just leaving the interpretation of P1 and P2 which are quite clear?
IJ: has been much discussion about this in UA and Judy Brewer thinks that there must be some statement in the guidelines to be able to use them as a "stick" to encourage improvement.
DD: if we make different levels of compliance, it will not help because people will not be able to decide where compliance lies.
IJ: in discussions with Judy: it would be necessary for someone to point to the user agent guidelines and decide if their product is accessible.
CMN: maybe we should bounce it back to CG and Judy and Ian.
DD: next CG meeting is the 2nd of Feb and it will be raised again.
GV: suggests we take a hard look at all P2's and decide if they are all required (and P1 too)
JW: suggested the addition of "unless undue burden" to statement of conformance.
Action: To determine if conformance shall include only P1's or P1's and P2's, the group needs to take a critical look at all P1's and P2's to decide if they are all required
GV: which item is to be P1?
CMN: happy that this is no longer an open issue because of previous discussions about FRAMES
Consensus: issue closed.
GV: B3.1. could be a priority 1, and the only reason to avoid simple writing is that is if the context of the site or document precludes it.
DD: but a person with a cognitive disability would understand it in with longer, more laborious effort.
GV: disagrees. Cites studies that show that Persons with Cognitive Disabilities need information faster or they will loose the train.
WC: not bump up to 1, leave at 2, but include a statement like:
1. We should have a general statement in the introduction that says something about important formation/functionality, and 'presentational candy' - in the latter case it is acceptable to let the content/function disappear so long as it doesn't break the document.
2. The only general page authoring mechanism is to ask the author to provide redundant forms of information and hopefully the user agent will allow the user to filter the representations that is most effective for themselves.
All: no one can figure out what the above statements mean.
DD: agrees with the concept of leaving the priorities where they are and work out some better statements than 1 and 2 above.
GV: doesn't agree that this will solve the problem. If you stay at the current level there is nothing you can do to make pages more in-accessible to persons with cognitive disabilities.
Issue still open.
GV: hard to make it a priority 1, but hard to make it a priority 2.
JW: we need a stricter definition of content and audience.
GV: Shall we try a split priority and see what we can do with it. Then we will sleep on it for a week.