WCGL 22 April 1999


  1. Chuck Letourneau
  2. Gregg Vanderheiden
  3. Jason White
  4. Wendy Chisholm (scribe)
  5. Eric Hansen
  6. Charles McCathieNevile
  7. Judy Brewer
  8. Phill Jenkins
  9. Daniel Dardailler

Table of contents

  1. Summary of action items and resolutions
  2. Amount of changes
  3. Layout tables
  4. Linear versions of tables

Summary of action items and resolutions

  1. Resolution: use language as proposed for 5.3. leave 5.4 as is.
  2. Action: GV, PJ, DD will check with groups offline for advice - do thorough research to determine if the "until user agent" clause of 10.3 has been reached with the current availability of new technologies.

Amount of changes

GV - if make too many changes will have to go back to Proposed Recommendation. We need to figure out what we have, what changes need to be made, if they are substantial then we have to come back around. If we have something that could be made better but we can live with it, then we ought to pass it on. We need to address the comments that have come in from the AC (advisory committee).

PJ - Does AC respond before goes to Tim? May they change input based on clarification?

JB - AC as a whole - no, unless it is resent as another Proposed Recommendation. We have to avoid creating an unchartered working group. However, for any w3c organization that has commented during PR, the working group chairs and editors and staff can be in discussion about issues the ac member has raised. no new issues can be raised at this point. At this point, the working group is providing advice to the director in his decision about taking the document to Recommendation.

GV - a company may continue to clarify its comments.

JB - was not aware until last week the number of changes and discussions that the working group has been making. there are some kinds of changes that are easy to make at this stage, and others that are more difficult.

GV - we need to figure out what are the really important issues and which are not important. important usually require strong justification. unimportant usually easier. if we can't go forward without it and putting it in substantially changes the document, we'll have to take it out of PR. However, if we can clarify an issue rather than substantially change it, that would be preferable.

did everyone receive the open issues list?

EH - no, using phone line for the call.

Layout tables

[GV read Charles' - proposal 5]

JW - last one already in 5.4.

CL - checkpoint elsewhere?

CMN - it generalizes, and allows table markup to slip to p3. some of the problems with table markup are p1.

JW - agree w/substance of CMN proposal. reflects discussions. would like to see a few changes. the 1st item can't be p1 because of defn of p1. references to 3.2-3.4, should be under x.3. other than that, agree.

JB - will GV and WC share specifics of AC comments?

[GV reads proposal 6]

[GV reads other AC comment - once SS supported, prohibit use of tables for layout]

JW -

/* missed discussion while testing structural markup used within table. not a problem in ns3 or ns4 or MSIE4. */

JB - changes in priority in either direciton considered substantive (internal to W3C). WG could argue against. priority levels new to w3c.

PJ - then agree with JW.

JW - P2 inaccessible.

JB - way address by clarifying issue w/out changing priority?

GV - anyone think that 5.3 shouldn't be p2?

JB - 3/24 version?

GV - using tables for layout. it looks like won't rise to p1. there are things that will linearize.

EH - it shouldn't be there at all. we should not prohibit things that don't necessarily interfere with accessibility.

JB - I would like to make sure we're focusing on addressing issues in AC comments.

JW - also have to address discussion going on in working group.

GV - [reread AC comment - once positioning is fully supported, prohibit use of tables for layout.]

JW - accept that comment, include that condition and substance of charles' proposal.

GV - very seldom have i seen tables used for layout where no side-by-side text.

JW - still think 10.3 likely to be a problem. proposing to leave as is, change 5.3 and 5.4 to accept AC comment, and add clarification along Charles' proposal.

JB - do one at a time?

GV - AC kills 10.3, and qualifies 5.3.

PJ - not only prohibition, means that every page have to have a text version.

GV - yes. more tools coming out every day that will allow to unwrap.

Jason's suggestion interesting. "Once positioning and style sheets are fully supported. prohibit the use." says, allows tables today. no restrictions on use today.

PJ - that linearize, is important.

JW - 5.4 linearization and structural presentation issue. 5.3 address AC issue.

5.4 - include words, "tables are used for layout,..." include the 3 conditions, 1. don't use visual markup for presentation, 2. use proper markup to represent structure, 3. ensure when linearized is presented in logical order. p2

JB - what extent can the solution be a clarification w/out substantive change or are we out of range?

GV - for 5.3 - have clarified.

JB - yes, clarification by qualification.

PJ - yes agree.

CMN - is a major change. ensure that the content linearizes properly. not major change.

jb please restate

GV - once positioning in style sheets is fully supported, do not use tables for layout.

JB - then CMN says it leaves a danger zone.

CMN - no, a major change.

JW - /* proposal. missed, since looking up current table text */

WC - [in techniques, ...]

CMN - "5.3 do not use tables for layout unless contents make sense when linearized." the structural stuff is covered by guideline 3.

EH - yes. is closer.

GV - nicely states that must make sense.

JW - need to highlight issue of structure.

GV - when you linearize something does that mean that the cells all show up as one big paragraph or do they treat as paragraphs anyway.

JW - either way, the problem is still there. if meant to be a list, or heading followed by paragraphs, then the detail is lost. it is represented by table elements.

CMN - 5.3 do not use tables for layout unless contents make sense when linearized this includes use of proper markup - refer to guideline 3.

JW - "this include proper use of struct. markup in table cells, refer to 3.3, 3.5, 3.6 3/24/1999" need a note under 5.3.

JB - what do you mean? from my understanding that is not a solution.

GV - not discussing techniques. you will see "notes." so when ss positioning is supported, tables should not be used, see 3.3 in 3/24.

JB - everyone comfortable?

EH - don't like restricting use of tables. clarifying question?

/* further discussion about structure, linearization */

Resolution: use language as proposed for 5.3. leave 5.4 as is.

JB reflect advice on rewording, as well as why the solution is proposing not a substantive change.

/* group reiterating that we need to focus on AC issues */

Linear versions of tables

GV [read AC comments - in summary - remove it because there are solutions that will work today. then current wording]

PJ - IBM raised the issue. this means I have to create a text-only version of our entire site. when they make relative tables and style sheets will get wrapped text. it's always going to occur. we don't need to require webmasters to require non-table version of tables when we have tools to unwrap.

CMN - your assertion is that the "until" requirement has been met.

GV - not true for all user agents, but enough that it appears that the until clause is met.

have we met it?

JW - there are some that exist. i suspect that there are a lg # of people using traditional sr's who don't have linearization tools. linearization tools being developed by ER group not widely available yet. so "until user agents clause" still in effect.

GV - what if an unwrapping tool available?

PJ - what about lynx? freely available that works with screen reader.

JW - if not entirely up to author (i.e. a UA component) then never a p1? this checkpoint becomes redundant.

GV - if unwrappers for 2 major browser, what say?

JW - add opera and make 3. then until clause is met.

JB - not sure if would. if a convenience factor. if have to spring for new version.

GV - no, a plug-in that works with old ones. definitely reasons not to upgrade.

JB - screen readers very sensitive to upgrades.

GV - so a linearizing plug-in for older browsers.

JB - does clarification help solve this issues?

CMN - nope. we've either reached it or not.

JW - the Until user agents clause has to be defined in terms of plug-ins.

JB - the working group is in agreement of checkpoint, but not in agreement of ... /* daniel joins */

GV - 10.3 major amount of work for sites. we are close to satisfying the until user agent clause by weeks or months with plug-ins to linearize. therefore, it would be unfortunate to set this in stone and we've passed the milestone. it seems that we ought to consider deleting 10.3 based on the AC comment.

JW - check with the tools group?

GV - we have most of the players in that here now.

JB - there are a lot of variables and not an appropriate way to make a decision. we're basing it on another group's work.

GV - take question off-line. preliminary investigation found that achievable goal.

PJ - doesn't lynx meet the requirement?

JB - does it work behind any firewall? any office setup?

CMN - can use the dos version.

JB -are the setup requirements reasonable?

PJ - easier than a screen reader. <grin>

JB - but...

PJ - most people I know have lynx.

CMN - yes, most people I know who are blind can not survive without it.

JB - so we're making it more of a need?

PJ - just not reasonable.

CMN - only those foolish enough to use tables - a large proportion of the world.

JW - it's only the short term issue that we're concerned with. I only use browsers that support tables well.

JB - what i was starting to say before phone confusion - is there a different approach we can take to address this issue? the until user agents part is clear, but when is it met is not clear. we are extremely close, but in transition. therefore, some may feel it is clearly met, others may not. want a solution that does not go over the substantive threshold. perhaps, add to the note, "please check the status" and link to more guidance on until user agent part. OR, within the process context, if we can clearly demonstrate and that is not controversially agreed, that technology met, then in our advice to director, we could say that change be made. there needs to be enough agreement to avoid controversy. not clear that this is the case.

GV - think it is clear that some people believe threshold has been met, others that don't. are ie and ns the problem? and if have plug-in be solved?

CMN - on the basis that opera, lynx and w3 already do it. are the communities affected able to handle the problem?

JB - in japan i found that the latest versions of products that we think have access features, only have the access features in the english version. therefore, many latest versions don't support what we think they do.

PJ - [in response to gregg] yes, reduce controversy. believes already satisfied.

CMN - point is to make things accessible. alternative choice - leave checkpoint in. if feel is met. then claim that, and therefore don't need to change pages.

GV - [reads defn of "until user agent" clause]

PJ - yes is criteria, and a "check here for when we're there"

JB - process - we're at the end of 2 hours.

GV - do we meet the "UA clause" today? implies drop 10.3.

JW - for myself, yes. for other people, not sure.

GV - if plug-ins are available that linearize tables, do we have consensus that the until user agent clause is met?

JW - if widely available, yes. at moment, largely satisfied.

/* jason leaves */

GV asks round table:

do you believe it is met now?

do you believe it is met if other two covered?

CMN - not sure. abstains. won't block a consensus.

PJ - there today.

CL - won't block consensus.

WC - today.

JB - would like a complete statement why it has changed and lays out the issues of what people have access to and the different variations of why we think this is covered. that people have reasonably easy access. so, it doesn't matter to me if everyone says yes, as long as they can summarize why the technologies has changed and why we are at a place where "until" has been meant.

GV - other comments?

JB - have a high convincing threshold because difficulties people have with upgrading.

DD - abstain, but not enough information.

EH - tend towards yes, but relying on other's expertise.

GV - sounds like the three of you want more information and the write-up that JB describes.

JB - concerned about needing to go to the IG. if the group does a write-up and go to a user group for feedback to make sure we didn't miss something. then feel comfortable. don't want to miss something from the user end.

PJ - yes, that's why i put on liserservs

JB - check with groups offline for advice - do thorough research. does not pose a process problem.

Action: GV, PJ, DD will check with groups offline for advice - do thorough research to determine if the "until user agent" clause of 10.3 has been reached with the current availability of new technologies.