WCAG WG Weekly Telecon

5 May 2005


See also: IRC log


Andi, Becky_Gibson, Bengt_Farre, Dave_MacDonald, Gregg_and_Ben, JasonWhite, Joe_Clark, John_Slatin, Loretta_Guarino_Reid, LucaMascaro, Makoto, Matt, Michael, Michael_Cooper, Microsoft, Mike_Barta, Roberto_Castaldo, Sebastiano, Tim_Boland, Yvette_Hoitink
Roberto_Ellero, Doyle, Neil_Soiffer, Roberto_Scano, Takayuki_Watanabe
Gregg_Vanderheiden, John_Slatin
David_MacDonald, Yvette, ben




<ben> minutes)

<jslatin> hi andi

<Andi> Hi John

<Yvette> Hi all

<rcastaldo> HI folks :)

<David> test

<rcastaldo> I'm trying to connect with dialpad... having some problems

<bengt> they update the software just try to exit and start again

<LucaMascaro> tks

<Yvette> np

<rcastaldo> Ciao italians :-)

<ben> scribe: David_MacDonald

<Sebastiano> Ciao Roberto!

Agenda review (including information about face to face) (John, 5

<David> js: techniques taskforce

<David> js: tell zakim to ping us a 20minute tells us times up, a good reminder

Techniques report (Michael, 5 minutes)

<David> mc: shadowing issue summaries along with guidliines issues

<David> mc: 4.2 we go less than half way through lots of stuff

<David> mc: some case issues into the guidelines. would be helpful, if the same person does the guidelines review as techniques reviews.

<David> mc: lots of discussion etc... on of the things we missed was paying close attention to the baseline. take the three baselines. basic graphic browser, user agent. identified for each techniuqe within each baseline whether it is sufficient or opptional, not harmful,

<David> mc: first this look....a proposal to remove a techniques might because on one baseline and not another.

Guideline 2.4 proposal (Yvette, 25 minutes)

<ben> yvette's proposal: http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0371.html

<David> js: move yh 2.4

<ben> annotated proposal (wiki): http://esw.w3.org/topic/May_2005_Guideline_1%2e3_Proposal?action=show

<David> yh: I don't see the message tonight about SC L1 about reading order. has simple examples of content that is not accessible.

<David> jw: the problems I raised still stand, that will have to be treat as 1.3 issue or worded different. If you reword it it doesn't belong on level 1

<David> jw: see mailing list for more

<David> yh: some people think it is about aggregation, it is better to get consencous on problem then get solution

<David> js: it boils down to what happens when a screenreader goes into say all mode...the problem is that I want that reading to make sense, so I don't want a paragraph of text to be interupted by avigation bar orother artifact that a sighted person wouldn't look at in the middle of their content readitng, on a page not constructed properly, that doesn'thappen for a blind person

<David> jw: everything I said in message is still true

<David> joe: topic dogs us unecessarily over the years... if waiting for freedom scientific to do stuff properly, not our problem

<David> joe: doesn't have to be in exact same order.

<gregg> ask jason

<David> joe: doument html read order

<gregg> ask joe

<David> joe: that is nvalid html and we tell people to use valid html.

<David> joe: not a real feal...if a lot of valid html, and valid css, lets put some examples together

<joeclark> Also, a case of content interposing itself at the wrong point in the read order is probably a case of a bad user agent or incorrect HTML in the first place. Valid, semantic HTML has an intrinsically comprehensible read order.

<David> gv: I'm still wrestling through first one.

<David> gv: if reading order not im0ortant we could skip over this guideline

<David> gv: we have to be very careful about level ones. and with these kind of comments about it we should reconsider

<David> js: gregg mentioned, and I say there are other technologies ie math ml, flash, pdf, order is extremly important. willing to grant that jaws etc have inadequate features regarding this

<David> js: i have heard many examples where menaingful reading order of text was a clear messed up order that no sighted reader would have

<David> js: not the intended behaviour behaviour by the author, I believe that it was not a validation problem, there are real issues. how to disentagle them...is a porblem

<David> yh: sc at level one, major overloap with 3.2 link text... this should be at level 2,

<David> yh: this could be a problem for content management systems.

<David> jw: back to problem of John, the cause is not solved by writing markup by writing to spec,

<David> jw: the problem is covered under 1.3

<David> jw: but introducing a linear order requirement should be level 3 in 2.4 at best....not agood way to go . oet's move examoles to 1.3

<David> gv: 2 things, 1) right now sc 2 is too general, if we say the content or parts of content are arranged in a linear sequece to understand their secquense proerly then them should be probgramatically determined

<David> gv: the comment was made thae sc 3 could not be done for some technology so it should go to L 2 , but you can't have something at level 2 that can't be done

<David> bc: jason raised good point when he talks about an artifact of presentation, is that something that we really want. it the presentation is ajusted to present things in an audio form.

<David> asw: if we take 3.2 wording it will be ok at level 2,

<David> gv: I'm sorry I have a family emergency with my orphaned nephewsand I have to go, can someone jump in.

<Yvette> scribe: Yvette

<ben> ACTION: john, yvette, joe, michael - revisit 2.4 wording and repropose [recorded in http://www.w3.org/2005/05/05-wai-wcag-minutes.html#action01]

<Tim> Testing linearity seems easier than testing "making sense"?

jc: rewritten 1.3 to explain about web standards (valid CSS, HTML and generic JS)
... we even require valid code in WCAG 2
... HTML will be majority of web content and will have structure, PDF might have structure but we will require structure
... This is catching everyone up to standards and telling them how to do it
... boils down to "write according to web standards"

js: what's the def?

jc: in all web pages you have 3 layers: structure, presentation, behavior
... structure - HTML/tags. Presentation - CSS, behavior - JS
... other technologies might not have all three layers but there still might be presentation and behavior
... 'information' is redundant and circular. Information is purpose of website. We are Web CONTENT accessibility guidelines so no need to name information explicitely

js: proposal before us. Some discussion on the list. Comments?

gv: I've only found a proposal for the guideline text, not for success criteria

js: there was one but forgot to include in the agenda

<ben> http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0248.html

js: cudos to ben for being the archive wizard

<Zakim> Yvette, you wanted to say "keep overlap with 1.3"

tb: would all technologies follow that model?

jc: one of the few experts on MathML is intrinsically structural because it's markup language
... HTML requires MathML to mark up equations
... ambiguity is part of math, don't think it's really a accessibility problem

<Zakim> Michael_Cooper, you wanted to say MathML is very baseline-dependent

mc: MathML has some UA support but won't be part of every baseline.
... it's baseline question

jc: My issue with that is that you suggest there is an alternative but there isn't. For real mathematical equations there isn't a real alternative

mm: but there is (la)tex which has been around since the digital ice age

gv: Joe, you took 'relationships' out of SC 1. Did you mean that relationships are covered by structure?

jc: I was one of the people who wondered about purpose of 'relationship'
... that question was never really answered so I took it out

js: so you believe that all the relationships are covered by structure?

jc: Yes, all the structures are already in HTML

gv: You keep referring to HTML but we must assume we don't know we're using
... If they want to use something else we have to wonder what they need to do in order to conform

<joeclark> The post I forwarded from Jacques Distler on MathML's "alternatives":

<joeclark> http://lists.w3.org/Archives/Public/w3c-wai-gl/2005JanMar/0066.html

gv: saying all relationships are covered by HTML structure doesn't address that because that might not be true for other technologies
... does anyone disagree that relationships are a subset of structure?

asw: relationship between form element and label, would that be structure?
... web pages are not just documents

jc: I would be happy to change that into whatever the group wants

asw: what about 'content' that was in the original?

jc: fine by me

gv: add note in guidedoc that we consider relationships as part of the structure
... Let's define relationships to be included in structure and change document back to content

js: modified proposal for SC 1 "structures within the content can be programatically determined"
... anyone against adopting that new formulation?

different one

jw: someone should write a definition of structure and make sure it includes all those relationships

jc: if someone can come up with a definition or link to definition, I can pass it on to standardistas and ask their opinions

<ben> http://www.w3.org/WAI/GL/WCAG20/#structuredef

jw: I'll take a look at the definitino
... current def is fine, so am happy with the proposal

js: any objections to unanimous consent?

<ben> resolved: accept Joe's proposal to revise 1.3 L1 SC1 to read: Structures within the content can be programmatically determined.

<rcastaldo> good

js: next SC

asw: did we have consensus on guideline?

js: no

asw: Had problem understand what was covered in SC 2 that wasn't in 1

<ben> proposal for SC 2: 2. Structural markup or coding is used to encode semantics to the extent possible for the content.

jc: 'the extent possible' is human testable just like correctness of alt-text
... common misunderstanding that standards compliance isn't semantic (for example: page with just <div> and <span>)

correction: common misunderstanding that standards compliance = semantic

jc: semantic is step beyond

asw: still don't think that's testable
... don't understand why we need it. If structures can be determined, that's enough
... it might not be understood as well for other technologies

jc: it's possible in tagged PDF

<LucaMascaro> use just <div> and <span> is a violation of the validity of the code therefore the DTD, because we not use the correctly elements

yh: forgot what I was going to say

jw: serious problems with this one
... it's redundant with SC1 1.3 and requirement of writing according to spec and 'to the extent possible' is not testable
... it is not testable in different technologies.
... we need SC about that it has to be relative to the technologies that author is user

gv: just to clarify: we have no requirement to write good alt text because that's not testable
... we would like to go further but don't have any objective way. We just specify what we can in a testable way. Especially in level 1 and 2 we're very careful about that because people might be required to conform
... is there a word missing in SC? "structural semantics"?

jc: Mean semantics as normal in web standards field, using right element to mark up content
... I want two things to happen:
... Not allowing people to use just <div> and <span> to mark up page
... I don't want standards-compliant people to be harassed if they use <b> and <i>
... web standards should be the standards


js: what I would like is that we get another subgroup to work on 1.3
... take into account Joe's proposal, responses and 4.2 discussion

Can someone take over scribing after this agenda item?

js: Joe, could you focus on 1.3 more than on 2.4

jc: sure

js: Joe, Gregg and Becky to work on 1.3

<ben> ACTION: joe, gregg, becky to work on revised proposal for 1.3 [recorded in http://www.w3.org/2005/05/05-wai-wcag-minutes.html#action02]

js: stay in touch with 2.4 to address the overlap

<ben> scribe: ben

Guideline 4.2 revised proposal (Loretta, 25 minutes)


lgr: looked at gregg's proposal to replace the word "baseline" - am liking the word less and less
... UA assumptions is not the same as baseline

gv: what I meant was assumptions you're making about user agents

lgr: agree with sentiment, but am soliciting alternative phrases

gv: you said, you wanted to focus on techs, you're talking about UAs that use technologies?

lgr: the technologies for which user agents exist

js: think there was consensus that we should talk about technologies rather than user agents

gv: yes, but 2 types of techs. techs user has or techs authors use

lgr: reads proposed def (assumes are supported)

gv: supported and turned on?

js: and enabled?

lgr: avail. to user?

gv: if turned off, then it's avail

lgr: looking for a word that implies that tech is both "there" and "on"

js: supported and active? or you could define supported as meaning tha tUA is actually using and processing it

second point is that it wasn't clear from revised proposal for definition that this was the min. that can be assumed

js: not talking about larger set, but minimum assumptions - think that needs to be put in explicitly

lgr: it is there, "minimum set of technologies"

js: any preference about alternatives jason mentioned?

gv: in this conversation, we've drifted between UA and users
... not sure which we want to use in definition
... any feature that can be turned on and off would have to be assumed to occasionally be turned on and off for whatever reason - would have to be not have to be turned off for accessibility reasons

jw: supported in UA and active in those UA is the sense that we want - that's the assumption from the author's perspective

asw: seems we're nitpicking - if you assume it is supported, you have to assume it's turned on, otherwise, what's the point?

gv: may need to be turned off by users with disabilities

mb: seems to be a distinction between supported and available - either say it's reasonable to assume something is available --or-- we know the user has it and is using it
... any consensus on which we mean or do we mean both?

js: my sense was that wording jason mentioned might be nitpicky, but covers both of those possibilities

mb: concern is that means that it is responsibility of person making claim to know whether user has something turned on and off, which author can't know
... that's very restrictive

asw: back to HTML only sites

mb: if you're saying that you're picking techs based on a reasonable belief that techs are available

lgr: remember, this is about trying to capture for WCAG the fact that someone will consider and make an informed decision about what is reasonable to assume - a lot of what we're churning on here is how someone makes a decision

jw: loretta's point captures it well, basically defining min. set of techs that author assumes are active in UA. reasonableness of that assumption is not part of the consideration here.

gv: one of the things we keep saying is "that can be assumed" question is by who? authors? companies? I think we should say a min. set of techs that are assumed to be supported. by who depends upon who is setting the baseline
... or an established set of techs that the author assumes can be supported
... so it would be "a standard set of technologies that the author can assume are supported by user agents"

js: don't think we can say "established" because there will be situations where nothing has been established
... current discussions in the TTF, list, etc. have been talking about 3 baselines, so word minimum may be problematic

gv: no, all 3 assume a different minimum set

<rcastaldo> I've got to leave the call now

<rcastaldo> By everyone :-)

gv: dropping the word "minimum" might be a good idea

lgr: think word "baseline" has wrong connotation. think "minimum" is critical

jw: was going to make the same point, emphasizing the importance of "minimum" in this context - agree with gregg's earlier assumption to say "are assumed"

js: propose we adopt this because it's better than what we've got now

gv: if we did, we might say "supported/on" as a footnote to remind us to fix it

<scribe> ACTION: loretta to wordsmith definition of baseline and post to list [recorded in http://www.w3.org/2005/05/05-wai-wcag-minutes.html#action03]

resolved: accept "

The minimum set of technologies that are assumed to be supported/enabled by user agents in order to access all information and functionality of the Web content.) as a working definition for baseline.

definition of technology

accept definition as proposed ("Technology" means a data format, programming or markup language, protocol or API.)

lgr: added a couple of proposals
... new SC propsed: [9] New GL 2.4, Level 1 SC


Changes to content, structure, selection, focus, attributes, values, state,

and relationships within the content can be programmatically determined.


Issue: how much of this is just a user agent requirement? When does

the author have responsibility for any of this?

js: reactions?

yh: agree with loretta's doubts that if it's not possible to determine programmatically if something changes, then no UA will be able to show it. if a change to content isn't determinable, then no browser could show it. I really don't see what problem this is going to solve

gv: if each one of those things can be accessed, then UA could compare "then and now" and programmatically evaluate whether a change has been made
... usually this kind of a clause is surrounded by a requirement for notification, but I don't think we want to require that AT be notified at level 1 (we'd have to think about how that would be done)

lgr: there is at least one case (ex. flash) where you have in fact tell things to update themselves - not sure if this is just a UA issue

gv: sounds like a UA problem

jw: agree that this is a UA problem, UA must be able to detect a change, otherwise it can't change it - it's an internal issue w/in a UA. Only issue I can think of is if things are being updated in sequence, you might want a transaction or commit after a change so it doesn't become apparent to UA until all changes have been made - that's where you might want something in content to flag that a series of changes have been completed

gv: having spoken against it, I was going to comment for it (slightly) - one of the things that author does do is have a choice for choosing a tech where this is possible. (ex. captions are synchronized) - they could choose a tech. that doesn't support synchronization. author is reponsible for using a tech for which there is a UA that can do something.

lgr: sounds like this should be withdrawn

asw: what about applets? where do they fit in?

gv: that's our big 4.2 issue

lgr: would keeping this help us address applets?

gv: I wouldn't drop it at this point because we haven't figure out all of 4.2 yet

when we have content that delivers the UA along with the content, we do have to figure out where that gets distributed

js: that is what the discussion is about and the proposal is an attempt to distribute 4.2 issues so that they are addressed in other guidelines
... think we should give loretta as much useful feedback as possible so we can get it done - can loretta (and team) take another shot at it?

jw: think that whatever we do, this proposal for 2.4 is really a 1.3 issue because it's about programmatic determination, not about anything else

lgr: I would have said it's about orientation

gv: yes, it's a 1.3 category issue

js: something that you just said suggests that we need to revisit guideline wording ("orient themselves" may not fit under 2.4)


gv: question: 4.2 is so complicated, I wonder if we should try to tackle at the face to face

Guideline 2.5 - preliminary discussion of issue summary/proposal (Andi: 25 minutes)

next week: 2.5, 3.1

asw: 4.2 was coupled with baseline issue, which was a blocking issue that we needed to solve before moving forward - are we declaring that we've closed the issue?

gv: we have a definition that's better than what we had, but still needs work

asw: if we've agreed that baseline shouldn't be in guidelines (in principle), I'd hate to defer progress until next f2f

gv: think we've been working on assumption that baseline won't be in the guidelines and we'll work off that assumption until we're proven wrong - is that what everyone else thinks?

js: think we should continue discussing 4.2 in coming weeks even though progress is slow

gv: suggestion to find those who are most interested in addressing this one and take an offline phone call/work session to try to see if we can help work it through
... would like to ask for volunteers to do a multi-hour work session with loretta to help her out with this

lgr: we might want to sunset that group and see who wants to continue

<scribe> ACTION: loretta to follow up with subgroup (plus gregg) for continued 4.2 discussion [recorded in http://www.w3.org/2005/05/05-wai-wcag-minutes.html#action04]

Summary of Action Items

[NEW] ACTION: joe, gregg, becky to work on revised proposal for 1.3 [recorded in http://www.w3.org/2005/05/05-wai-wcag-minutes.html#action02]
[NEW] ACTION: john, yvette, joe, michael - revisit 2.4 wording and repropose [recorded in http://www.w3.org/2005/05/05-wai-wcag-minutes.html#action01]
[NEW] ACTION: loretta to follow up with subgroup (plus gregg) for continued 4.2 discussion [recorded in http://www.w3.org/2005/05/05-wai-wcag-minutes.html#action04]
[NEW] ACTION: loretta to wordsmith definition of baseline and post to list [recorded in http://www.w3.org/2005/05/05-wai-wcag-minutes.html#action03]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.122 (CVS log)
$Date: 2005/05/05 22:06:59 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.122  of Date: 2005/03/31 04:43:41  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/somethinbg/something/
Found Scribe: David_MacDonald
Found Scribe: Yvette
Inferring ScribeNick: Yvette
Found Scribe: ben
Inferring ScribeNick: ben
Scribes: David_MacDonald, Yvette, ben
ScribeNicks: Yvette, ben
Present: Andi Becky_Gibson Bengt_Farre Dave_MacDonald Gregg_and_Ben JasonWhite Joe_Clark John_Slatin Loretta_Guarino_Reid LucaMascaro Makoto Matt Michael Michael_Cooper Microsoft Mike_Barta Roberto_Castaldo Sebastiano Tim_Boland Yvette_Hoitink
Regrets: Roberto_Ellero Doyle Neil_Soiffer Roberto_Scano Takayuki_Watanabe
Agenda: http://lists.w3.org/Archives/Public/w3c-wai-gl/2005AprJun/0392.html
Got date from IRC log name: 5 May 2005
Guessing minutes URL: http://www.w3.org/2005/05/05-wai-wcag-minutes.html
People with action items: becky gregg joe john loretta

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]