See also: IRC log
Jan set the stage for the morning session
Part B is support for the production of Accesible Content. We will be using http://www.w3.org/WAI/AU/2008/WD-ATAG20-20080728/WD-ATAG20-20080728.html
For conformance, you can limit the claim to a limited set of of the fomats that the tool covers.
Began with a discussion of the Web Content Accessibility Benchmark Document
Must be published on a public website with a license that permits it to be copied
The name and version of the web contet technnology or technologie and Version and URI of the Web Content Accessibility Standard being referenced
<Jan> While the Working Group highly recommends the W3C-WAI Web Content Accessibility Guidelines as it was developed under the same W3C process as ATAG, other recommendations, standards, and regulations with the same goal exist in jurisdictions and organizations around the world.
Language changed in the Note on other accessibility standards that allows for developing of guidelines other than W3C
3 target levels, A, AA, and AAA
Assumptions about user agents available to authors and end users
Item 4 needs to be revisited and clarified "Assumptions about User Agents"
<jeanne> Enter issues at http://www.w3.org/WAI/AU/tracker
Proper Benchmarks for each normative requirement of the accessibility standard at the target level defined
Issue: Is ATAG requiring more than what is being required by WCAG?
<trackbot> Created ISSUE-147 - Is ATAG requiring more than what is being required by WCAG? ; please complete additional details at http://www.w3.org/WAI/AU/tracker/issues/147/edit .
Issue Assumptions about User Agents
Issue: Assumptions about User Agents
<trackbot> Created ISSUE-148 - Assumptions about User Agents ; please complete additional details at http://www.w3.org/WAI/AU/tracker/issues/148/edit .
Discussion resumes at Level A Success Criteria for Guideline B.1.1
Level AA and Level AAA Success Criteria for Guidleine B.1.1 also described in this success criteria section
<Reed> and then sub sections
Issue: Need to clarify the distinctions between success criteria for Level A, AA, AAA success criterias for B.1.1
<trackbot> Created ISSUE-149 - Need to clarify the distinctions between success criteria for Level A, AA, AAA success criterias for B.1.1 ; please complete additional details at http://www.w3.org/WAI/AU/tracker/issues/149/edit .
<jeanne> JT: THe goal is to present a format where the technologies are more accessible. Presenting it that way, may get around the problmes we are having. How do we encourage more accesisiblity as we move through A to AA to AAA?
<jeanne> CS: IN WCAG, we expect that majority of people will never make a conformance claim, but will use it to help them build accessible sites. ATAG can't be used as tool for building an accessible product.
<jeanne> CS: Recommend techniques that explain this.
<jeanne> JT: Give the authors informed choice of what is the most accessible technology to produce web content.
<jeanne> MS: We need a more work-flow oriented approach.
<jeanne> JT: We should talk about what work flow process implement B.1.1 instead of going right into the benchmark.
<jutta> The goal is: Support Web content technologies that enable the creation of content that is accessible. The success criteria that support this are: to provide accessible WCT among your choices of WCT, provide a choice of WCT with accessible WCT given priority or selected by default, inform authors about which WCT are more accessible, inform authors when they are choosing less accessible WCT.
RS concern the word content has 2 meanings
e.g. a table or the whole document
RS suggests content elements
MS could also have the conversion of a whole document
DS guideline doesn't indicate whether the content comes from another tool initially
RS is worried that the tool is going to save the old file with accessibility information without giving the user to not do this
RS is worried that the tool is going to save the old file with accessibility information without giving the user <the option> to not do this
RS example - ministry of defence - may have information that has security issues if it's preserved
e.g. alt text on secret images
adding an option to query the user for the information that is going to be lost / modified
JS concerned that preserved output should be available for end users
JR end users implies AT users as well
adding AT users to "end users" definition
MS concerned whether B.1.2.2 (a) is technically possible for the tool to decide what information is accessibility information
RS e.g. styles information is difficult for the tool to decide on
JT the user could choose to decide whether to receive notification of info to be deleted or not
JR the basic point is respect for author generated content
JT the tool doesn't necessarily need to determine what to delete or not as long as it notifies the user
RS low cost option for developers would be to provide the option to not delete information
adding a no deletion option - user can turn off any deletion by the tool
no open issues for this guidelines
CS some concerns around the use of the word "prompt"
prompt is quite an aggressive word
implies something being in your face in the UI
brainstorm for word to replace prompt
dialog, fields, query, inform, opportunity for input, alert (probably not), cue, provide input
AM suggest encourage, request
guide is another option
<AnnM> you've all gone rather quiet - it's really difficult to hear the discussion
changing to guide and cue rather than prompt
JS 2.1.2 if the user turns the prompts off, this isn't the tools fault
another scenario is plugins that may disable prompts in the tool
RS suggests the example for image insertion should be a little broader
RS we don't want to stipulate UI design, e.g. where the alt text option should be placed
checker needs to be granular
<jeanne> http://www.w3.org/WAI/AU/tracker/issues/78 - All SC needs to be evaluated for testability.
<jeanne> http://www.w3.org/WAI/AU/tracker/issues/79 - B.2.2.1 makes HTML a gating factor.
<jeanne> issue 79?
<trackbot> ISSUE-79 -- B.2.2.1 Makes HTML a gating factor â€“ needs to be more general. -- OPEN
DS we shouldn't specifiy too much how the checks are carried out
e.g. stepping through them all compared to an overall analysis
<trackbot> ISSUE-151 -- B.2 Are we requiring validation be built in. Coordination with Note -- OPEN
Range is two dimensional. Coverage is already in there (e.g. element, entire file etc). Needs to also cover what checks to do, e.g. don't check for captioning if there are no videos etc
DS may need to consider the different roles of authors when checking
e.g. there may be someone else who does the checking rather than the original author
<jeanne> AR: At RNIVB< we have many authors, but nothing get published until it has been checked by someone else who knows what to look for. It helps with education, as well.
RS concerned about conformance claims that may burden a third party without their knowledge / consent
adding a note that the burden is on the conformance claimant
rather than the developers of any of the software components included in the claim
Definition of Authoring tool
Dana: Should defintion of authoring tools include test tools?
Mike: I don't like that. Are CMS even authoring tools?
Cynthia, Jan: yes. templates count
Should we use tool/system to make it more clear that CMS counts?
Change defintion to create, modify, or aggregate content
Dana: Can an agregator fix broken
... CNN.com aggregates content from anywhere
Reed: what about Google? content from anywhere
Ok, those don't count, covered by WCAG
Dreamweaver doesn't edit images, but needs to be able to repair it by editing its metadata
Jutta: Fine balace. Many authors are
... we need to be very careful to include things that are creating content through distributed authoring.
Jan: start from where we dont' want
to be and work back
... all agree that a CMS where people are putting in pages and images, we want that to be an authoring tools
... on the far side, soemthing like google search that grabs urls
Jutta: display of linsk in google counts
Cynthia, Reed: But that's WCAG
Jutta: Google doesn't interact with authors. Let's not take effort to exclude google. Content it automatically generates must be accessible.
Dana: What about a learning management system? creates new content and aggregates existing content
Reed: not sure I agree that we shouldn't take extra steps to narrow the scope
Jutta: Before we add extra complexity to definitions, let's see if we are alreayd excluding it
Jan: Is there anything on the outside
that we want in
... quickly think of the case of a learning management system that has parts that create original content, but author can also drop in a pdf that's a historical thing. Is that different than dropping pdf into a page?
... draw a larger circle and bring in other tools. If that LMS is accessible, it would ahve a means of making content that the author brings in accessible
<Reed> Cynthia: a lot of that is already covered under WCAG, what we need to cover here is how do we ask for alternate versions?
<jeanne> CS: A lot is covered under WCAG, and the only part that should be included here is how to ask for or include an accessible version.
Jan: is the PDF any different than an image?
Dana: images are easy to make accessible. what about a video, or a query response?
Jutta: Query should have a style that is accessible. The captions should be there for the video. The authoring tool should prompt for it.
Reed: assumes that the author has access to the original and can fix it.
Jan: if the author has it. It's about not cutting off the possiblity.
Reed: 3 authoring tools feeding into a CMS. 2 of them have accessibiltiy support. Is it the CMS job to have to ask for or repurpose that content.
<jeanne> Reed: The developer of the content management system is going to have to fix the output of the previous authoring tool that didn't have the accessibility.
<jeanne> Cynthia: That is required by WCAG of content aggregators.
<jeanne> JT: Educational institutions are required to go back and make content they aggregate accessible.
<jeanne> Reed: the CMS can not go back to the originating tool to repair the content.
<jeanne> Dana: We don't want to create a standard that is limited to specific file types. People will use this to exclude software, to say that they are not doing the right thing.
<jeanne> Jan: Anything that is external would get a "get out of jail" card from ATAG.
<jeanne> Dana: If the product can recognize the product, it should.
<jeanne> Cynthia: WCAG would make that level 2 or 3.
<jeanne> Dana: Where the tool can identify a known format that it can modify or repair, it should.
<jeanne> Greg: We are getting outside the scope of the tool.
<jeanne> Jutta: Don't forget, we have the option to ask the author.
Reed: parts of content management systems that include creation of new content
<jeanne> Cynthia: Let's look at the tasks to be accomplished rather than the types of software?
Jan: how is an IT person setting up a system an author?
Jutta: Choosing a style template, creating templates, creating content infrastructure that influence many things down the line
Jan: Decided that there should be a banner ad
Jutta: setting up style for pages, creating templates
Reed: hwo si that not covered by WCAG
Jan: it is covered by WCAG
Dana; setting up templates is something that we want covered
Jutta: ATAG should include influencing the decisions an author makes
Jan: aggregating content from CNN
<jeanne> Cynthia: WCAG says that they are still responsible for streamed or aggregated content.
<jeanne> Greg: We are putting a big burden on aggregators to evaluate the accessibility of the imported content.
<Andrew> suggestion from Jutta:
<Andrew> The criteria for the bounds of an authoring tool for the purposes of
<Andrew> ATAG are: 1. the application processes that interact with a human
<Andrew> author, and 2. the authoring choices that author is making or 3. the
<Andrew> authoring choices under the control of the authoring tool.
<Andrew> Authoring choices include choice of style sheets, templates, scripts, etc.
Jan: get back to linking in sydicated dynamic content
Jean: get back to B2.2.2
Jan: predicatabilitiy for human author, checking stuff that the author brings in, requires human interaction
Jutta: doesn't matter if it's dynamic or static, but it's decisions that the authro can make
Reed: if the tool includes
functionality to modify the content type, then it should check
... but if it's pulling in functionality that it can't edit.
Jan: Dreamweaver can't edit gif, but can make it accessible
Reed: Big scale difference
Jutta: not unreasonable to require it pull in an OCR system
Reed: yes it is unreasoanble. Licencing, scale, feasibility, etc.
Mike: Minor issue: note at end of B2 says nothing is actually reqyiured??
Jan: only manual checking required at minimum, not automatic.
Jutta: only cover things the author
has the ability to modify
... add a preamble to B2 that talks about the bounds of supporting the author. Only covers things that the author actually has a choice about.
Make more detailed handling of embedded documents a best practice
Question: should it be a best practice, a lower priority SC, or soemthing else
Work on a preambel
<Reed> The criteria for the bounds of an authoring tool for the purposes of ATAG are: 1. the application processes that interact with a human author, and 2. the authoring choices that author is making or 3. the authoring choices under the control of the authoring tool.
Cynthia: each function of the tool is
responsible for helping the author to make the best content for
... how do we help the author choose between the products of other tools?
Moving on to B.2.2.2
Reed: should be able to check for things that I can modify
Jan: then how do we handle the alt text on a gif in Dreamweaver
Reed: put that as an example of modifying
Dana: dynamic source
Jan: handle it separately. unknowability can happen with any format
How do we handle "can't control"
Reed: tool can't decide for me what
source to use
... what realistically can the authoring tool do to promote that?
Jan: Point you to a way to make a text version is practially too difficult.
Jutta: making choices when setting up templates. Many decisions where the content and organization. Authoring tool provides the facility for those choices. Do it at the level of what choices we make with this authoring tool.
Reed: too much duplication. CNN aggregator system. My feed providers should have done the right thing. I shouldn't be removing their accessibility info, but is it my job to re-check them?
Jutta: Not saying what the authroing tool should recheck, but how to support the author
Jan: as we go up the practical scale, have a way to specify the equivalent.
Dana: if the orig content has accessibiltiy feautres, and if the tool can recognize and modify it, then it should. If nay one of those is missing, the tool can't do it.
Jan: example of video presentation you don't own.
Reed: Author 1 making a movie. Movie making tool should have prompted me to add captions. Author 2 is using a CMS and pulls in movie. A2 shouldn't be responsible for captions.
Jan: ask A2 if there are captions. If not, is there a version that does?
Cynthia: WCAG says that those movies pulled in will either be known to be accessible (such as by contract) or will be exempted in a statement of partial conformance.
Jan: exmaple of scanned PDF of declaration of independence in a history class
Reed: let's try a different approach.
If I'm a CMS system that wants to be accessible. What do I have to
do to become compliant? Is that list too big?
... Level A: provide a dialog that says favor the most accessible version
Cynthia: does that help teh author?
Greg: What about an educational portal pulling in images from an archive, where those images don't have alt text?
Cynthia: Guides and Cues are UI that influence author behavior. Checks see if that guide/cue worked.
Checks are much more explicit, and a different kind of programming
Checks happen after insertion, Cues before
Jan: Fixing note at the end of B2
Moving onto B2.2
Reed: We are forcing CMS builders to build the worlds biggest checking tool
Jutta: check for the decisions the
CMS supports. Teh CMS is not what's used to create a paritcular
element, but does make decsions about thing that effect
... for each level of authoring, it supports the decision.
Cynthia: Is the choice of formats to embed one of the choices that an author using a CMS makes?
<jeanne> B2.2.1 The authoring tool or an external checking or validation tool checks that the choices made by the author are accessible prior to publishing.
<jeanne> B2.2.1 The authoring tool or an external checking or validation tool checks that the choices made by the author are accessible prior to publishing. The tool only must check for objects or elements that the tool has the ability to modify.
Jutta: modify or augment
Jan: Dreamweaver can't change gifs but can change html to make it accessible anyway
The tool must check for objects or elements that the tool has the ability to modify. The tool should check for things it has the ability to augment.
Jan: can we do something with remembering what decision the person has made? If I drag something onto my document I would be cued. Once I've been in the dialog, I've made a decision.
Jutta: add reasonable to augment. create some defintion of ability to augment.
Jean: what are the similarities between the things that are reasonable to augment?
Jutta: capability to augment, ability to augment, can a tool augment. A CMS can't augment a video with synchronized captions.
Cynthia: but it could link to a transcript.
Jutta: can we define what is reasonable?
Cynthia: I think we'd get disagreement about whether it's reasonable to requier that CMS user type a transcript, even though it's technologically feasible.
<jeanne> ACTION: Jan will draft a proposal for reasonable augment by 7/28 [recorded in http://www.w3.org/2008/07/28-au-irc]
<trackbot> Created ACTION-2 - Will draft a proposal for reasonable augment by 7/28 [on Jan Richards - due 2008-08-04].
This is scribe.perl Revision: 1.133 of Date: 2008/01/18 18:48:51 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/weel/well/ Succeeded: s/Juta/Jutta/ Found Scribe: Greg Inferring ScribeNick: Greg Found Scribe: Andrew Inferring ScribeNick: Andrew Found Scribe: cynthia Inferring ScribeNick: Cynthia Scribes: Greg, Andrew, cynthia ScribeNicks: Greg, Andrew, Cynthia WARNING: No "Present: ... " found! Possibly Present: AR Andrew Andrew_ Andrew_Ronksley Andrew_Ronksley_ AnnM AnnM_ CS Cynthia Dana Danalouise Greg Greg_ Issue JT Jan Jan_ Jean MS Mike Reed Suggestion gpisocky jeanne jutta michael_squillace trackbot You can indicate people for the Present list like this: <dbooth> Present: dbooth jonathan mary <dbooth> Present+ amy WARNING: No meeting title found! You should specify the meeting title like this: <dbooth> Meeting: Weekly Baking Club Meeting WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Got date from IRC log name: 28 Jul 2008 People with action items: jan WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]