19 October 2000 minutes

available at: http://www.w3.org/WAI/GL/2000/10/19-minutes.html

19 October 2000 WCAG WG telecon

Summary of action items
·       Action JW: Propose text to cover the issue about associating text 
equivalents with what they represent.
·       Action GR: Try to abstract JW's proposal.
·       Action WC: rework proposal for checkpoint 3.1.
·       Action JW: Take issue of maintaining info about assistive 
technology and alternative browsing to Education and Outreach via the 
Coordination Group.
·       Action CS: write up a browser profile for what authors should support.

Participants
·       Claus Thøgersen
·       Marshall Jansen
·       Matt May
·       Jason White
·       Dick Brown
·       Loretta Guarino Reid
·       Cynthia Shelly
·       Charles McCathieNevile
·       Kynn Barlett
·       Gregg Vanderheiden
·       Gregory Rosmaita
·       William Loughborough

Next face to face
Two Day Preference: Mon/Tues or Thurs/Fri. (Feb. 26-27 or March 1-2)
Will your group be flexible about the meeting days: If you choose Mon/Tues; 
would Thurs/Fri be acceptable?
The number of people in your group that will attend the f2f meeting __
Will folks in your group attend the Wednesday Plenary?
What other W3C groups, specifically, would your group like to meet?
Are there other groups with significant membership overlap with yours which 
should NOT be scheduled on the same days?
GV I teach on Monday, so I prefer Thurs/Fri.
About 10 - 15 people will attend.
Mobile, device independent authoring, other WAI groups, XHTML, XForms, 
Voice Browser, XLink, XPointer.

New draft
WC Not out yet. It's close.
JW Then we'll vote on the list. I'll take a stab at finishing up proposal 
for checkpoint 1.1. Please raise remaining issues with checkpoint 1.1.
GV If you post wording and know there are things that were not addressed, 
please mention this.
WC That discussion is in the minutes from the face to face.

Text equivalents be identified in markup
JW Impressions of whether people want a requirement. A separate checkpoint?
LGR Is the concern that it is hard to confirm that text equivalents exist?
JW It is hard to generate something with text equivalent, especially if a 
document generated for user preferences.
GV For example?
GR In UA context, the ability to embed text descriptions into clip art. You 
can also find copyright info. If UA has a way to extract, and AT's are 
being encouraged to do that for clip art they are distributing.
GV If no alt-text why wouldn't the tool look inside? If it's there, it can 
easily find.
GR The request is from an authoring standpoint. I think this was raised for 
clarification. How do people believe we should deal with this issue? Should 
we throw it to UA or ATAG? Gets to the heart of "web content" is it the 
component or document or application.
CT This idea of section 5 is problematic wrt UA.
GV If it's something like a picture you could grab the info from the data 
format, but if you have turned graphics off, then the UA would not know to 
fetch it.
CT I understand the issue to be: how much do we want to know about a structure.
JW Who wants this requirement? Each text equivalent must be marked up as 
per guideline 2 from the accompanying text.
CMN Yes, I would like to see such a proposal.
DB Still having troubling understanding it? Must be distinguished?
JW Can do in SVG or in HTML w/alt. However, this needs to be distinct in 
the markup.
CMN Example is: alt in HTML - this is an alt for this image. an example of 
not doing it: having a chart w/alt that says, "great britain chart" and 
then in the content a description. no way to associate them.
DB If someone has in the text that would not be explicitly linked. In HTML 
4 could use longdesc.
GR In another application, they use captions to describe photos. The 
alt-text is then usually "photo" or "black and white photo."
CS An interesting idea, sounds like a feature to add to HTML.
WC I think that it is clear from 1.1
GR It can be interpreted to mean that if you do OCR.
JW Or if you generate something with descriptions with no markup, you're 
rendering it as text.
GV Two parts:
alternative text not distinct from text
marked up but not obvious
alt-text will be distinct from, if not always rendered and not alt. Knowing 
that it is there is what we need to focus on. "If alt-text provided, it 
must be obvious from the markup to indicate its presence."
Action JW: Propose text to cover the issue about associating text 
equivalents with what they represent.
Action GR: Try to abstract JW's proposal.
CS This does not work for all interfaces. If a voice interface, it doesn't 
matter if a graphic button exists somewhere.
CT What if you can turn it off?
CS That will cause problems for people with cognitive disabilities.
CT You can not decide for the user.
KB This assumption comes with the idea that there is an optimal 
presentation. From our work, there are different ways of presenting 
information. I don't need to present to a visual user if they have said, 
"don't play me sound" to let them know that there is sound here. I spoke 
with Ian about this because it sounded odd. He specifically said it has to 
be clear in the markup or the data model. It may be on my server - an 
explicit representation between this image and this text. He said that as 
long as in the data model, I would be covered. This does not have to be 
sent to the user.
GR That is a basic underlying principle.
KB Then the requirement is still odd to me.
CS I agree with Kynn there are other ways to get around the problem.
JW There are still issues.
CS I think we are talking about 2 issues.
Should things be associated
Informing the user of other forms of the information

Errata to 3.1
CS As long as not a paragraph.
CMN Problem for magnifiers.
WC Can use Opera to magnify the alt-text of images.
CS Section headings are usually large fonts.
WC Text in images to create logos? Does anyone disagree?
CMN I am not convinced.
JW One could avoid the implication that you can interpret to mean it does 
not exclude every image.
GV If you can't do something so it will work with the browsers on your 
site, does 3.1 say "it doesn't matter, AA means you must use markup language."
JW Did we add into the Errata, Ian's proposal in 11.1? That would take care 
of it.
WC No.
CMN It rules out things that can be done using images that should be done 
in markup. Use MathML to represent math. What can't be done using markup? 
You don't have strong control over button appearance - how good is css support?
WC /* restate my proposal */
KB Does proposal cover WAI logo?
WC Yes.Let's keep checkpoint as is, but write a clarification that image ok 
for logo, navigation buttons, image maps.
GV "until widely supported" - what if 2/3 of browsers support it. Does that 
mean we switch? We have a question for how long it is that people are 
required to do things.
WC Since 1.0 errata, i think we can use the until user agent language 
because 2.0 should be out before until user agent is met.
JW Don't think resolve in 1.0 w/errata change, therefore "when an 
appropriate markup language exists" means "when supported by user agents."
MM: A lot of companies won't want to put SVG out for public consumption. 
Once it's out there it can be stolen. A lot of companies use graphical 
content to protect info so that things can't be perfectly copied. Would 
people want to adopt SVG?
CMN I don't think that holds. If someone puts an imperfect logo out there 
is making that logo available whether it is SVG, or gif, or whatever.
GR A legal issue.
CS What about the word "appropriate." What CMN finds appropriate is 
different than what a legal person finds appropriate.
KB I think most people will want to do specific things that CSS and HTML 
will not be acceptable solutions. MathML is not acceptable. There are 
things you can't do.
GV Will be a list or tie back to 11 - where possible to do that using xyz 
then you must. The word appropriate is vague. It is not from the rule but 
the explanatory text underneath it.
CMN The alternative interpretation is for cases where CSS works on Netscape 
and Explorer etc. then the answer is to apply 11.4 and supply 2 versions. 
That clearly meets the guidelines as written.
Action WC: rework proposal for checkpoint 3.1.

Until user agents proposal
CMN How do we know when user agents are sufficient. I have proposed a set 
of conditions.
·       Things have to work in 2 browsers. One browser must be free and 
work across MacOS, Linux, Windows, etc.
·       It must work with X number of assistive technologies.
One issue is, how will this pan out internationally. There were two 
approaches proposed.
say that X number of months after these are available and have been shown 
to work we expect people to use them. is it 6 months after they become 
available or localized to a language? we need input on that from 
non-english speakers. from the european blind union, i hear that there are 
whole countries where people who are blind use DOS. Therefore, not 
supporting DOS is an issue.
CS If designing for English, then only concerned about English tools?
CT Who will document what is being used in different countries? And keep it 
updated?
CS Someone launching a site in a particular area would do that research.
GR in U.S., baseline is DOS or Windows 95. Discussed with RNIB, people 
upgrade to Windows NT.
CMN Available technical support is a major issue.
DB I could not find anything in the existing guidelines where we mention 
that products are free. Do we really want to do that? What about the hardware?
CMN That's another issue.
CT The only thing that is free is unix things (re: AT)
CS The Narrator that ships with Windows 2000 is free, but Windows is not.
CMN Should we be requiring that if you need something free that you live on 
a linux system. or we require that there be one free solution per 
platform.. Or does everyone pay the same.
WC What info will help us answer this info?
CMN There are some things that we should just answer.
JW We need to do it in a neutral way to suppliers.
CT We need a way to collect data. Can EO help here?
CMN Probably. They maintain a policy page and an assistive technology and 
alternative browsing page. We should ask them because they have wider 
international contacts. The I18N is in a sense simple. There are only a few 
questions about how to apply the rules we decide on. Then we have to answer 
the same questions in different languages.
Action JW: Take issue of maintaining info about assistive technology and 
alternative browsing to Education and Outreach via the Coordination Group.
GR An opportunity for EO to contact disability groups to collect information.
CMN We could look at the components people are using and say "yes" or "no".
CS Like Javascript?
CMN Right: we require CSS Font, don't require CSS positioning, yes 
JavaScript, etc.
CS That seems a lot easier.
WC Create a profile.
CS Then people pushed to meet that reference.
JW That's UA.
CMN They don't specify technologies.
Action CS: write up a browser profile for what authors should support.

$Date: 2000/10/19 23:10:58 $ Wendy Chisholm

--
wendy a chisholm
world wide web consortium
web accessibility initiative
madison, wi usa
tel: +1 608 663 6346
/--

Received on Thursday, 19 October 2000 19:12:15 UTC