Re: General Exception for Essential Purpose

"Leonard R. Kasday" wrote:
> 
> Ian,
> 
> To help me understand your point of view, would you say whether , for each
> of the cases I described, graphic text should be allowed and whether an
> alternative page is needed or not?  I repeat them below for convenience.
 
> Here, "chilly font" is a font with icicles dripping from the letters,

Cool! As it were...

> that's hard to read for people with low vision even with
> magnification.  However, it is not copyrighted, or part of a logo.

Assume the following user needs (there may be others):

 User need 1) Text content must be available as 
              visually-displayed text, speech, and braille
              to users who are deaf, blind, and deaf-blind 
              respectively. I assume users have access to 
              software to render all three of these.

 User need 2) Users with low vision require control of present
              parameters of text content (e.g., speech characteristics, 
              color, font family, and size of rendered text).

Call these UN1 and UN2. 

I don't think it's useful to distinguish the cases of author
intent as suggested by the cases below (i.e., to have WCAG requirements 
that identify different author intentions such as "for art", 
"for branding", etc.). In one model of accessibility guidelines
(heavily influenced by a recent conversation with Charles):

1) User needs come first.
2) Then you meet them by assigning responsibilities to
   authors, spec writers, and software developers.
3) Then you shuffle these responsibilities based on the real world.

So in all your scenarios:

I) If the author uses an image to encode the chilly effect:

  1) To meet UN1, the author must provide a text equivalent 
     and it should be well-structured. Call this author
     responsibility 1 (AR1).

  2) To meet UN2, the author must encode the image in
     a format that allows user control of presentation
     (e.g., SVG). Call this author responsibility 2 (AR2).

II) If the author use text + style/markup to encode the chilly
    effect, nothing else is required.

UN1 or UN2 or both will not be met if either:

   a) No user agent implements completely or properly the
      specifications used by the author. This is user agent
      responsibility 1 (UAR1). User agents have other 
      responsibilities such as making available the alternatives, 
      etc.

   b) The author hasn't met AR1 or AR2. This may be because
      the author has neglected to include a required text
      equivalent, or because no specification exists that
      allows the author to meet AR1 or AR2.

When the ideal world breaks down (due to inadequate specs,
bad authoring, or bad implementations), it becomes necessary
to shift responsibilities (e.g., from author to user
agent for bad authoring, or user agent to author to 
compensate for poor implementations, etc.) So if AR1
is not met (author forgot to include a text equivalent),
then the user agent may help somewhat by generating repair
text.

However, it may not be possible to shift responsibilities 
enough to meet some user needs. Web content requirements 
can only go so far, because they are predicated on some 
conditions for  a perfect world. This WG has started looking
into a model where the world is not perfect, and building up 
from minimal requirements for user agents. This will benefit
WCAG 2.0.

Looking at what can be done today (independent of any
specific authoring scenario):
 
 1) Since CSS2's Web Fonts are not implemented, 
    authors cannot realistically use them with HTML for 
    fancy effects such as the chilly effect. However, 
    CSS's generic fonts mean that for many other font styling 
    effects, style sheets can be used.

 2) CSS is implemented enough to allow users
    to control text color through style sheets. But if you
    want to control font family and color, you lose.
  
 3) SVG is not yet widely supported. The first
    public working draft of SVG was in February 1999 and
    WCAG 1.0 was published in May 1999. I don't think
    3.1 in WCAG 1.0 was designed to address the user 
    requirement to control presentation of content in general,
    but it does include some requirements for user control
    of text presentation. 

 4) Authors can encode text in raster images at
    a variety of sizes and make several images
    available to users. This could be done
    on alternative pages since user agent's
    wouldn't be able to negotiate the different
    sizes; users will have to choose manually.
 
 5) If the author encodes chilly information in
    a single raster image, UN2 will not be met.

I think WCAG 1.0 mostly focused on user needs, and
so I don't think it can be twisted too far to account
for what the author can really do today in light of
support for markup languages in deployed browseres.
I still think WCAG 2.0 shouldn't include author intentions
(e.g., it's ok to do this for branding) in the formulation
of user requirements, but it should go farther in modeling 
(with how much flexibility for evolution?) the real world.

I hope I answered your question adequately. I didn't mean
to skirt it. <smile>

 - Ian

> ----
> 
> 1. a site teaching kids to read, where the word "cold" is written in chilly
> font.
> font ok?
> alt page needed?
>
> 2. a schedule of public meetings of a township zoning board, where all the
> winter dates are written in chilly font.
> font OK?
> alt page needed?
> 
> 3. an artistic page, with poems about winter, incorporating chilly font.
> font OK?
> alt page needed?
> 
> 4. A hypothetical web site for a company "frosty cola".
> - a splash screen with chilly font
> font OK?
> alt page needed?
> 
> - an employment listing with job listings in chilly font.
> font ok?
> alt page needed?
> 
> Should WCAG 1.0 say any to distinguish these cases?
> Should WCAG 2.0 say any to distinguish these cases?
> 
> If so, what?
> 
> Len
> 
> At 01:51 PM 10/27/00 -0400, Ian Jacobs wrote:
> >"Leonard R. Kasday" wrote:
> > >
> > > This is a proposal for a general guideline which will I think, help resolve
> > > the image text issue, and other issues as well. Its a modification of WCAG
> > > 1.0 checkpoint 11.4 on alternative pages.    It also needs to be added to
> > > WCAG 2.0--I don't see an equivalent to 11.4 in 2.0.
> > >
> > > Here's straw wording of what I'll call the "essential purpose" guideline.
> > >
> > > <guideline>
> > > If a web page's essential purpose prevents you from satisfying a
> > > checkpoint, you can consider that checkpoint passed if the user can
> > > conveniently access an alternative page on which the checkpoint is
> > satisfied.
> > > </guideline>
> >
> >"Essential purpose" refer to author's intent. The Guidelines were
> >designed based on user needs. The tension between the two is
> >the source of many issues.
> >
> >In order for the author's intent to translate into content that
> >is accessible to users with disabilities, you must have
> >markup languages that allow authors to express their intent with
> >content that can still be manipulated to meet user needs. You
> >must also have user agent support for those markup languages.
> >
> >WCAG 1.0 was clearly designed to address user needs first
> >and author intention second. I think that it may be best to
> >leave it that way (even in clarifications), and that
> >WCAG 2.0 focus more on the author.
> >
> >Users have requirements. Responsibilities to meet those requirements
> >lie with specification writers, authors, and software developers.
> >As Charles has pointed out, in an ideal world, it may be easy
> >to divvy up responsibilities among those three camps. But in the
> >real world, the challenge is to identify who must do what, who
> >cannot do what, and what requirements may not be met by a given
> >set of guidelines. I don't think WCAG 1.0 was designed to, or
> >is flexible enough to, meet the needs of all users with disabilities.


-- 
Ian Jacobs (jacobs@w3.org)   http://www.w3.org/People/Jacobs
Tel:                         +1 831 457-2842
Cell:                        +1 917 450-8783

Received on Friday, 27 October 2000 17:15:07 UTC