Re: PROPOSAL: Resolution of PR#207

Jon Gunderson wrote:
> 
> There has been much discussion related to checkpoint 2.1.  These are some
> of my observations:
> 
> A. Consensus on access to all human readable (if a definition could be
> found) content through the user interface
> 
> B. Consensus on access to all alternative equivalents through the user
> interface
> 
> C. Currently the group has identified the primary use for access to machine
> readable content through the user interface is for providing access to
> poorly authored pages or technologies that do not fully support accessibility.
> 
> The following items summarize the chairs view of the issue:
> 
> 1. Seems to be most people in the working group feel that people should
> have access to all the information the author is providing through the UI.
> 
> 2. Difficult to draw lines between human and machine types of content
> 
> 3. Access versus usability issues: Most techniques for access and
> understand some types of author supplied information would require a higher
> than average knowledge and skill in the technologies that are being used.
> 
> 4.Source seems to be a least common denominator, but is not very usable to
> most people.
> 
> 5.JW and AG feel that access to element attribute information is an
> important issue for XML (chair conversation)
> 
> 6. Changes to this checkpoint may impose new requirements and may require
> stepping back to a previous stage in the recommendation process.
> 
> =======================
> PROPOSAL:
> I proposed the following:
> 
> 1. Checkpoint 2.1 remains the same

Ok.

> 2. The minimum requirements are:
> 
> A. All views (or view ports, IJ help here) rendered by the user agent
> conform to the guidelines.

A "viewport" is the piece of the UI. A "view" is the slice of the data.
What does it mean that a viewport conform to the guidelines?
If you mean that the user agent must meet the user interface
requirements of UAAG 1.0, I agree. But that's no more than conformance
to the document. So I don't see "A" as a minimal requirement.
 
> B. All author supplied information must be available through at least one
> of the view ports offered by the user agent.  This does not mean that all
> of the information must be in one view (i.e. a source view).  But the
> combination of views offered by the user agent must provide access to all
> the information provided by the author. 

That seems reasonable as an elaboration of the checkpoint, but it
doesn't suggest a minimal requirement. It says "Whatever you do, all
the content must be available at some point through the UI.".

Perhaps we can resolve the issue by telling developers informally 
what is expected:

 Note. A user agent must make available all content through the
       user interface. A user agent is expected to process and 
       render content according to specification and to make available
       all content intended for the user (including equivalent
       alternatives for content) through viewports and other
       components of the user interface. A user agent is not required to
       render all content in the same viewport. A "source view" does not
       satisfy this checkpoint for content that may be rendered
naturally 
       through an ordinary viewport since users are not expected to 
       understand a markup language in order to use marked up content.
       However, at times (e.g., when the user interface or an assistive
       technology do not provide sufficient access)
       some users may need to view markup, style sheets, scripts, or
other       
       information in "raw form", so a user agent should offer a source
       view (or better) to ensure access to all content.

I would note that the above proposal is very similar to the one 
proposed by Charles [1] on 3 April:

<BLOCKQUOTE>
Note: For information intended to be human-readable, (for example
equivalent
alternatives) a source view alone is insufficient, and access must be
provided as part of the normal user browsing interface.
</BLOCKQUOTE>

I would note further that a source view alone would often mean that the
user agent fails checkpoint 6.2 (conform to W3C specs). A source view
of HTML is not an HTML view but an SGML view. Therefore, if the user
agent
provides no other view of the document, it cannot claim to be a
conforming
HTML user agent.

For me, the most difficult part of this issue has been "where to draw
the line" between what MUST be rendered through an "ordinary" viewport
and what must not. Drawing that line has clearly proved
difficult ("process according to specification" does not suffice, it
would seem). I think that technically, a source view would satisfy
this checkpoint minimally, but as many have argued, that's not
really sufficient for ensuring access.

I have been reticent to include a priority 1 requirement that users
have access through the user interface to content not meant to be
viewed.
In part, I don't think that many users will benefit from this
information
(since most users will not know enough about markup to make use of a
source
view). Secondly, all content must be available through and API anyway.
Thirdly, if some content is not available to any user, all users suffer
equally and thus a broken user interface in this case is a general
problem,
not an accessibility problem.

However, Charles and others have argued that the user agent must make
available all content natively. It is very difficult to draw the line
between
content intended for humans and content intended for machines; the line
moves according to user needs and may not be clearly specified anyway.
While in theory a source view is not required by this checkpoint
(a perfect user interface would make it unnecessary), in practice
we are requiring that every user agent implement a source view as a 
basic repair strategy for failures of the user interface. A source 
view does not suffice for general rendering, but without one, 
a user agent might not provide access to all content through the UI.

- Ian

[1] http://lists.w3.org/Archives/Public/w3c-wai-ua/2000AprJun/0007.html

> For example, non-source view ways
> of providing access to author supplied content not normally rendered
> through the UI:

These are all good techniques.

> 1. A context sensitive menu capability may provide option to view the
> author supplied attributes for the selected object in a view port.
> 
> 2. A document information dialog box may provide access to meta element
> information of a document.
> 
> C. A source view is typically only one of the views a user agent offers and
> is not very usable to most people.  But it would satisfy access to all
> content, but all other views of the document must still be accessible.  The
> only way a source view would be the only way to satisfy the requirements of
> 2.1 is if it was the only view offered by the user agent to anyone (not
> very likely for anything considering conforming to the UA guidelines).
> 
> Please comment:
> 
> Jon
> 
> Jon Gunderson, Ph.D., ATP
> Coordinator of Assistive Communication and Information Technology
> Chair, W3C WAI User Agent Working Group
> Division of Rehabilitation - Education Services
> College of Applied Life Studies
> University of Illinois at Urbana/Champaign
> 1207 S. Oak Street, Champaign, IL  61820
> 
> Voice: (217) 244-5870
> Fax: (217) 333-0248
> 
> E-mail: jongund@uiuc.edu
> 
> WWW: http://www.staff.uiuc.edu/~jongund
> WWW: http://www.w3.org/wai/ua

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

Received on Thursday, 20 April 2000 12:49:57 UTC