W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 2004

Re: [2.4] Updated proposal

From: Jason White <jasonw@ariel.its.unimelb.edu.au>
Date: Thu, 4 Nov 2004 13:23:12 +1100
Message-ID: <16777.37520.466597.517045@jdc.local>
To: "John M Slatin" <john_slatin@austin.utexas.edu>
Cc: <w3c-wai-gl@w3.org>

John M Slatin writes:
 > 
 > <proposed> 
 > 
 > Guideline 2.4. Provide tools to help users find content, orient
 > themselves within it, and navigate through it.
This seems reasonable enough.
 > 
 > > Level 1 success criteria for Guideline 2.4
 > 
 > 
 > 1.	Structures within the content can be determined
 > programmatically. 
 > 
 > Editor's note:This is very close to the wording of Guideline 1.3 L1 SC 1
 > ("Structures and relationships within the content can be determined
 > programmatically"). Several people who participated in the 28 October
 > call felt that there should be a Level 1 success criterion for this
 > guideline, and that it should be identical or nearly identical to 1.3 L1
 > SC1. There was some discussion about whether to use exactly the same
 > wording, or just link to 1.3 L1 SC1. But there was also some feeling
 > that the reference to "relationships" expressed in 1.3 L1 SC1 would not
 > be appropriate here. So the proposed wording is slightly different from
 > that in 1.3.

Since the distinction between "structures" and "relationships" is very
ambiguous as it is, using one phrase in 1.3 and another here only
creates confusion. I suggest either linking to 1.3 or using identical
wording. I disagree with the comment that there are "relationships"
relevant to 1.3 that aren't relevant here, as this suggestion has not
been backed up by clear arguments, definitions or examples.
 > 
 > 
 > Level 2 success criteria for Guideline 2.4
 > 
 > 
 > 1.	Documents that have five or more section headings and are
 > presented as a single perceivable unit include a Table of Contents with
 > links to important sections of the document. 
 > 
 > 	Editor's note:This success criterion may become unnecessary if
 > the proposal to make UAAG 1.0 the baseline is adopted. User agents that
 > conform to UAAG 1.0, checkpoints 9.9 (Allow structured navigation
 > [Priority 2]) and 10.4 (Provide outline view [Priority 2]( could create
 > a Table of Contents automatically using structural markup within the
 > document. (There are extensions for Firefox that provide outline view,
 > and there are similar extensions/plug-ins for both IE and Mozilla.
 > However, the UAAG checkpoints referenced above are Priority 2, and we
 > have not yet discussed the level of conformance to UAAG 1.0 that would
 > be required in our baseline.

I agree with the comments made in the note. Also, is there any
cognitive rationale for 5 headings as the threshold?
 > 
 > 2.	There is more than one way to locate content, including but not
 > limited to link groups, a site map, site search or other navigation
 > mechanism. 

This runs into the problems I pointed out at last week's
teleconference. If conformance is always claimed at the "delivery
unit" level, site-wide scope is not possible. This can be fixed by
rewriting the success criterion along the following lines:

"There is more than one way to locate the content of each delivery
unit", or similar wording. In other terms, this delivery unit can be
reached by more than one method.

As a more global remark, we need to agree on a small number of terms:
"content", "delivery unit", "authored unit", etc., and use them
consistently throughout the guidelines to replace ambiguous terms such
as "site", "page", etc.

There is another problem with this success criterion: how does it
apply if the Web content is a user interface? For example, a delivery
unit might be an error message in an application, reachable only if an
error is signalled. Thus there is only one way of "locating" the error
message - by triggering the error, which the user might not be able to
do (e.g., "file system not available"). Yet, this would violate the
above success criterion.

Unless a suitable restriction can be worked out, this criterion is not
applicable to all Web content and should be moved to level 3.

 > 3.	Blocks of repeated material, such as navigation menus with 8 or
 > more links, are marked up so that they can be bypassed by people who use
 > assistive technology or who navigate via keyboard or keyboard interface.
 > [V] 
 > 
 > 	Editor's note:General Techniques might include something about
 > satisfying this criterion through metadata, use of a future role
 > attribute, etc.
 > 

What's the rationale for the "8 or more" threshold?

 > 
 > Guideline 2.4 L3
 > 
 > 1.	When content is arranged in a sequence that affects its meaning,
 > that sequence can be determined programmatically. 
 > 
 > 	Editor's note:This is an improvement upon the wording proposed
 > on 27 October, but it still needs work. The problem is how to specify
 > that it applies to content that is intended to appear as a sequence
 > without requiring a test for intention.

I don't understand what this intended to achieve beyond guideline 1.3
or the proposed 2.4 level 1. Under what conditions is 1.3 level 1
satisfied and this success criterion not?
 > 
 > 	Editor's note:The criterion about reading order would probably
 > be more appropriate under Principle 3: it could be argued that reading
 > order is irrelevant unless it affects the user's ability to understand
 > the content. This in itself is not necessarily an accessibility issue.
 > It becomes an accessibility issue if a user with a disability (such as a
 > visual or cognitive impairment) could not reliably derive a meaningful
 > reading order from the default presentation. If we want to retain this
 > criterion and keep it under 2.4, we need to craft wording that ties
 > reading order to users' ability to operate and use the content.

I still think reading order should not be treated separately from
structure/relationships and that if the former are exposed to the user
agent, the latter can be made reasonable in the generated presentation.
 > 
 > 2.	The tab order follows relationships and sequences in the
 > content. 

Can you think of something less keyboard-specific than "tab order"?
Perhaps "the order in which elements receive focus when navigated sequentially"?

 > 3.	Images have structure that users can access. 

Am I correct in interpreting this in conjunction with 1.3 as implying
that if I am writing in SVG, I need to expose the structure for level
1 conformance; and if I want to conform to level 3 of WCAG 2.0 I need
to use SVG or another image format that supports structure? If this is
the correct interpretation it is fair enough, but it is important that
authors using SVG understand their obligation to expose structure
under 1.3 and that they don't misinterpret this success criterion as
an escape route from that requirement.

 > 4.	Perceivable units have descriptive titles 
I need the definitions of "perceived unit" and "delivery unit" before commenting.
 > 5.	Text is divided into paragraphs. 

If I write a two-sentence document, does this mean it needs to be
enclosed in some sort of paragraph element? What happens if my markup
language doesn't support paragraphs, e.g., the text is inside an SVG
DESC element? How does this apply to user interfaces?
One answer might be that it is at level 3 for exactly these reasons
and we shouldn't worry about all of the circumstances in which it is inapplicable. 

 > 6.	Documents are divided into hierarchical sections and subsections
 > that have descriptive titles. 
 > 
 > </proposed> <current> Guideline 2.4 Facilitate the ability of users to

I suggest removing as many of the editorial notes as possible before
incorporating this into a draft. The notes just distract the reader
from the text of the success criteria, and most are not questions to
reviewers of the sort that would be left in a public working draft.
Received on Thursday, 4 November 2004 02:24:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:32 GMT