IBM comments on WCAG 2.0, July 30, 2004 Public Draft

Section "Conformance"
   This section needs to describe how conformance claims can be made for
   sites that aggregate content from other sources.
   Web sites do not have a single URI that identifies all of the site,
   rather pages are identified by URI and not all pages can be referenced
   by a unique URI.

Section "Overview of design principles"
   Principle 4 - "Content must be robust enough to work with current and
   future technologies." Working with future technologies is an impossible
   bar to meet.

Guideline 1.1
   SC Level 1 - if a technology doesn't support explicitly associating a
   text alternative with non-text content, it should still be conforming to
   provide a text equivalent another way.
   SC Level 1 - suggest modifying "Text-alternatives are explicitly
   associated with non-text content..." to "Each instance of non-text
   content has at least one text alternative that is explicitly associated
   it...."
   SC Level 1, a and b.  For an image button containing text, the alt text
   should match the text in the image. SC 1a says that for graphical
   buttons, the text alternative should describe the purpose or function of
   the button. Does alt text that matches the text in the image button meet
   this success criteria?
   SC Level 1 "how to" link: <noembed> is not widely recognized.
   (
   http://www.w3.org/TR/2004/WD-WCAG20-HTML-TECHS-20040730/Overview.html#noembed
   )
   We need a success criterion that deals with non-text content that
   requires a signficant amount of text content in order to be equivalent.
   For example, a complicated chart or graph may require a page of text or
   a large table as an equivalent text alternative. In HTML, you "can"
   provide a longdesc attribute on the img element for this. But,
   alternatively, because longdesc is inconsistently supported by screen
   readers or because you might want to make the detailed information
   available to everyone, you can simply describe the chart in the text on
   the page, add a text link next to the image that links to the detailed
   text equivalent, or make the chart itself an image link which links to
   the detailed text equivalent. This is certainly accessible but it would
   not pass the checkpoint as currently worded because it is the detailed
   description that is really the equivalent and it is not "explicitly
   associated" with the non-text content.
      Suggest modifying Level 1 to: "Each instance of non-text content has
      at least one text alternative of fewer than 150 characters that is
      explicitly associated it...."
      and adding another Level 1 success criteria: "Text alternatives of
      more than 150 characters are provided either inline or via an
      adjacent text link."
   Accepting this proposal will make examples 2, 3, and 4 valid. See
   comments below.
   The definition of non-text content excludes scripts, applets, and
   programmatic objects as they are covered under guideline 4. It should
   also exclude content rendered via plug-ins such as Flash and PDF.
   Although not as mature and well-supported as HTML, it is technically
   possible to create accessible Flash and PDF documents and it is expected
   that the technologies will continue to improve.
   Benefits - Users with cognitive or reading disabilities may actually
   prefer images to text in some cases (for example, an icon representing
   sports or a flag instead of a text link with a description and a state
   name).
   Example 2 and 3  - Short label + longer description.  There is nothing
   in 1.1  that would "require" both a short label and a longer
   description.
   Example 4 does not use an explicit association but an implict
   association using a "close" link. This makes sense and should be
   allowed, even encouraged in the guideline but currently is not.
   Example 4 refers to "after" without defining what "after" means,
   directly below the content, or following in reading order?  Would a
   blank line invalidate the "immediately after" requirement?

Guideline 1.2
   SC Level 1 would disallow video conferencing since few video providers
   have the capability to include real-time captions.
   SC Level 1 item 4 Exception - How do you decide if something is
   primariliy non-vocal?
   SC Level 1, item 6 - Need to define "respond interactively."

Guideline 1.3
   SC Level 1 - this is not possible in all cases on the web. Relationships
   between elements are not adequately supported. You may not know one
   object controls another that is in an entirely different part of the
   document hiearchy. For example, DIVs which are rendered as menus but
   which are stored at the end of the page and rendered as context menus
   for each node which launches the menu.
   SC Level 1 item 3 and Level 2 item 1 are confusing. Some people don't
   understand the difference and are also interpreting this to mean you
   can't use color to enhance usability. Suggest the following rewording:
      SC Level 1 item 3 - Color may be used as an enhancement, but not as
      the only way to convey information. Any information conveyed through
      color is also available either through additional context in the
      content or by programmatic interpretation of the coding.
      SC Level 2 item 1 - Color may be used as an enhancement, but not as
      the only way to convey information. Any information conveyed through
      color is also available through additional context in the content
   SC Level 1 item 3 - There is also confusion about whether or not this
   requirement applies to links; that is, visited links are displayed in a
   different color than non-visited links. But this situation should not be
   non-conforming. Users can customize the color of visited links either
   through the browser UI or through custom CSS.
   SC Level 2 - A specific common example of this is required form fields
   which are labeled in red text but also have an asterisk next to the
   label. This would be a good example to add to the guidelines or the
   Gateway document.
   Example 1 refers to missing information in forms but what about forms
   data which is invalid? XForms deals with this problem by providing a
   validate declarative function in forms. Recommend the group looks at the
   declarative events in XForms to make sure you have not missed anything.

Guideline 1.5
   This guideline is marked as a level 2 guideline, but there is no level 2
   success criteria.

Guideline 2.1
   SC Level 1 - this is confusing as worded. Some interpreted this as not
   allowing for device specific content for a device that does not have a
   keyboard. For example, a one-way communication system that pushes
   messages to users but does not require interaction. Suggest rewording as
   "When the content provides interactive functionality, all of the
   functionality is operable through a keyboard or keyboard interface where
   the functionality or its outcome can be described in a sentence."
   SC Level 1 and 3 - need to provide an example that demonstrates the
   difference betweent these two.
   SC Level 2 - define "the more abstract event".
   Benefits says that individuals with severe physical disabilities benefit
   because they can use speech input to enter data and manipulate
   information.  There are plenty of people who use speech recognition
   because of issues with using their hands but they don't consider
   themselves to have severe physical disabilities. Suggest: "Individuals
   who are unable to use their hands can use speech..."

Guideline 2.2
   SC Level 1 uses a time that is at least 10 times the original setting.
   10 times the default can still be a very small number. If the default is
   1 ms, the "accessible" time is 10ms, hardly an improvement for a motion
   impaired user. There should also be a minimum absolute time as well.
   Something like "adjust the time limit over a wide range which is at
   least ten times the length of the default setting and a minimum of 30
   seconds in duration".
   It is not clear how SC level 1 item 1  and level 3  item 1  differ.
   How does this guideline apply if there is no reading or interaction
   required, for example a server-side redirect (which includes an HTML
   page with most servers but users are not required to read this because
   the user agent connects to the new location automatically) or a
   client-side redirect depending on factors which the user cannot
   influence (for example the availability of a plugin)?
   The time expiration warning is an extreme change in context which
   another guideline does not allow. This seems to contradicting
   requirements.

Guideline 2.3
   General Flash Threshold - "rectangle meter" should be "square meter",
   the unit is cd x m^-2  (or more reliably written as cd/m/m).
   Spatial Pattern Thresholds - The description of "clearly discernible
   stripes" lacks a thickness dimension. Provide a thickness dimension so
   it is easier to visualize the type of pattern that causes problems.  As
   stated, one could ask whether tabular data with alternating row
   background colors (for legibility) falls into this category of
   troublesome pattern.  Where is the threshold?
   Spatial Pattern Thresholds - How is the pixel size relevant relative to
   screen resolution? Can a device with less than 1024x768 pixels cause
   photosensitive epileptic seizures?  Doesn't the impact depend on the
   size and distance of the screen and the absolute size, not a pixel
   value? Would a criteria definition that specifies the visual angle be
   more accurate than the current way that specifies a pixel rectangle when
   using a particular resolution? Expressing this as a percentage of screen
   space and giving an example would probably be less confusing.
   Editorial Note: The flicker test tool should be available according to
   the draft but no link is provided yet.

Guideline 2.4
   SC Level 2 item 2 - if skip links must be visible, this should be moved
   to a Level 3 success critera.
   SC Level 3 item 5 - Define what is meant by the "statement."
   SC Level 3 item 1b - Table of contents is very document centric and a
   site map is a concept that cannot be applied to portal sites where users
   can personalize the content. Suggest "table of contents (for pages),
   site map (for sites), or search function (for sites).
   There is a lot of confusion about the difference between guideline 2..4
   and 1.3. The Level 2 items seem to be about being able to get a high
   level view of the information or group it into categories. This sounds
   like "understanding". While the level 3 items seem to be about the
   exposing the structure independently of the presentation (guideline
   1.3). I suggest that we move 2.4 to principle 3 and reword the guideline
   to "Organize the content in a way that allows the user to understand the
   high level concepts or functions without having to read sequentially
   through all the detail." and then move all of the "coding" requirements
   in Level 3 to 1.3 Level 3. 1.3 is about making the structure available
   programmatically through the code. Reading order, accessing the
   structure of a diagram, and tab order seem to be about making sure the
   structure is available programmatically.

Guideline 2.5
   SC Level 3 item 2 - It may not be desirable to spell check terms and
   offer alternatives for all text input, for example for userids and
   passwords/passphrases, Web based content management systems, text in
   languages not understood by the Web application etc.  "Text" should be
   limited to long free-form text and the user should be able to disable
   spelling suggestions, I would hardly want to use a Web mail system that
   does not understand German and get helpful suggestions for each and
   every word :-)

Guideline 3.1
   SC Level 1 item 2 - Meta information and alt attributes are examples
   where acronyms cannot be expanded programmatically.
   SC Level 2 item 1. The meaning and pronunciations of all words in the
   content can be programmatically located. One could read this as
   suggesting publishing a full dictionary and linking every word of
   content to the dictionary. Is that the intent?
   SC Level 3 item 3 - define what is meant by the "statement."
   Example 1: The first example refers to "page title". This could be
   interpreted to mean the <title> element in HTML which allows only text
   context. The <acronym> element cannot be used in a <title> element.
   Suggest rewording the example as "An acronym in a heading" to eliminate
   the confusion.

Guideline 3.2
   SC Level 3 item 3. This is extremely expensive and impractical to
   implement.

Guideline 4.1 Examples
   Please publish all links to IBM Web sites as http://www.ibm.com/path,
   not http://www-3.ibm.com or any other server name. As an aside it would
   make the document more accessible if links were spelled out, so moving
   the mouse over the link was not required to see where the link points
   to.

Andi
andisnow@us.ibm.com
IBM Accessibility Center
http://www.ibm.com/able

Received on Friday, 10 September 2004 22:05:11 UTC