Content/Structure Separation

Home > PDF/UA Home > Working Draft > Breadcrumb placeholder text

Reference WCAG 2.0 Guideline 1.3

1.3.1

Informative Text
Relevant definitions:
"programmatically determined means that the specific value can be determined in a
standard, machine or software readable form."
"Programmatically located means that the value can be found, though there may be multiple
possible values (for example, providing a dictionary definition for a word that has
multiple meanings). This term contrasts with programmatically determined, where the
specific value can be determined."
"structure = 1.he way the parts of an authored unit are organized in relation to each
other and; 2.he way a collection of authored units is organized in relation to a delivery
unit and; 3.he way a collection of delivery units is organized."
"Examples: 1. book is divided into chapters, paragraphs, lists, etc.
Chapter titles help the reader anticipate the meaning of the following paragraphs. Lists
clearly indicate separate, yet related ideas. All of these divisions help the reader
anticipate changes in context. 2.
bicycle
is divided into wheels and a frame. Further, a wheel is divided into a tire and a rim. In
an image of the bicycle, one group of circles and lines becomes "wheel" while another
group becomes "frame.""
"authored unit = Some set of material created as a single entity by an author. Examples
include a collection of markup, a style sheet, and a media resource, such as an image or
audio clip."
"delivery unit = A set of material transferred between two cooperating web programs as the
response to a single HTTP request."
Analysis:
As distinguished from "programmatically located," the definition of "programmatically
determined" conveys unambiguous specificity. Since this WCAG guideline applies the term
to "structures," the intent appears to be that assistive technology does not have to guess
about content structure, or use heuristics, but can obtain authoritative and precise
information about it. The example of a bicycle image suggests that even images can have a
structure with labeled parts. To require labeling of not only an image as a whole but
multiple parts of it, however, seems beyond normal expectations for accessible web
content, so that example may be misleading.
Note that the guideline uses the plural "structures" whereas only the singular
"structure" is defined. Should the guideline use the singular form instead?
Also note that the guideline does not explicitly refer to all structures within the
content. Does this mean that some, but not all, of the structures need to be determinable
in a programmatic manner?
The structure of content is often hierarchical in nature, e.g., chapters, sections, and
subsections of a book. Might PDF content also have a non-hierarchical structure? If so,
should that structure also be subject to this guideline?
To meet this guideline for PDF content, tagging seems necessary (not surprisingly).
Tagging marks parts of content with structural roles.
The
set of tags is organized in a hierarchical tree. This tree defines a logical reading
order for the document.
Knowledge of structure benefits both comprehension and navigation.
Comprehension of content is aided by an understanding of how different ideas are sequenced
and subordinated to one another. Navigation of content is facilitated by boundaries and
levels among parts of the document, enabling assistive technology to jump either to
particular parts or by particular increments.
Some structure may be arbitrary, but still necessary. For example, a sidebar may be
arbitrarily placed between two paragraphs in the tag tree.
If not placed somewhere, assistive technology may never find it.
Minimal tagging requirements might include the following points. All textual content must
be tagged. All significant, non-decorative, non-artifact images must be tagged. Word,
line, and paragraph terminating characters must be explicitly drawn. All headings and
subheadings must be tagged and numbered hierarchically. Long documents must have a table
of contents with hyperlinks into the body. The tag tree must be representable as wellformed
XML validated against a schema defined by the PDF/UA specification. Adobe Reader
implements an API for accessing the DOM of a tag tree, but presumably, other APIs for
determining structures programmatically could be developed by third parties.

1.3.2

When information is conveyed by color, the color can be programmatically determined or the information is also conveyed through another means that does not depend on the user's ability to differentiate colors.

Comments on the Reference (Success Criteria)

Accessibility Issue

If meaning depends on color (e.g. push the yellow button, or required steps are in red),
then users show cannot perceive color will not be able to understand the content.

PDF/UA Implications

1. Review spec to identify where the color of content can be identified reliably. (I know
   
2. that text color can be identified. Need to review other content.)
   
3. Advise authors not to depend on color for meaning any place else?
   

Requirement(s)

Instruction to Authors

Informative Text