WebAIM - Web Accessibility In Mind

E-mail List Archives

PDF is only partially accessible

for

From: Wayne Dick
Date: Dec 3, 2009 3:12PM

Accessibility of PDF

Based on many of the responses from the WebAIM list, as well as from my
own research, I conclude that well structured and tagged PDF is
partially but not completely accessible. It excludes people with
moderate low vision (20/70-20/160) from its accessibility supported
group. See the American Optometric Association article on low vision for
a complete classification of visual impairment from near normal vision
to total blindness, http://www.aoa.org/low-vision.xml . I came to this
conclusion with regret, because Adobe has been a model of corporate
responsibility regarding accessibility support for several years.
Still, I think PDF needs to go one more step to be classified as
accessible without qualification.

The group with moderate low vision needs a restyled page that can be
read with their eyes. Alternative text is not useful for most of us.
We can navigate documents with visual cues. Most of us use and like
mice. In fact we can read and use almost everything but the body text.
Tools like JAWS are an annoyance because unnecessary sound distracts
us from the hard job of seeing with our limited vision. Zoom products
are great for people with severe low vision, but they disrupt the space
so much that we lose our visual ability to navigate using normal visual
tools. What we need is enhanced text. That means size control that is
not uniform for all tagged text types, e.g. paragraphs should be
enlarged more than headings. We need clean and simple font families
like Arial, Verdana or Tahoma. We also need spacing control over
letters, words and lines. PDF does not seem to be able to provide these
transformations.

This is a direct deficit according to WCAG 2.0 Criteria 1.3.1. The less
modern criteria of 508 1194.22(d) may let this slip by, but the intent
of Section 504 and the ADA would consider this a deficit because it
excludes a well defined disability group. Here is my analysis. Well
structured and tagged PDF is not perceivable for people with moderate
low vision at a level that facilitates effective use of print media for
acquiring knowledge. This is because critical parts of well structured
and tagged PDF documents are not adaptable to the needs of people with
moderate low vision. Because WCAG 2.0 gives a direct analysis of why
this is a problem I will use this standard. PDF fails to support
Guideline 1.3, Criteria 1.3.1, Sufficient Technique G140.
Adaptable:

Guideline 1.3 Create content that can be presented in different ways
(for example simpler layout ) without losing information or structure.
Understanding Guideline 1.3 . PDF fails to support this Guideline
because it fails to support 1.3.1 completely at the text level.
Info and Relationships:

Criterion 1.3.1 Information, structure, and relationships conveyed
through presentation can be programmatically determined or are available
in text. . PDF Fails this Criterion because text objects in PDF cannot
be restyled visually an effective format for people with moderate low
vision.

Sufficient Technique G140: Separating information and structure from
presentation to enable different presentations . -- Not supported
completely by PDF.

The text contents within paragraphs, lists and table cells appear to be
immutable regarding the style properties: font family, spacing of
letters, words and lines. That means the user cannot adjust necessary
presentation style properties to support reading. This interferes with
the person’s ability to separate figure from ground.

Serif fonts create a complex visual environment. Visual complexity is
the enemy of people with low vision. Generally, people with moderate
low vision benefit from a simple font line Arial, Helvetica and Tahoma.
Choice of font family allows this group to use their limited sight
effectively.
Closely packed letters make reading long words difficult. The middle
letters blur out. Closely packed words make separation of words
difficult. Words can get blurred together and become unreadable, or
they can disappear from sight. Closely packed lines make tracking
difficult. One wanders from line to line in the middle of lines.

Once again complexity is the problem. Separating text objects without
sufficient spacing creates a difficult complex environment. Often
publishers of large print books skimp on spacing and destroy the
effectiveness of large print.

Variable sizing is also critical. In a large print world with increased
spacing, screen space is precious. Enlarging headings by the same
factor as paragraph print is wasteful. There are many other cues that
can designate the relative importance of text that do not require size.
Borders, color for people who can see it, emphasis, strength and font
style can all be used as visual level markers that can be used by this
group and does not waste screen space.
Accommodation vs. Accessibility

There are accommodations and alternative media than can give a
significantly inferior experience for this group. This is a well known
fact because HTML and most Word Processing formats (doc, docx, rtf, odt)
can product the exact formats that respect these needs of moderate low
vision completely. So, objectively, PDF is not as accessible as word
processing formats or HTML. From my point of view, life would be much
better off if every posting in PDF was also posted in one of the
completely accessible formats. Accessibility for reading documents
means that the user with a disability can use a program to produce a
format for reading that is just as usable as the format developed for
the primary audience. Anything less is not completely accessible. That
is why PDF is only partially accessible. People with moderate low
vision cannot get an equally effective experience when they are forced
to read PDF as their only choice. Unfortunately, large institutions
have jumped on the notion that PDF can be made completely accessible,
and the presence of stand-alone PDF postings is accelerating. This is a
sad occurrence for those of us whose quality is directly reduced by this
trend.
What can Adobe and PDF Users Do

Adobe needs to address this problem. They have addressed tagging,
reading out loud, reflow, color, tables and most other problems. They
can take one more step.

Until this issue is resolved to the satisfaction of people with low
vision, Adobe should not encourage the belief that their product can be
made totally accessible.
What can institutions do?

Until Adobe completes its work with PDF to provide accessibility for all
people with low vision, institutions should not engage in postings of
stand-alone PDF. All postings of PDF documents should be accompanied by
documents in a completely accessible format. Today we know these are
obtained from word processors and W3C mark-up languages. TeX, LaTeX and
MathML are good for mathematics. PDF is completely inaccessible for
STEM publications.

Wayne Dick PHD.
California State University, Long Beach