SGML: PDF AND ADOBE ACROBAT VIEWERS
------------------------------
Article: 9788 of comp.text.sgml
Newsgroups: comp.text.sgml
From: Liz McQuarrie <lmcquarr@adobe.com>
Subject: PDF and the Visually Disabled
Message-ID: <1995Jul6.145312.22490@adobe.com>
Sender: usenet@adobe.com (USENET NEWS)
Organization: Adobe Systems Incorporated
Date: Thu, 6 Jul 1995 14:53:12 GMT
There have been discussions on this newsgroup
about issues regarding the accessibility of
PDF for the visually disabled. The attached
document outlines the issues and Adobe's
plans for making PDF accessible. (This document
was also posted to comp.text.pdf.)
-----------------------------------------------------
THE ACCESSIBILITY OF PDF AND ADOBE ACROBAT VIEWERS
FOR THE VISUALLY DISABLED
June 30, 1995
Liz McQuarrie
Adobe Systems Incorporated
Adobe's Portable Document Format, the native file format of the Adobe Acrobat
products, is a final form description language for documents that is not tied
to any operating system or application. PDF provides the document layout
richness of Adobe PostScript and allows publishers to retain the look and
feel of their publication. On the World Wide Web, PDF is becoming
increasingly popular for documents that need the layout richness that
HTML currently does not provide. Corporations are also using PDF to
disseminate electronic documents over corporate networks, via e-mail, or on
CD-ROM.
For the visually disabled, however, there are currently some accessibility
issues associated with PDF and the use of Adobe Acrobat viewers (Reader and
Exchange) for viewing PDF files. This document describes Adobe's plans for
making both the Adobe Acrobat viewing products and the PDF file format
accessible for the visually disabled.
Overview of Acrobat and PDF Accessibility Plans
Adobe has the following plans to make PDF and the Adobe Acrobat products
accessible to the visually disabled:
Offer an Accessibility Plug-in for Adobe Acrobat on Microsoft
Windows (available in late 1995). The plug-in will allow users to access
Acrobat for Windows through cooperation with Windows screen-reading
programs, as well as output formatted ASCII that is compatible with
alternative output devices, such as Braille printers.
Enhance the PDF format and Adobe Acrobat for Logical Document Structure
(available in 1996). Adobe currently has an active project under way to
add knowledge of a document's logical structure to the PDF file format and
to build solutions based on these extensions. For visually disabled
users, the PDF structure project will deliver several accessibility
solutions.
Adobe Acrobat Viewers and Microsoft Windows Screen
Readers
Screen-reader programs allow the visually disabled to interact with the
Windows interface by interpreting what is happening on the display and
outputting that information to speech synthesis or refreshable Braille output
devices. Although screen-reader programs for DOS have been around for many
years, screen-reader programs for Microsoft Windows that provide reasonable
performance and functionality have only been introduced in the last year
or so.
Limited Screen Model and PDF Documents
The delay in the availability of Windows screen readers is due to a number of
factors. First, it is quite an intellectual challenge to present what is
happening on the Windows display to a visually disabled person in a way that
allows the user to be productive. This is because with Windows (or any
other GUI-based operating system), there may be multiple applications open
simultaneously, each of which has a title bar, menus and several child
windows for their data. Each application also has many visual controls to
present: icons, scroll bars, and dialog boxes that include check boxes,
buttons and list boxes.
To keep track of what is happening on the screen, Windows screen readers build
what is referred to as an off-screen model in memory, and then allow the user
to navigate through the screen model. For example, from the screen reader,
the user can select a particular application window to interact with. Then
once an application is selected, the user can navigate and read text in the
window. Reading text usually involves using commands like "Next Line,"
which reads the next line down the page, and "Previous Word," which reads the
word immediately to the left of the current word.
The current screen models for Windows for handling the text that is in an
application-specific window are very similar to the models developed for DOS.
These models can typically only represent text in simple documents, where the
document contains a single column of text that flows left to right and top
to bottom on the page. They cannot properly represent documents with more
complex layout, such as a multi-column newspaper or documents with text
that is not horizontal, like a tax form that contains vertical labels.
When the screen-reader program encounters such complex layout in an
application, it will likely make mistakes. For example, in a multi-column
document, the screen reader would probably read the first line in the first
column, followed by the first line in the second column, running the columns
together as if they were on the same line. Because many PDF documents have
such rich layout , the use of Windows screen readers with Acrobat on these
documents will not produce useful output for the visually disabled person.
For the future, Adobe is participating in discussions with Microsoft and
Windows screen reader vendors regarding a new Application Programming
Interface to screen readers in Windows that hopefully will more fully
support complex documents.
Conflicts Between the Acrobat Viewers and the Screen
Readers
The second factor that has delayed development of capable Windows screen-
reader programs is that there are no easy operating system "hooks" (or entry
points) in the current operating system (Windows 3.1) for developers to use to
build the screen readers. (Note that Microsoft is currently adding such
hooks to Windows 95.) The net result of this oversight is that developers
of Windows screen readers have each independently invented techniques for
capturing screen state for alternative output devices in often incompatible
ways.
Additionally, the Windows version of the Adobe Acrobat viewer really pushes
the Windows environment technically in order to render visually beautiful
documents with fonts that are very close to the author's original fonts.
This fact often causes the Acrobat viewer products to be incompatible with
many of the screen readers. Screen readers are not able to read any text
in the Acrobat document windows, regardless of the layout.
The Adobe Acrobat Solution for Windows for the Visually
Disabled
To make PDF documents accessible to visually disabled users of Microsoft
Windows, Adobe is developing a special plug-in for Acrobat. (A "plug-in" is
a piece of software that works with the standard Acrobat product and enhances
its capability.) This plug-in is called the "Accessibility Plug-in".
The Accessibility Plug-in for Acrobat will present an alternative view of the
open document in a separate window. This view will contain the text in as
close to reading order as possible, including presenting multi-column
documents as a single column, and "straightening" all non-horizontal text.
This alternative view will use only those text attributes that screen readers
currently understand, such as bold and italic.
The Accessibility Plug-in will interact with the primary Acrobat document
display in such a way as to make interaction as seamless as possible and will
allow the use of all standard Acrobat features, such as access to annotations
and hypertext links. Additionally, the Accessibility Plug-in will allow
users to export from PDF to ICADD-compliant ASCII, as well as to formatted
ASCII. ICADD (International Committee for Accessible Document Design) has
proposed using a set of tags based on SGML (Standard Generalized Markup
Language) with documents to enable them to work with alternative output
devices, such as Braille printers.
A Beta version of the Accessibility Plug-in for Windows will be available in
early fall, and will be distributed free of charge through a number of
channels, including the World Wide Web. The Accessibility Plug-in will work
with both the Acrobat Reader product, which is free, as well as Acrobat
Exchange.
Adobe Acrobat and Other Platforms
There are currently two other platforms that have screen reading programs
available: DOS and the Macintosh. (Screen readers for X-Windows systems on
UNIX are still under development.) DOS screen readers only work with the text
display of DOS. A third-party organization is currently contracting out the
development of a Acrobat Viewer for DOS that will work in text mode. The DOS
viewer will be available in the public domain. Note that this viewer is text
mode only, and is not related to the Adobe Acrobat 1.0 Viewer for DOS, which
is a graphics mode viewer.
The Apple Macintosh has not been a popular platform for the visually disabled.
This is probably due to the fact that many of the screen reader software
vendors started working with DOS, and then grew into Windows when it became
available. There is, therefore, only one screen reading program available
for the Mac: outSpoken from Berkeley Systems. Unfortunately, outSpoken for
the Mac currently has a bug when operated with Acrobat that causes Acrobat
to exit when the user attempts to read the document. Berkeley Systems
intends to fix this bug in their next release. When an updated outSpoken,
however, is available, it will also have problems with complex documents that
contain multiple columns and/or rotated text. Plans for an Accessibility
Plug-in for Acrobat on the Macintosh are still being formulated.
Extensions to PDF for Document Logical Structure
The second area of interest regarding PDF and the visually disabled concerns
document logical structure. ("Logical Structure" refers to the organization
of the document, such as title page, chapters, sections, subsections.)
Although standard ASCII may be adequate for disabled access to short
documents, access to longer, more complex documents is greatly enhanced by
an understanding of the document's logical structure. Many advocates for the
visually disabled feel that documents based on SGML are best suited for
accessibility because of SGML's strength in representing document logical
structure in a portable way. Documents based on SGML, however, continue to
be a small percentage of those that a visually disabled person may encounter
on a day-to-day basis.
Adobe currently has an active project under way to add knowledge of a
document's logical structure to the PDF file. This project will benefit all
Acrobat users -- there are several important problems that can be solved when
the logical structure knowledge of the document is available. An example:
the ability to search for documents that contain words in specific chapters
or headings. For visually disabled users, the PDF Structure project will
deliver several accessibility solutions, which are described below.
Document Structure Knowledge Leads to a Better Accessibility Plug-in
The Accessibility Plug-in to the Acrobat Reader will be greatly enhanced by
knowledge of the document's logical structure. To begin with, by knowing a
document's logical structure, the Accessibility Plug-in will be able to do a
much better job of displaying the document in reading order. Consider, for
example, a PDF document that is a newspaper, where the first page contains
the beginnings of three articles, each of which is continued on a separate
page. With a knowledge of the document structure, the Accessibility Plug-in
will be able to follow a single article from beginning to end, much like a
sighted person would flip the newspaper page to continue the article until
complete.
Second, many navigational features for both sighted and non-sighted
individuals can take advantage of the document's logical structure. An
example: next and previous section links.
Finally, the plug-in can do an even better job of exporting to an
ICADD-compliant tagged ASCII file by supporting a richer set of ICADD tags
(ICADD ASCII can then be used with alternative output devices).
Availability of Structure in PDF
A shipping date for products from Adobe that include document logical
structure in PDF is currently scheduled for calendar year 1996.
Comments or Questions?
Liz McQuarrie
Adobe Systems Incorporated
lmcquarr@adobe.com
415-962-3915