|Home | About | Journals | Submit | Contact Us | Français|
The American Physical Society (APS) has been a leader in using markup languages for publishing. ViewPlus has led development of innovative technologies for graphical information accessibility by people with print disabilities. APS, ViewPlus, and other collaborators in the Enhanced Reading Project are working together to develop the necessary technology and infrastructure for APS to publish its journals in the DAISY (Digital Accessible Information SYstem) eXtended Markup Language (XML) format, in which all text, math, and figures would be accessible to people who are blind or have other print disabilities. The first APS DAISY XML publications are targeted for late 2010.
Computers have revolutionized information accessibility for people who are blind or have other serious print disabilities. When information was distributed primarily on paper, it was necessary for a sighted human being to intervene in order to make that information available to people with print disabilities (PWPD). Information could be read directly to PWPD, taperecorded for them, or converted into braille and, for detailed graphical information, into tactile graphics with braille labels. Such intervention is no longer necessary for a wide variety of information that can be read by PWPD on computers using speech or braille screen readers. Email, web textual information, and many other types of electronic literature are now directly readable by PWPD.
Unfortunately there is still a vast amount of literature that is distributed electronically but that cannot be read adequately with screen readers. Very little scientific literature is fully understandable because the mathematics and critical graphical information are presently not directly accessible by any non-visual means. The new Enhanced Reading project has the goal of making a number of professional physics publications available in the DAISY1 (Digital Accessible Information SYstem) eXtended Markup Language (XML) format that makes all text, math, and figures fully accessible by PWPD. This development will blaze a trail that will make it possible for many other publishers easily to make their publications available in accessible DAISY XML format.
information is spoken in semantically meaningful strings
IBM was one of the pioneers in the use of markup languages for technical publishing. In the 1980s IBM began standardizing the composition of its technical manuals in all IBM manufacturing plants and laboratories, using a generalized markup language (GML). The goal was to preserve the original files in an electronic format that permitted easy maintenance and reusability. An offshoot of this effort was a research project to develop methods for reading, searching, and navigating such documents on computers. The result of that research was the IBM Book Manager computer application. By 1987 IBM had more than 3,000 manuals available in Book Manager format, along with support services utilizing the GML technology.
In the late 1980s IBM developed a speech screen reader as a hardware PC attachment that enabled blind programmers to navigate and read IBM Manuals. IBM was surprised when many non-blind programmers requested and used the screen reader, explaining that they could read and comprehend better when they could hear as well as see the words. Today it is well understood that many sighted people benefit from hearing as well as seeing words. This is almost always true for people who are known to be dyslexic, but it is also true for many who are not.
One of the authors (R. Kelly) was a leader of the IBM developments and in 1993 moved to the American Physical Society (APS) to lead their journal modernization. APS is a major publisher of scholarly physics journals. Its Physical Review journals and Physical Review Letters (PRL) are generally regarded as the most important physics research journals in the world, between them publishing more than 18,000 articles each year, out of approximately 36,000 articles submitted for review, by authors from nearly every country on earth. In 2008, roughly one-third of the corresponding authors were from the United States and Canada, slightly more than one-third from Europe and slightly less than one-third from Asia. APS also publishes a number of other specialized physics research and physics teaching research journals.
APS has been a leader among scholarly publishers in modernizing its publishing procedures. In 1994 APS started experimenting with Standardized General Markup Language (SGML), in part to facilitate moving PRL online in PDF form. PRL went online in 1995, followed by all the rest of the APS journals; by 1997, all were online. In 2004 APS began using an XML workflow for its journals. Once a submission has been accepted for publication, text and math is converted to XML. The XML file is then used to create the article for printing.
The XML source file can be repurposed for many other uses, including online offerings and Web 2.0 initiatives. In 2007 APS and ViewPlus began a collaboration to repurpose APS XML to enable publication of accessible online journal articles. Transforming APS XML to the accepted accessible DAISY XML format was easy. Making APS Encapsulated PostScript (EPS) figures accessible was the major challenge, but it was one for which ViewPlus was prepared. ViewPlus demonstrated that figures can not only be accessible but can also contain a great deal of metadata that greatly expands their usefulness. It quickly became apparent that DAISY XML publications, read with a good online reader application, could provide greatly enhanced access and information for sighted physicists as well as for PWPD.
the audio/touch method has been well tested and found to provide excellent accessibility
Research on graphics accessibility conducted in the Science Access Project at Oregon State University2–4 first posed the question as to whether it was possible, even in principle, to create mainstream graphics in a form that could be automatically accessible to PWPD. It was clear that well-designed viewing software and additional hardware would be needed even if the mainstream graphical information was in an ideal form. The researchers were strongly influenced by the pioneering work of Parkes,5,6 who had introduced the audio/touch technique in 1988.
The audio/touch method has been well tested and found to provide excellent accessibility. The user has a tactile copy, typically mounted on a touch tablet attached to a computer. When a text label on the graphic is touched, the label is spoken by the computer. When a graphical object is touched, the title of that object can be spoken. Audio/touch software typically permits the developer to incorporate a number of levels of information so that any graphic can be made extremely accessible. The American Printing House for the Blind7 and several other agencies used Parkes’s method to create audio/touch educational graphical materials for the blind. These were quite expensive, owing partly to the need for special hardware and special computer applications, and to the difficulty and expense at the time of making the tactile copies.
Introduction by ViewPlus of the Tiger embossing technology in 20008,9 solved the problem of making a usable tactile copy from mainstream graphical information. When the World Wide Web Consortium developed the Scalable Vector Graphics (SVG)10 language, the Oregon State research group immediately recognized that SVG was the key to making mainstream graphics that could also be accessible. SVG is an XML object-oriented graphics language in which labels and description properties are permitted for every object. When objects are meaningful and identified by their labels, then the SVG file can be accessible to all readers, including those with print disabilities. SVG text includes the Unicode characters, not just images of text. Consequently SVG text can also be accessible by non-visual techniques.
SVG was the key to making mainstream graphics that could also be accessible
The research project was taken up by ViewPlus after that company was spun off from the Science Access Project. Its IVEO SVG Viewer11–16 was introduced in 2005. The authoring/conversion application IVEO Creator Version 1.0 was also introduced commercially in 2005, and Version 2 was introduced in 2008. The Creator and more recent Creator Pro applications can be used to author simple diagrams and to convert virtually any electronic format to SVG. They also permit authors to enrich the SVG image with additional metadata, including image description, title and descriptions both for the figure and for individual graphical objects in it. SVG text is more or less automatically accessible.
These SVG files can be displayed visually with any SVG viewer. PWPD can use the IVEO Viewer, which can be downloaded free from the ViewPlus website17 to view by audio/touch. The image may be printed from IVEO Viewer to a ViewPlus embosser to obtain a tactile copy. When the tactile copy is placed on a touch pad, the user feels graphical objects and text (which has a distinctive tactile feel even though it is generally not readable with tactile sense alone). When text is pressed, it speaks. When a graphical object is pressed, its title is spoken. A graphical object may also have a description field (of any length), which can be read by pressing the appropriate menu item or hot key. The audio information can be browsed if the text is long, and it can also be delivered to an online Braille display, attached to the computer. The Braille display makes the information accessible to people who read Braille, including those who are deaf and blind. Spoken text is reproduced in a status bar whose font size is user-selectable, a great help for many people with low vision.
Figure 1 shows the IVEO Viewer being used by a blind reader. An IVEO SVG example is shown in Figure 2. Figure 2(a) shows the original figure that is seen on screen and printed on a color tactile image. Figure 2(b) shows the embossed image, where dark squares represent high dots and lighter squares represent lower dots. The ragged embossed text patterns seen in Figure 2(b) are generally easy to discern by touch as text. When pressed, the text will be spoken. Text is organized into semantically meaningful spans in the file. When any text element in a span is pressed, the entire span is spoken, so that information is spoken in semantically meaningful strings. Some simple diagrams and flowcharts in math, physics, and computer science are fully accessible if nothing other than text labels is accessible. However, most graphics are more accessible when important graphics are labeled and when the graphic has a good description.
DAISY1 is an organization of non-profit agencies for the blind and dyslexic. It was formed to develop international standards, infrastructure, software tools, and training documentation to promote information accessibility by PWPD. DAISY initially developed an XML format that is rich enough to convey most non-scientific literature, and it is now a subset of the mainstream open e-pub XML format.
The original DAISY XML specification did not include any capability for mathematics or for accessible graphics, beyond showing them as images with ‘alt text’ labels. A MathML Working Group developed an extension to permit MathML to be included in DAISY XML, and their recommendations were adopted in 2007. There are presently several methods for displaying MathML non-visually, including several commercial DAISY book reader applications18,19 and the Design Science20 MathPlayer plug-in that displays MathML in Internet Explorer. MathPlayer is accessible to speech screen readers.
After ViewPlus demonstrated with its IVEO SVG technology that SVG graphics could be excellently accessible, a DAISY SVG Working Group was formed to develop DAISY guidelines and to extend SVG specifications; the authors of this paper are members of this working group. The working group is building on the success of ViewPlus IVEO developments and intends to make SVG a high-quality format for mainstream information accessibility as well as for PWPD.
The DAISY Working Group is developing guidelines on such things as how to describe figures for the figure description field. It is also extending SVG by defining a number of attributes for special purposes and by adding DAISY namespaces to SVG. Examples of useful features that are possible with special attributes are the capability of making some objects visible on the printed copy but not embossed, or objects that are embossed but not printed in ink. There are many instances were it is convenient for the author to create tactile images that are not well represented by standard embossing rules relating color or intensity to embossed dot height. A portrait is a good example. By default a white nose would be embossed with low dots whereas the blind user would expect a nose to protrude from the paper.
tone graphs give semi-quantitative intuitive information by mapping frequency to value
Additional namespaces will be used to hold the numerical data displayed in graphs, both x–y graphs and two-dimensional displays such as those seen in common Geographic Information System (GIS) displays. Having data stored within the image file gives a major improvement in searchability, since sophisticated search algorithms could carry out filtered searches, making it interesting for many purposes. Having numerical data within the image file also improves its accessibility for blind people, since users can display data in numerical form or as audio tone graphs. Tone graphs give semi-quantitative intuitive information by mapping frequency to value. Normally intensity rises as the value becomes more positive and decreases as the value becomes more negative. Audio tone graphing provides good accessibility for both x–y graphs21 and two-dimensional graphics.22
The aim of the initial phase of the APS ViewPlus collaboration was to demonstrate proof of principle by transforming PRL articles to DAISY XML, with figures in accessible SVG form. This was done, and these prototype articles were demonstrated at a number of publisher meetings in 2008 and 2009.23 APS then invited several other organizations to join it in the newly formed Enhanced Reading Project. Presently that group includes the math software company Design Science, Inc.,24 the archiving company Portico,25 and the composition vendors American Institute of Physics (AIP),26 Aptara Corp.,27 and Beacon PMG.28 AIP is itself also a major publisher dedicated to publishing in DAISY format. The long-range goal of the Enhanced Reading Project is to improve the publication process, enabling DAISY XML publication at no increase in cost, and to provide an excellent reading experience for every reader, whether printdisabled or not.
DAISY XML and well-designed online reader applications can provide major advantages to all readers. For example, articles can be displayed on computers and mobile reading devices so that font size and contrast can be easily adjusted. The reader can choose to browse visually and in speech, or just to listen to all or part of the paper being read aloud. He or she can copy-and-paste MathML expressions into other applications, such as Wikipedia and math computation programs. Figures can be zoomed and portions of figures selectively displayed. Underlying data in figures can be displayed or transported into other applications. Rich descriptions within the figure files can be displayed or heard in speech. Papers can be linked online to references or to community networks for discussions and criticism. Most of these advantages are not possible with the PDF versions that are presently available, and certainly not with paper copy. The Enhanced Reading Project hopes to develop such advanced reading capabilities and to explore future enhancements made possible by XML, many of which may be as unforeseen today as the Internet was 20 years ago.
The APS/ViewPlus first phase project was funded in part by a Small Business Innovation Research grant, Number R43EY018799, from the National Eye Institute.