Ron Stewart, Director
Northwest Center for Technology Access
Oregon State University
Introduction
One of the most significant challenges facing disability service and access technology staff in higher education is how to provide equal and equivalent access to math and hard science materials for our students with print-related disabilities. For many years, truly effective tools have not been available to create and distribute fully accessible materials for any content that is visually complex or that relies on symbology as a primary means for conveying information.
For most of the last decade the staff and faculty from the Science Access Project, in cooperation with those of the Technology Access Program, at Oregon State University, have been conducting the research and development of a variety of technologies that focus specifically on access to mathematical and hard science content for the print disabled. The members of SAP work primarily from the research and development perspective and TAP focuses on the production and effective use of these technologies by their intended audience.
The outcomes of these activities have taken two paths. Those research projects that showed a commercial viability continue to be further developed and marketed by ViewPlus Technologies, the TIGER Graphical Embosser being the best known example. On the other side of things, projects that showed a strong potential for application in the teaching and learning environment, but little or no commercial viability, continue to be developed and disseminated by the Technology Access Program. WinTriangle is the current focus in this area by TAP, and will be a primary focus of this article. More information on WinTriangle can be found at http://tap.oregonstate.edu/WinTriangle/WinTriangle.htm.
The development and use of WinTriangle is just one viable technology for the creation of accessible math and science. Many others are just beginning to emerge, and their functionality and usability will only develop as they are widely adopted. This article will explore the process for creation and distribution of these materials, and will introduce the reader to several other ways in which to produce fully accessible content. This effort is still highly resource- and staff-intensive, but the outcome results in the ability of our students with print-related disabilities to be full participants alongside their peers in the exploration of careers in math and the hard sciences.
The E-Book Process
The creation of accessible math and science materials is best accomplished as a part of a comprehensive alternative format production operation, or E-Book production program as they are coming to be called. The ultimate goal of this effort is the creation of fully accessible materials for all content areas. The scale of this operation is really determined by the extent of the need, and may actually be most effective as part of a cooperative effort between several campuses, schools or higher education system. Viable models exist for both stand alone and cooperative efforts. Additionally a national effort for the development and dissemination of best practice-based programs in being coordinated by the Association of Higher Education and Disability, in cooperation with textbook publishers and third party providers of E-Books such as RFB&D and Bookshare.
The first question we want to ask ourselves as we go down this road is, "what is our end product going to look like?" The answer is, "that will depend on the needs of the student user." Will you be providing an Audio Book, with or without navigation, or a Braille-ready full-text E-Book. These outcomes are at opposite ends of the development spectrum, but a fully-developed production program will create source files that can be delivered as either. In the long term we will be able to easily obtain the electronic source files for almost any book from its publisher, but that is a few years down the road. What is currently extremely uncertain is whether we will be able to obtain the math and science content in an easily translatable format; the odds are, probably not for many years to come.
There are a lot of other questions we will need to deal with as we continue with this process, and once again the answers will depend on what the desired end product is, or in reality what the student is going to want to be able to do with the materials you are providing. This is also a resource question, as what you will be able to produce will to a great extent be determined by the level of staff and equipment resources you have access to. It is becoming very clear that books on tape or an unedited digital file such as an MP3 file are no longer acceptable accommodations, but this is still what the majority of programs are providing.
What file format are you going to use to develop and disseminate your materials? This has profound implications for what your users will be able to do with them. PDF files are probably fine for your users with learning disabilities, or those who are going to be using a book reader product like the K3000 or WYNN, but these same files will be totally unusable by your blind or low vision users who are utilizing a screen reader or downloading the files to a BrailleNote. The format in which you produce your graphics (such as charts or images) and formulae will also determine the suitability of the final product for use by the greatest number of students. Depending on the nature of the material, it may be best used as a tactile image, or as a tactile image with Braille tagging and annotations. Perhaps a description may be the best format, as we have found over the years, but then you also need to make sure that its relationship to the rest of the content is clearly indicated and is not provided as a contextless and therefore useless piece of trivia.
How much editing are you going to do once you have scanned the original text into the computer, and run it through an Optical Character Recognition (OCR) program such as Abbyy Finereader or OmniPage? Maybe you are going to do, as many have done in the last few years, and use the K3000 software to do all your document processing. This will be fine for your LD users, but once you convert these files to Rich Text Format (RTF) for your Blind users you will lose almost all structure and editing of the text. What your user will get is what I call a "stream of consciousness file": one long piece of textual information that is almost unusable as a learning tool. Use of files of this nature, while working well for your largest population of users -- those with cognitive processing issues -- are only effectively usable with a screen reader or turned into Braille after an extremely high level of editing.
Non-Textual Content
Now that you have been totally overwhelmed and are wondering what this has to with math and science access, let's delve into this in depth. First, I would like to look at the bigger topic of "Non-Textual Content", or in simpler terms anything that cannot be readily accessed as text by a screen reader or as refreshable or standard Braille. This includes, as I have already indicated, images and charts, but more importantly, math, scientific notation, and programming languages. The nature of the curriculum and the pedagogical intent of the instructor will need to be your guide for how this material will be dealt with.
For a lot of content that is primarily for the purpose of developing a conceptual understanding, an expanded word description is often best. When we first started OSU's alt format production program as a research and development project, we produced a lot more tactile material than we do now, along with some very lengthy descriptive supplements. What we heard from our students was "this is great, but can we do something more usable?" By working with our faculty and the student users over the years we have developed much more exact protocols and procedures governing how we produce this content. These documents are available on our program website at http://tap.oregonstate.edu/altformat.htm. These documents are constantly a work in progress due to changes in technology and our attempts at increasing program efficiency.
I cannot over-emphasize the importance of getting faculty guidance as you approach the decision to convert or to describe. Most of those of us who work in the field of AT do not have the level of content expertise to make informed decisions about complex mathematical or scientific information. In solving this problem we have developed very effective working relationships with many of our campus departments, Math and Physics in particular. This is primarily the result of our seeking out their expertise as we produced these materials. A secondary benefit of this effort is now many of the faculty are willing to do their own conversion of materials, such as tests and simple handouts. We provide them the training and the tools; they do the work, and then provide the material directly to the user. That is what I call true “Universal Designâ€: give a person a fish and they eat for a day, teach them to fish and they eat for a lifetime!
The Nuts and Bolts
Very specific techniques and equipment are required for the effective production of accessible math and science. However, what is even more important is an effective management and control process to insure that material is converted based on a priority system, that redundant backup systems exist, and that very strict formatting and editing procedures be followed. I have learned all these things the hard way, and if you take a look at our procedure and protocol documents you will see the extent of the exactness that is required in order to be true to the original materials and to provide truly usable access to our clients.
The staff used to do your editing and conversion are the key to the success of your program. All of our programs run on the skills and energy of student workers, something that is readily available at a four-year school. Each of the component programs is overseen by a Graduate Assistant. This is an effective model for a university, and most two-year schools primarily rely on overworked and underpaid full and part-time staff. Regardless of the organizational structure, staff need to be well trained and detail-oriented if your conversion program is going to be successful.
We typically only hire students that are at the advanced level of their degree program, juniors, seniors and ideally grad students. Each must pass a proficiency examination on their computer skills, and on their basic knowledge of math and science. We also test them on their ability to convert information on a basic level. They undergo a several-week training and oversight period, and typically are only allowed to work independently after having been employed by the program for several months. We pay well, about 25% better than most other student employers, and our retention level is very high. Each school situation is going to be different, but most find this to be rewarding and demanding work.
Required Hardware and Software
While all the same basic hardware and software is required for producing accessible math and science as is used for all e-text production, some very specific things are required for this work. In addition to the usual office suite you are also going to need graphics creation and manipulation software, and something to deal with charts, graphs and flow charts. You will also need a way to produce tactile images and annotated tactile images, and there is not a better product on the market for this than the TIGER Embosser. Standard Braille embossers will also work for low resolution and uncomplicated images. We actually use the TIGER for all tactiles and annotated tactiles, graphs, charts, etc., and a Juliet Pro for all true Braille production. The Juliet Pro is faster and the Braille quality is better than that produced by the TIGER. Our users have always complained about Braille done on the TIGER, so we try to use the best tool for the task at hand.
A wide variety of software is used depending on the end product, and we are continually trying new things and contacting vendors to see if we can get a feature introduced or a change made so it better meets our needs. Obviously this works much better with smaller vendors than large ones. I am just going to deal with the math- and science-specific software at this time, and if you want more details please contact me. If I receive enough feedback it may generate another article. For creation of math and science materials, what you use is really determined by what the final product is going to be, as you’ve heard before. Tools for the production of Nemeth are very different from those used to produce RTF-based information to be used with WinTriangle. How you handle the symbology is going to be different, and what you can do with the end product is different. In the long run a lot of this is going to go away as we gain greater access to XML-based files, but that is also a topic for a different day.
Math and Symbology Production
One of the very useful tools for lower level math is actually the built-in equation editor in Word, and it is dealt with relatively well by most screen readers. Once we get beyond about the sixth grade, however, the complexity of the materials and the conversion tools become very specific, and as you can probably guess, much more expensive. Before we can even begin to employ any of these tools we need to look at the process once more. The following steps are critical to your process, and to recovering from errors and mistakes, not that we ever make those.
You need to start by scanning the materials, and this should always be done by section or chapter. There are ways to automate this process but generally are more trouble than they are worth for our purposes here. You will then save the original image files (.tiff), run the OCR application on the files, and save as a text file as well. If you are going to be producing for WinTriangle or the TIGER, it is saved as Rich Text Format (.rtf) and if for use with MathType, as plain text (.txt), or for production as Braille then as a document file (.doc). Now we are ready to begin the editing process.
Regardless of the final product desired, we are trying to maximize the usability; you now need to standardize the file, by standardizing the font, the page layout, and the format of information presentation. What do I mean by this? You need to format the entire document to a common set of parameters. How are page numbers and headings going to be dealt with, what size are they going to be? How are images or formulas going to be identified? What about endnotes, footnotes and sidenotes, and what about all that non-standard text presentation that we often encounter in "modern" textbooks? Finally, once again we need to decide how the images, formulas, figures, and graphs are going to be dealt with. In a nutshell, what we are trying to do is provide a non-visual context for navigation and use from a visual presentation. The final stage in the process is to remove all soft-formatting, such as automatic word-wrap, paragraph indentation and line breaks.
I have briefly dealt with image handling already, but now let's get into the specifics. Deciding how you are going to deal with the "images" is an important decision and primarily focuses on whether each one is going to become a textual description or a tactile image of some type. You need to ask yourself an important question, and if you do not know the answer then you are going to ask the faculty member who "owns" this content. The question is "what is going to best convey the meaning?" with our possible responses being "enlarged print, described image, standard tactile, or annotated tactile."
If you choose to describe, and most often it will be the most appropriate decision, it needs to be a complete description. You need to include all relevant aspects, forces, direction of forces, and all significant details such as arrows, direction of arrows, data points, waves, form of the wave… I think you are starting to get the idea. Specific information also needs to be used when dealing with tables, problems and theorems. The presentation context needs to be uniform, and simple tables are often best converted to descriptive text. Make sure that you are maintaining a logical navigation order in any table structure, and that all material connectivity is maintained. For example, a problem should be followed by a solution, and a theorem should conclude with a presentation of the applicable proof.
Production for WinTriangle and TIGER
WinTriangle, I feel, actually provides the most usable and accessible end product. However, to get to this result also requires the greatest amount of time and energy. Documents that are produced by editing directly in the WinTriangle environment are self-voicing and all elements are fully accessible in both electronic and hard copy once properly edited. The caveat here is that the user needs to learn the DotsPlus version of Braille, and the document must be embossed on a TIGER. For a proficient Braille user this is not a difficult task, and to develop audio-based proficiency with WinTriangle takes just a few hours. It is also important to note that Nemeth users despise WinTriangle, but given the months it takes to learn Nemeth and the hours for WinTriangle, the choice seems apparent from a service provision perspective.
Documents created for use in WinTriangle require the use of specialized font sets and macros for editing, all of which are free and readily available for use from the WinTriangle website. Additionally, a special set of keystrokes and symbol usage must be maintained for the documents to present properly. Fractions, sub- and super-scripts, and vectors all require special procedures to make them usable, and all equations need to be linearized, or put into a single line. All self-voicing software like WinTriangle can only read one line at a time, from left to right. When reformatting equations the established parameters found in our editing documents must be carefully followed. Linearized notation, appropriate element labeling and use of specific notation are all essential to the production of usable material.
Editing for production on the TIGER also requires some additional steps, and the documents are not usable with a standard screen reader, nor can they be printed on a standard embosser. (If standard Braille with Nemeth math presentation is required, use Duxbury Braille Translation software and the embedded Scientific Notebook editor to produce the content directly in the Duxbury editing environment.) As has been mentioned before, all Braille elements are produced in the DotsPlus format, and the TIGER-specific font set provided by ViewPlus must be used for correct conversion. The most limiting factor by far is the cost of the TIGER embosser, but given its unparalleled ability to produce graphical information it is a must-have for any comprehensive program. Specific templates must be used to produce usable content, and all math operations need to be structured in a software product called MathType. An added benefit of this document production is that these documents can be converted for use as web-based documents with the MathPlayer plug-in.
One of the most important aspects to remember when using MathType to edit formulas is that it does not always work, and you may have to revert to another editing strategy or to descriptive text. Roman numerals are also handled in a proprietary way, and related graphics have to be prepared separately and then merged back into the final document. Figures and tables will also have to be enlarged due to the use of limited font sizes and product-specific markup.
The Future of Math and Science Access
The math and science production program at OSU and the use of WinTriangle are on the leading edge in providing accessible math and science content, and are the result of years of development and refinement to get them to their current state. Neither the software nor the process is ideal; it is still a very time- and resource-intensive process, but worth every penny if we can provide truly equivalent access to the material. When we first started, it took an average of 1000 hours for the conversion of a high level math or physics text. In the last year we have reduced our content production values by over 50%, and see incremental reductions each year as the technology gets better and we get better at using it. Current estimates place the cost of fully converting an abstract calculus text to a fully navigable E-Book at about 400 hours.
Further development continues on WinTriangle, as it has been moved to the open-source community and will continue to be managed by TAP. One of the most promising efforts is the work from Harvard University, in collaboration with one of our very successful users and graduates, on the development of a LaTEX-to-Triangle converter. Information about this effort can be found at http://WinTriangle.com. Given that TEX is the primary working environment for many in the math and physics community, we are hoping to shortly see a click-and-drop solution for converting TEX to RTF in a couple of short steps.
The use of Scalable Vector Graphic (SVG) technology will make the use and conversion of web-based graphical content much easier to use and more accessible by persons with disabilities. An innovation from ViewPlus, IVEO, takes SVG-based content and allows you to quickly and easily add audio notations. The user then uses a touchpad and computer to navigate content through an information-rich tactile environment--a technology with great potential at all levels of the educational spectrum. Another web-based innovation, MathML or web-based mathematical markup language, works in conjunction with MathPlayer and browser plug-ins to provide self-voicing math content in websites, currently an insurmountable challenge in online course delivery.
As you can see, the equal and equitable access to all content can be fully realized by all persons with disabilities. The challenge is now yours and mine to develop and implement these programs on our campuses, and train our users appropriately. Further innovations are surfacing on an almost daily basis, and perhaps in the next decade we will eliminate the remaining barriers to instructional access.